{"id": "train_025000", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "ff4bab78e80c9a9d9bd830443d2b123386cce057"}
{"id": "train_025001", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "4255382c306fde439a6ff9b30ec5ffcb2d91648e"}
{"id": "train_025002", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "cfb18bd5b0d09644057ca0eb40d569540e648422"}
{"id": "train_025003", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "f338aa46d8a81ef9032bc14a792395847a85520c"}
{"id": "train_025004", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a16c9b651dd2aec1573bf2a8709c7e0b0fc29364"}
{"id": "train_025005", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e6e4af2472d05beeb792804f810baa4a22133f17"}
{"id": "train_025006", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "2545ccf251a15b4fac1a9737916c9d9e2214308e"}
{"id": "train_025007", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "7b17dfe8535df15bc842d9e7a21a68eb61840317"}
{"id": "train_025008", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "52d50722dc906c197ddfae27efc96bd889e1b7e0"}
{"id": "train_025009", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "62d98eb261727f4d51eeb0f797f569b6d1c891fd"}
{"id": "train_025010", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "3316b54f5280a2a35117814980051840b0a09765"}
{"id": "train_025011", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ce566587b46d9dcdff60738c1ae45eb3cb4e0200"}
{"id": "train_025012", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "c095edccfd191e392c70ec4fe01f75825f2e594c"}
{"id": "train_025013", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f86bffc93999a5e125facfd376c201d9d9773724"}
{"id": "train_025014", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["coding", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "65a5e77374b984001aa1ed2abbb976e3915bcde4"}
{"id": "train_025015", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "6f6a418467f8013c11939c8cb1dff37d3b2d5cf2"}
{"id": "train_025016", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "b5432b6069bc1908a552cb6fbd2912da5517e5f8"}
{"id": "train_025017", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8ea0d31afd0a7647dbef7423f01cdbf8d3d35a0a"}
{"id": "train_025018", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "fc50f527cd983137614dc677692d954e1550d54d"}
{"id": "train_025019", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "626cb477fac7e47a44e52258cee5b9d808e78e62"}
{"id": "train_025020", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "2ba02266219edfe39b4160b5d789701f84811c35"}
{"id": "train_025021", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2e4476e11ff4e78df74de6be019038392170d945"}
{"id": "train_025022", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "526bc0fdd979102f4bfcf8dc8bf545def6244e1e"}
{"id": "train_025023", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "1296c188e410f505d89ab983586b6af829a6d49a"}
{"id": "train_025024", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "36ae91617ca72d828427514769cd69724fa881e9"}
{"id": "train_025025", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "5a6a3f27d786629f31694e2f9f97237fa51e8a87"}
{"id": "train_025026", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2dc9388e3b9ecab34bc12f0758ab172bef77687b"}
{"id": "train_025027", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "85a394627d11fb90c2adc6515da15390bbb41ceb"}
{"id": "train_025028", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "609ac42ae6ed03508e82da9bbe417359e57383c5"}
{"id": "train_025029", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "8b1c0b1622d24ca7d65cc66fb66350a555af24de"}
{"id": "train_025030", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "c19189f159581c0901b85b72cf232ab480dd26de"}
{"id": "train_025031", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "b613d721fa27246dd6678f891b86e236f6dcb648"}
{"id": "train_025032", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "e84d279fa854069d11ee228c616a943c19f400db"}
{"id": "train_025033", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ec4d20419fe820dd1aa38c1e7fe1238df6cc52f4"}
{"id": "train_025034", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "22333c25a1ee1657dea6fb81083e093900b4ed7b"}
{"id": "train_025035", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "fdf87da2d0a66f82b4766676511756aa9d27f345"}
{"id": "train_025036", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "a659e7e923c4f470c12f510e51c3949c3d64d07f"}
{"id": "train_025037", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2b59c3068a99fdfa99bd6a0b96e4da2e651990a7"}
{"id": "train_025038", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "20804c9150d2ba783c4f854e2666e209664cc5c5"}
{"id": "train_025039", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "33a9bc1160df62e000f9577a05fdeaf19be57f42"}
{"id": "train_025040", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "f69d163c0f569c4ed11b58ce5545919f492954a8"}
{"id": "train_025041", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "aac8def98862572d2c86cbc4b22f466f4d7ac6da"}
{"id": "train_025042", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "275cd483294cef0ddeff0bb5d8cb38a6995b0391"}
{"id": "train_025043", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "ffdaf7fe937aca6f8475b0fdd60d8435e93e38d8"}
{"id": "train_025044", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "d19ea25f81f6778f238eea8c3390737a268bbd13"}
{"id": "train_025045", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "5a43e9cb9a374118be9c908282acbe5b31162c4a"}
{"id": "train_025046", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "3e14721dcfd915628a81f972b399eb8da79524ab"}
{"id": "train_025047", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "68d526e3f63dc4b5963fd494539c21d85c8a3a4e"}
{"id": "train_025048", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0ed89a760652a938a4d470e89c71a0177dae13f5"}
{"id": "train_025049", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "67d760708ea8936231e5788ecdb8aff4ec6daf0c"}
{"id": "train_025050", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "abe6fa1b3dd64a0c24302a547b585f4863c65728"}
{"id": "train_025051", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "40c0a1e6717cbde661ec64ec73ceb601cea30841"}
{"id": "train_025052", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["research", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "d5cad8ccdd4c5ec2f7b7a67c45cd21174f19d207"}
{"id": "train_025053", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "bd6fb5b87806e7ed8eb656db6566c6c0f41ee4bf"}
{"id": "train_025054", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "556a1a3b2ec2c116fc700b2f91115d5cb446c755"}
{"id": "train_025055", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "848f7859bc35584c329b40229d002ac166f07c5b"}
{"id": "train_025056", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "91256f7969bb765b60284e8b08710de18f114c7a"}
{"id": "train_025057", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "f28e71735e730c0883a0e41f491c6442326bb766"}
{"id": "train_025058", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "594309bf3f38a9e2d840d32da57b5a3163b4a808"}
{"id": "train_025059", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "90cb27146f11f4cf4c0639916e935fa33e94ea93"}
{"id": "train_025060", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "b2d1f307a0709e33d583c2ccc3d6ed2830d74786"}
{"id": "train_025061", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "33e43ef6dccbaec29d55b7330986a2117024466d"}
{"id": "train_025062", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "46eaf85ef898306bc5b01c7be8f09b9fad95b89c"}
{"id": "train_025063", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "6bae441cc17740590b119c80623200895232d517"}
{"id": "train_025064", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "23d09f9d5c4460363565463f026d0b8e720042aa"}
{"id": "train_025065", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "ae1b8cc806f7297eab520e0f378f7f036a5267d3"}
{"id": "train_025066", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["research", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "70d78986822c7500b779b5d520adf99efda79cdb"}
{"id": "train_025067", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "63c8ad236fb291c06f34abc363461054a45f4165"}
{"id": "train_025068", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "6603666bff0c080511b921a8b277b97ebdcaabea"}
{"id": "train_025069", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "17974173bd1bd9dc73eafdf47990521403507703"}
{"id": "train_025070", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "4a9b0723ceff247c443ac2f2b70535bc6a1f5e29"}
{"id": "train_025071", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "dc53f4aeffc2ea27fe26a643f2e4d986febd7ff6"}
{"id": "train_025072", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f205f4db6a06c571548a2fe5225dd09dac616910"}
{"id": "train_025073", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "a7ec826fe63894d38054484c7befc88ab0290110"}
{"id": "train_025074", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f3f029f1d487cfb8842bc7d4dca60d2f40fc6efb"}
{"id": "train_025075", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "002254e65a281ee7b9c5cb79d10b541480549749"}
{"id": "train_025076", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "1e445e8dbd566013686669bbed7052b2c15e557b"}
{"id": "train_025077", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["coding", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "d64efabe726093192460ae46f0411fec82f73cb1"}
{"id": "train_025078", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "966d03d39e8482a7e6932a476f022378878ea637"}
{"id": "train_025079", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "a6280dfc2367a1ca22aee7b9052f15178637045b"}
{"id": "train_025080", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "77c6209b4bc9e738e67f17b10c3334c1c2635266"}
{"id": "train_025081", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "718d79fd0863bff974d5518d5fe35fc4cfef6e64"}
{"id": "train_025082", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["performance", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "0edf02213cb2f8e6382ef7ca720ee722a094c968"}
{"id": "train_025083", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "9a7eb9e9405a78216f3aaa3fc7c4126de0cff7f2"}
{"id": "train_025084", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c2313b35e965fee0070c06cb3ede4c89639c95a7"}
{"id": "train_025085", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "3c87174a2e8ef2c212989d0e5dfb75e869e3fe97"}
{"id": "train_025086", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["evaluation", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "840e2a84535ef453a05d93073d5cdfc2b9a242d2"}
{"id": "train_025087", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "9b355aa04352046e3a13e1043bf5f0cab2685788"}
{"id": "train_025088", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "62758fc927b1f40968c0b7a6ad41dca98d415c1a"}
{"id": "train_025089", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "7c90998d0030a3e0e4ba65f6c0706a342a8b45e6"}
{"id": "train_025090", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0a03e8cfe249685e21566f58211f88fa9dc08b61"}
{"id": "train_025091", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "68eb2e270b8495b11914c6b2eb68e968e2282919"}
{"id": "train_025092", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "ef32ebe578a9d7d16085b65db9a7e09ce66f750c"}
{"id": "train_025093", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "48d0f8ba61d01a4b82982b7e34de3f9b139f8743"}
{"id": "train_025094", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "4beada11c2131682bf2e0e32853b4f5e8eea8a80"}
{"id": "train_025095", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "22606fc156cd3bbf823d12248b5b3bebfc876fdb"}
{"id": "train_025096", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["coding", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "791971c1fec150cf757ec4bcbaed24e73ac0f56e"}
{"id": "train_025097", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "91c8d404a5abcd9c8e722f6e99b2759096f013c4"}
{"id": "train_025098", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "371f3288855f70b212cc416068717f95ba3ba970"}
{"id": "train_025099", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "cd2f0bde2cfe9a3c01ebcc0cfe4ff3387b3cfd7d"}
{"id": "train_025100", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "e318b1818e9f222d59741820e34fc21e2696ede1"}
{"id": "train_025101", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "a582e6bb4ce2b542d8cfb76b45a96d001b14ecab"}
{"id": "train_025102", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "119279fe5f6dfac2af389452074718d6f436f658"}
{"id": "train_025103", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "d39ceaa4ca505cf43b29e68429b2e2d8c8619275"}
{"id": "train_025104", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "61a73b41560fa2a27ce320254e6ce229b3d1c498"}
{"id": "train_025105", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "8d137bd5f20cfc94286b806584d386e033b00d18"}
{"id": "train_025106", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "3a9b3331b324966b318d470a7170adc054e67d04"}
{"id": "train_025107", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "68712c78fce1dcd7098c6699aaff74578574e7c3"}
{"id": "train_025108", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "23d09f9d5c4460363565463f026d0b8e720042aa"}
{"id": "train_025109", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "1365fb806170aa2c82f80f2d82b96a740aa651aa"}
{"id": "train_025110", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "9668626a40b31e06e4e0c8bdd54870784314121e"}
{"id": "train_025111", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "2bd8b660e74824b8f67a7bb58f85164626a9e816"}
{"id": "train_025112", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "e59adea3a1c4ea8f9e75c9fcd7ac227e305700e5"}
{"id": "train_025113", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["research", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "87652bf061393eb92ece84923ee84a14d93f6a05"}
{"id": "train_025114", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "9391685319931c0df4c28709360f77c048107180"}
{"id": "train_025115", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f205f4db6a06c571548a2fe5225dd09dac616910"}
{"id": "train_025116", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "25e304abf5d25c9204059804843a295c8737969b"}
{"id": "train_025117", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "df17eb44f22e47b99e0e2c806d83d0cca14db9c6"}
{"id": "train_025118", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "c8542c5b5158130f1389d1a229e3a8e2e63d51e1"}
{"id": "train_025119", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2c96f4b713073a50453eb16b5aa1b61b337c7097"}
{"id": "train_025120", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "3113e5d80445c581af4cb9c85bb53de49e13bc41"}
{"id": "train_025121", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "b08b39c4d64d6b987472ad280f4bfd84e0b39caf"}
{"id": "train_025122", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["research", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "5d5dba248cddb54394e1d5ed8291f0b94fb59abc"}
{"id": "train_025123", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "cbc98211fef899e61e554a5a1f53af3b1b923378"}
{"id": "train_025124", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "ae03b75d1ab574885d8e1058ffe227b89bb3fb66"}
{"id": "train_025125", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "6908066078580ab42b46f19816605708e3973eb0"}
{"id": "train_025126", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6b6c9d209b57a28193ca3c3b4a4c90ce25ccf796"}
{"id": "train_025127", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "34c1ccb2b319ac54b771fb671e49b7571818f688"}
{"id": "train_025128", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "5c19c298a40521332a0a7a1e812cf9fb5faded80"}
{"id": "train_025129", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "cec3930ef1539c2668a6534459e61937e7285922"}
{"id": "train_025130", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "5bb96184c143f6b0addd4045ab98cde1620079b5"}
{"id": "train_025131", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "37f971bfd486f0802b9595fca7bd105979c4c66e"}
{"id": "train_025132", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "8f7a3628a73ea1b612eedac6265b80a87345290f"}
{"id": "train_025133", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "504bc69cc928ee2a32d52d22b4f362d5ca219b24"}
{"id": "train_025134", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "76acfaf04666d8cdf5288bfd697d83ff325e614d"}
{"id": "train_025135", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "06880ac1e401e382ef7af054699c57c988e0d0a5"}
{"id": "train_025136", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["security", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "9ec050578ee914e5d70ef897c6c29e1189dc20f6"}
{"id": "train_025137", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8b26ff2da70de2af107a91f66cfa25e84228d049"}
{"id": "train_025138", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "115f6e2d775d940cd51d27c4dc37e55b71257464"}
{"id": "train_025139", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "af0df436744dbee227025602be0c3fc3047a1af1"}
{"id": "train_025140", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "8d45ee23d46bcbca05a155d295992ddb3722b1db"}
{"id": "train_025141", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "967ae8353212f82c7e82d2832e9ec858c6677f24"}
{"id": "train_025142", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2821f9423350fc2b1ab92e8c07acf17d9d8ded5b"}
{"id": "train_025143", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "d2548e761d92fa6f53aa446ab768c1ec37d54c8b"}
{"id": "train_025144", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a3ecedc758aaf938aa3791d876a2499f4e5dbdb1"}
{"id": "train_025145", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "938117ae8b4dc738ca09042403e5b07265761500"}
{"id": "train_025146", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "79f985e3f3aca6ed482a17d1e0c4852bcabc9068"}
{"id": "train_025147", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "bdf6723d2cf4961032dbdb646a33a8e42dc8ed2c"}
{"id": "train_025148", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "d83e7f88b548409404a6d419cb14a1c411f82dee"}
{"id": "train_025149", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "764658f543ac8bfbc88cc7d8637a3089c29b563d"}
{"id": "train_025150", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "19f6abf555f095f87c55e17b0ad10dc890dd91c4"}
{"id": "train_025151", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "874fd218fd3d5dd9b95bd30bb3ac6922f2241393"}
{"id": "train_025152", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "29cc0a52d1b1400ccdd78752f283dcf7f4b335db"}
{"id": "train_025153", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "b691579cf5790e4223289606cb78d12f7f6c9f65"}
{"id": "train_025154", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "0776023f740e0efbb708f207a5a561af09a79817"}
{"id": "train_025155", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "88207863e2bae6bdfd00cd03165f3bbfd2a5c34a"}
{"id": "train_025156", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "50aa57d1287460ff41bec4ed9ad04252018ca431"}
{"id": "train_025157", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "2cfd709b5f0b7878635daedd111e50fa63a2bc2a"}
{"id": "train_025158", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["security", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "094fe32940d4dc835c614e5b923e73f84b48df20"}
{"id": "train_025159", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "662780786966621dbf773f2d851b04f09303ee3c"}
{"id": "train_025160", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "c716715859fa4f18eac6dc579f6faf4a6baf735a"}
{"id": "train_025161", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "3d408bb18f315ce30725f226f3103e3e7addd271"}
{"id": "train_025162", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ae7f3e83a61317918e97adc7c63ffafe2da869bd"}
{"id": "train_025163", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "899a21cd45c2f03b1ec13be9b1f72e031f3e4b07"}
{"id": "train_025164", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "64c59c1600454e1497b0e89a62a3ea9e510c89d3"}
{"id": "train_025165", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["performance", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "e9c15abc9eefe55c266945f616f50d76961ec359"}
{"id": "train_025166", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "8a9847ee9b07102b44524a88e5e2d7e2945b9447"}
{"id": "train_025167", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "49f53647eb63c1af104021c4bf86ed352ae94213"}
{"id": "train_025168", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a3a58f30ca68e9e5a9c6a401f8d3b4c4e90919f8"}
{"id": "train_025169", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "42505c6d293b92270e442d7ff2fd1ff27d9ad39e"}
{"id": "train_025170", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "13b0951c3a171fbe1f50dc10063e9c25a5cbab66"}
{"id": "train_025171", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["research", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "894e2194fab52525c78d34f32ec9426c9375b43c"}
{"id": "train_025172", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "3ae4e9a0d3675ddda147282cb4120508f0428afc"}
{"id": "train_025173", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "ffe8dd698941844c01298797fac3ca73fe14a305"}
{"id": "train_025174", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "7822553c4e576513d4f5940cb50500f42e84ad6d"}
{"id": "train_025175", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "554787c2558320679f6d042ed450e985a9332d7d"}
{"id": "train_025176", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "cc7e5d5f30b32bafec946cf0c6057b007a844519"}
{"id": "train_025177", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "bccced7dba4ff6b2ddc5e6ff3b6ef2100325d44b"}
{"id": "train_025178", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "5ebc23b65fab82ef7e06de7b6f6bf649f2e89887"}
{"id": "train_025179", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "cbdf9b06be1d0eb7ea3c9243e9dc68e7bc4530d9"}
{"id": "train_025180", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "508ed643f92c3c289d99f508d9611542e3e13ba3"}
{"id": "train_025181", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "bdce22dc73ecc87e8c8b2c810c183a08112be242"}
{"id": "train_025182", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "52ca68481091fc42a757e0e17d1e07616cc099f1"}
{"id": "train_025183", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "b603bb998eab2b3612c4b3d77e80753f3447c3de"}
{"id": "train_025184", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "07a2af5e8b18a939d7ae2b54485017afbc65257d"}
{"id": "train_025185", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "ff4e1bdf7d0ac5c1771b1646089459a9f17dbb5d"}
{"id": "train_025186", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "df4034ef59a61ad60ebc69c768da1e246b429215"}
{"id": "train_025187", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["data", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "1def973fdf19d624e61e6c2968812bcd79bfdb64"}
{"id": "train_025188", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "20f4f08036fb94fe1f71e70709db468dc380bb85"}
{"id": "train_025189", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e7f58d5195f7efb8fe0e52d887d225fba2249f64"}
{"id": "train_025190", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "0f5336922b7430d06ab0c1106cefd2ad2923bff2"}
{"id": "train_025191", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "a49da0eff2fe9ae97e3a0bf4a54a456b1ac39ab9"}
{"id": "train_025192", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "4efd9778151c92a48ed04a62275b8abf317c4e27"}
{"id": "train_025193", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "d9432175671be6a801fc73cb11f4aef1822d4699"}
{"id": "train_025194", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "6e24dee98e1af62dc9ee8b1f8bf41a078a889661"}
{"id": "train_025195", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "4862643bb76e797bf7d0e507d3a1bbf30cae121a"}
{"id": "train_025196", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "03304146714d1e931a26e023e4fa6b9885092cf7"}
{"id": "train_025197", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "4e78cd2fb7b16cc7e954ef89faab53b93a3b4adf"}
{"id": "train_025198", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "f97ea717caa30943a7bded196b55b56f9336c82c"}
{"id": "train_025199", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a0840c9572150c70a1c33a57c8f0e5aaa1d0c237"}
{"id": "train_025200", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "6a3241fc19a18814736c053d2fee58be00db08ef"}
{"id": "train_025201", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "961d0220e8ad017070ff1e72d7632dab3925ce68"}
{"id": "train_025202", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "0f6ad8b6a539bd466f089065b46bcf61fae7d1be"}
{"id": "train_025203", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["data", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "64f2871ef0aafc9710074b28bca62195808dd6d8"}
{"id": "train_025204", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a9578a6f1c6a2ae22cede2490a7a29fcb69e2653"}
{"id": "train_025205", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "ecf57d101fcb7225b1239090bf505098e26d41f6"}
{"id": "train_025206", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "366c667e83e68564812c987411a7ee8157d414fe"}
{"id": "train_025207", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ce98f526c6111622d0faae43749493c144a8a236"}
{"id": "train_025208", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "bce755450a5f875a84b6332d99b08337d6864827"}
{"id": "train_025209", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ad1e472ebb627422c0863192c51dcd0cc8184851"}
{"id": "train_025210", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "843dc72b06d4c5603aa4f28c779181b1d2a60b6e"}
{"id": "train_025211", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "d5fb23ef48cf2a42a30153397f9ccbe60da4e0e0"}
{"id": "train_025212", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a1154cc33b0a592621aa53b3410743206c0aa35c"}
{"id": "train_025213", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "cbc50b5e45e05128a9b204dddeba54de0bb6b12a"}
{"id": "train_025214", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["performance", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "b765cd6afb62606fe18fcd442ff4a5c0e76deff7"}
{"id": "train_025215", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "00c90377ec99444c16bad3aab5b5f7a9ea6b707f"}
{"id": "train_025216", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "7bdeb420ea6dd790e74a38a8ba903f93e25269b6"}
{"id": "train_025217", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "8a2eccb0cad8c138de9499cc6f659bcac361c60d"}
{"id": "train_025218", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "c2b7bcded075d67de720d35a4a25ba126fab855d"}
{"id": "train_025219", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2f4923325a009b2f877f27c6f6339a01d9a6baf5"}
{"id": "train_025220", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "559a01edd34635576b5305368d0d64beba01ee62"}
{"id": "train_025221", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "79e3400d505f0d7976b31e09bbc06e0a387dafb6"}
{"id": "train_025222", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "31506cb1d0f68317059ce232557d0a6718a8ed7e"}
{"id": "train_025223", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "9d9f0280f4a75701b5d0bc266909699d6fd5e44a"}
{"id": "train_025224", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "437c93cf32c47b162f1c780142f7751219c241b0"}
{"id": "train_025225", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "cede688cfddc8e5e472cb1ebde148c4423755180"}
{"id": "train_025226", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2c37560364ff125b94d3706bc128d4e31a2f75f2"}
{"id": "train_025227", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ed13eeee6b07411e5e195c156a407f431dd033fe"}
{"id": "train_025228", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "d4e7f1865afa57f4f68c7803a6ac3bf0d1134e5a"}
{"id": "train_025229", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "3a0c86c2a75fb51f6d1220fdb2929b4dc586aa0e"}
{"id": "train_025230", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "05cf77f4da3c8882c1ff212715279a43a00eaa45"}
{"id": "train_025231", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "46e0e54e693bf39137d83468cb28c8dacaaf23f4"}
{"id": "train_025232", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "9dd6e4246a82398840e1f794d3312d20d9302b9e"}
{"id": "train_025233", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b78924a3ba0390ade31a8ba720edf9a595bb91f1"}
{"id": "train_025234", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "bca5989806da4738951faaab5a3289cac53b453a"}
{"id": "train_025235", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "20e58a6e920c5bca771726dfbd0d31a608b9b186"}
{"id": "train_025236", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "7872fdd9ae2cfd48b3f0206131eb506621558e89"}
{"id": "train_025237", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "56df551d69997ef9d49cdf7156c65b6f25b4488c"}
{"id": "train_025238", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["data", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "19c6e025d39a7f8806fefb9d0955b7d749bfc1aa"}
{"id": "train_025239", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "269bcf764180e550d2894d87fd208d2e0d4be3ed"}
{"id": "train_025240", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "0c01ffe135fc7fb7e93e26461cb4ab52552fc3c5"}
{"id": "train_025241", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2809d5345ec2ec27cbbc0ce697f0f1bd7f8b98e4"}
{"id": "train_025242", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "65624845aaa9763bfd79592e725baff70d342c31"}
{"id": "train_025243", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "d6a23a74cad7604001a8f3e03f0f57b512205d62"}
{"id": "train_025244", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9d549b5b1f49ed1a33acc0d8926ec141ca69f231"}
{"id": "train_025245", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "bdc5826598fa896af658672efeea68e21de27938"}
{"id": "train_025246", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "b08560ac4f9168d2df0f1533b8f11f675faa8ceb"}
{"id": "train_025247", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "4dfa9c0e8c79e33456e10f9acdb0b883ad77b281"}
{"id": "train_025248", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "01da951f9a1dd085219b01b210f4916df612f15f"}
{"id": "train_025249", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c40a5689bd2eb7e3402bf13878698ab2fe73dccc"}
{"id": "train_025250", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "cd900239381b24e7fa0a57faaf0551a9e694b95a"}
{"id": "train_025251", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "60216b7d7633194db4caa9cd1a52cd3552e5d9d9"}
{"id": "train_025252", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "85498308f557a98f5d1782019f2a6e9db8802a0f"}
{"id": "train_025253", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "a5cfa2038956f6fd1cd00167e53001048dbf54a7"}
{"id": "train_025254", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "cb35ae136a998d4e40962d3c94e729c35a59b2a4"}
{"id": "train_025255", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "9676dde5ea1d92507f76fdd75ba970842ea022b0"}
{"id": "train_025256", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "1fe42918c69bfe6dd4352f80a960834c35dc53ac"}
{"id": "train_025257", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ca9890536023f9fd3952919024fa695e47a021ea"}
{"id": "train_025258", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "865563eff409c09392fda94e5cbe62c31334a310"}
{"id": "train_025259", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a88e02120226abcbbd7f466b3f588420fbd6e3e4"}
{"id": "train_025260", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "523ff0e730ba1696b3d2fc456155e6ffdf8b5d04"}
{"id": "train_025261", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "b9c66f0816a42b64c391a671e3988aefd061f170"}
{"id": "train_025262", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "9cd44e89796c73b3c48e6208a3c352626bb0f6c1"}
{"id": "train_025263", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "6fa1c0bed6d782161608a3590a83f4d0bf8604fd"}
{"id": "train_025264", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "2d0f4ee05fabed0d88bfe76f3482933ef6b83f0b"}
{"id": "train_025265", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "67be2ce19a65f4a5b7acdfeca03c6702201d15c9"}
{"id": "train_025266", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "236533a40e2c9002301bc23cf36bd9193cb0aa64"}
{"id": "train_025267", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "fb14b8ca586968424a38d90afc74ce5bf74d11e4"}
{"id": "train_025268", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "f793ee0d8a9d8a13cf7365878d3d0543a5aa6e8b"}
{"id": "train_025269", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "061fcfe72aee4f01f845f0ab731895b0253846b2"}
{"id": "train_025270", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "b7bddb7e983462fda6261c718c726c77d0d6848e"}
{"id": "train_025271", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "716f394920728c66079f43a6a420884b1045052d"}
{"id": "train_025272", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c73a234dbc6b57f3f9a23646a8b178123820f588"}
{"id": "train_025273", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2984d5a866468404d0861a73d8b6e8b8e3de0ba7"}
{"id": "train_025274", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "1cd28719665934b55db7b9c835cb3196529dacda"}
{"id": "train_025275", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "1e911462289e599ea446cfe91ecc8a6669c9f806"}
{"id": "train_025276", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "86056c4caa7412abb7246293b985975235be18cf"}
{"id": "train_025277", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "b8468d58a956a5877081f619e128725b2fd35ecd"}
{"id": "train_025278", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "335c00e01da1946707ec442fdeb33f3ff9b68291"}
{"id": "train_025279", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9f6878b293cf2962219880b8a5453f8dee3bdecd"}
{"id": "train_025280", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "f51e605971424d5c188907d3b2902e2092393bb7"}
{"id": "train_025281", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "d4b94e6725471a0cdb5b14c6e235e3885872eba5"}
{"id": "train_025282", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "76c9892c2bd8e61f0838cae655deb0d9803afe1f"}
{"id": "train_025283", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["data", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "6b0c8d0c55b2d9a2ae6c9409052ee243a3c92799"}
{"id": "train_025284", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "6a54dbc4aff1d8985044fdd45dbe87daf7265f41"}
{"id": "train_025285", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "aee04f5bf98ffbf659ef3c7c9a4e07cade92b67f"}
{"id": "train_025286", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "d114a40ab2442967accfd1d508db5373d53b7ae8"}
{"id": "train_025287", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b43770ec17ea63d2d99694ddedd70a21679167ce"}
{"id": "train_025288", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "cc25394b54a2b484194354f497c0d285c4fac3b2"}
{"id": "train_025289", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "4fa501f6cddcf421831b4c0a6e07c64fb2d315fc"}
{"id": "train_025290", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "54776fd26b1c36a4e9f9142abfdbb8f224453a7f"}
{"id": "train_025291", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c9e8081ccd358ec331de8ef68d4bcaf13e74c695"}
{"id": "train_025292", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "81ab9745f764b1e3979fca5fa63a178be84d1aff"}
{"id": "train_025293", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "6a4a09c677987c2b3f7b2e8515a523e40f3e3518"}
{"id": "train_025294", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "5bbb3495afda73c06c17adafb20b90f17cd3b7e2"}
{"id": "train_025295", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "40c7e8923bdc3352e80e0be5a8f5f9a9235c35c4"}
{"id": "train_025296", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e671c26e404a0679b50c009d6f354d459529fc3f"}
{"id": "train_025297", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "836f2a71bfe4d59eb391e1b34fe66d4d85d502cf"}
{"id": "train_025298", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "01252c19fb9d06c0669971b0c6d6561512783433"}
{"id": "train_025299", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f599c951c97f6f31c8927a1c78d4f5afed9af26a"}
{"id": "train_025300", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e80510af113eccc1dd930803af836df69e1f3038"}
{"id": "train_025301", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c99d57f9dd4b73b8de7c9d6142b77117d9ba7bab"}
{"id": "train_025302", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "b7f28158106a9db8a7c2b221ec1813f5ed9cde64"}
{"id": "train_025303", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "9e59f20082931f76b936e49794f1182ddb66c1ca"}
{"id": "train_025304", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ac5c60522c341a2cd333785ab6d68277f64070ff"}
{"id": "train_025305", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "e4563e512d0060dd1a8c4eadea57c81983ac7dda"}
{"id": "train_025306", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "fdc73edf8210b9aa37dda4f24caee0379cc5bc64"}
{"id": "train_025307", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "2af146be4e68ef71b815eeb735f7bd8858a74511"}
{"id": "train_025308", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "e569746691239795d6dc6c56675740a7cc9454b1"}
{"id": "train_025309", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "5034f48718b7d6743271a07b94e7cc7bb9cde93a"}
{"id": "train_025310", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "f1705c3afe9fce8538e268a86948df5956892cd2"}
{"id": "train_025311", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "58261500d9a756c2621d49e10b9f9aa15ceb8abf"}
{"id": "train_025312", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "32c63f69bf9a5dee88199dc4889c6b73f8733786"}
{"id": "train_025313", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "5a8f8f4ed66ee4b43f08834e541c8b843ef922bc"}
{"id": "train_025314", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["security", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "5705ef3eb4ad3a5941460535e813e238c10fa9a5"}
{"id": "train_025315", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "9f2b84ecc0e1c66fbc94dbc1c17f3f8e3d5bc4d9"}
{"id": "train_025316", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "d8847f03c0800f45b6b7d5eaa315540a131905a8"}
{"id": "train_025317", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["data", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "564d1850c3db42ded9e66110f0cd9e9c3acfa351"}
{"id": "train_025318", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "75d07132674ef89d28bc2250b6eb2042f44f5f5e"}
{"id": "train_025319", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "97265e9b20793ae5d4ecba0e43a0d66a08c6260e"}
{"id": "train_025320", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f8d5f3c0581824a9f96c2a4a85501d4682cf4e40"}
{"id": "train_025321", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "dbb56d174263954747cf273240477180965059a5"}
{"id": "train_025322", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "357a7ee954cc888d18082a1da8a80cc852fbaab7"}
{"id": "train_025323", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "e695e2971f7f617a4277a652742f47beb7f9faf2"}
{"id": "train_025324", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "711f4799ff1519d217c9be0062ef1124ac6acc0f"}
{"id": "train_025325", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "30c63915e2a431a60049b78cee673e5603b1c33b"}
{"id": "train_025326", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "47034eca184aa87e0b7f4f02b4b12ffdf5797d98"}
{"id": "train_025327", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "af3de257c7b0269f271f9a2ed4927cd846d54f73"}
{"id": "train_025328", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["research", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "d04a9af5a6e67d6e76b304cfda3eb3b27cd1dabd"}
{"id": "train_025329", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "b61425a320a3a8ae096f733d2e8b74e285058153"}
{"id": "train_025330", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["research", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "d69f464718321f985093c9c06d326ab414946461"}
{"id": "train_025331", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "e9cb4d4cd8de6fcc4d5cc2ab5c0eab42e7ae7c1c"}
{"id": "train_025332", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "cdb8a13e270e290d4803de156dabec8865f1ef2e"}
{"id": "train_025333", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "f93eaa471a7e4b0692b1830c206192b339ea08eb"}
{"id": "train_025334", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a3bcb6cc301ddd15898f5cf492c498a72467ce47"}
{"id": "train_025335", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "9638f31b1d86355deab19dd28e80adaef5a475bd"}
{"id": "train_025336", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "4dc55a38192f72f2584ca14d6cf72329381178d5"}
{"id": "train_025337", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "bc56a9afcef3ea7254ef6a760208f2ebdd51ad21"}
{"id": "train_025338", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "cd598faa60dda991a6ef7c145763f7966cb654d1"}
{"id": "train_025339", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2a4c28f357742b6b873a554a993399be62ff68eb"}
{"id": "train_025340", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "7f791ac4684c1092d0fc6365c79d0f4d97ec6de8"}
{"id": "train_025341", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "a46a18e22f6ccf0abbd595908198b06be92c6aa3"}
{"id": "train_025342", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "a93a90aa3b2b49cedb8a04512e9140a760af5529"}
{"id": "train_025343", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "0876d6937faa40f930730942357baa5ce0ffc8f8"}
{"id": "train_025344", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "4b745cd5ac6555ab831633386e4b4ec133063a9e"}
{"id": "train_025345", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "beaf07cd94f5b0b7419b40dcf2a541687343886a"}
{"id": "train_025346", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["performance", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "fece2a5de59e8b63f4c12aa64c6fafe1ac3d448a"}
{"id": "train_025347", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b484c1da7358bfe80fa1d4efa919148484db0d30"}
{"id": "train_025348", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "f5d5fca7d9335e2867172619b4a07969ab3e7954"}
{"id": "train_025349", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "cadb0979d1b540913665b85222a172d82b8e3822"}
{"id": "train_025350", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c142a55f9719e525d0f15910212b1c554b80106c"}
{"id": "train_025351", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "7b395e83e0f82c7afd5aa7e0b5f7e0592b956506"}
{"id": "train_025352", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "0dfcfc85180e52322d6cae2ccd98d14d8a08722e"}
{"id": "train_025353", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "0ebe5138ff9e100c4627555326c11661d3b5a63e"}
{"id": "train_025354", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "047638ab34e739fa9c22d22386d222daa5a0db1b"}
{"id": "train_025355", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "1b170f75f504f05ed9530d4be83701b586a11855"}
{"id": "train_025356", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c81e01670e6d64cb5c2b5c61ab01ca76cdc4a5df"}
{"id": "train_025357", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "1bba0756a3d4bd69e88a70d01f11eaa61ae8f274"}
{"id": "train_025358", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "3427ea83fb62ae1937c012cacd8be27718db2b1e"}
{"id": "train_025359", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2ae6a3d2e9b285236b58591d45c310e9acb54879"}
{"id": "train_025360", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "3fcfb570a349e19b0419c3c32faeb7d8d0c2f7a3"}
{"id": "train_025361", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ed125307f0f99b71c5479f8f290fd5804f3f0b60"}
{"id": "train_025362", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "d661ea578a800f7cdcfc797ccb62eae383afb48e"}
{"id": "train_025363", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "54b37f3f26a1782f822316f4243d8e0d12079f5a"}
{"id": "train_025364", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "389ff5de8cd88a981994da186c3eec2fde04f9e3"}
{"id": "train_025365", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f4593d81785d45a01c1bf2db9aad9ec53d2c98be"}
{"id": "train_025366", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c01dc841cfe96452c1eaef7c07fb34750f334521"}
{"id": "train_025367", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "b22e00424fa1dd4a49ed2e76ca3c883b930415bc"}
{"id": "train_025368", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e8aee0e4749f2f54cb5f4b41ef65849bcf750115"}
{"id": "train_025369", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "003247c03aac727caa81619533398cc0ca24415c"}
{"id": "train_025370", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "dfd7e6be45dd3a37f0c2e59a00316f21cbf25aa2"}
{"id": "train_025371", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0ec3b46a455cab005edf3f0604fda09d2085d432"}
{"id": "train_025372", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "0ebee5c84505e5854d3476e9607aee69c2be7c09"}
{"id": "train_025373", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "eeffedbcd8708460487cffbcab21836bfbb3c2f0"}
{"id": "train_025374", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "b4a753548dc31e4adecb47f1935f7f8dfe1bfcec"}
{"id": "train_025375", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "72491b00073f861742cd478154c154eef230e398"}
{"id": "train_025376", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "301ac523373da6c37e620328fb57003463f40cd4"}
{"id": "train_025377", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "8c86616364559ae4741be0e720e3dcf00ac40cc1"}
{"id": "train_025378", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "9de7345bb3d5f4b154eec762a0f4f9bc6b4a3d2b"}
{"id": "train_025379", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "93691cbf3cf750f24629f60fffec29fbff5d50e0"}
{"id": "train_025380", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "9ce9bbc9fc6c7143cb5ac911298e0a5638933067"}
{"id": "train_025381", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "5ec2d50a19af8dcee15c6c82e375314a13a7d6f9"}
{"id": "train_025382", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "5876c27b8b8384fd7539870f955c9c67a5c6f2c1"}
{"id": "train_025383", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "26d660a6626471eecc052a5127f0f4b6c7e87f22"}
{"id": "train_025384", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c9314f9fafb5b2c0c2fc440534904fc273000f06"}
{"id": "train_025385", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "e78376fcb9271790c240616f99e9b7593b52978f"}
{"id": "train_025386", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "e7a3b67119c0370a9bc0842ec1ed98b280956c47"}
{"id": "train_025387", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["data", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "dac02a1a978c39f81e1cfe2b5b820f66352c3de0"}
{"id": "train_025388", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "319741d31a105ca388d484b2d0927f3eee7ab66a"}
{"id": "train_025389", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "0925d32df13c8139447fbd569dc6c7cbe110db12"}
{"id": "train_025390", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "af2002ff5c86dec75f4a3abdf65a4d362d832842"}
{"id": "train_025391", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["evaluation", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "db8d27bd40f8aff2fa541d298e7a4c727a7e249b"}
{"id": "train_025392", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "37457e2d679284e2402b287ca489a2eec26f562c"}
{"id": "train_025393", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ca5cce49106c3e309cf238301c55778253be6ffb"}
{"id": "train_025394", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "eef79851d0679a571745af44232d740f72f9804e"}
{"id": "train_025395", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2a9d4835f9989093bc6ea0c82fb6c142003ff6d3"}
{"id": "train_025396", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["coding", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "a6280dfc2367a1ca22aee7b9052f15178637045b"}
{"id": "train_025397", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["security", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "4f0d1b7ef9f6ab81dc390b7b88366d8d3367a269"}
{"id": "train_025398", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "bcccc89ebf38e57919dafee549f787a9dd907e59"}
{"id": "train_025399", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e588ef1ff8be9adeb6b3327ed76fdf2ad26a1977"}
{"id": "train_025400", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "bd12357c2b6514036a0cc985afff23f5fa506855"}
{"id": "train_025401", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "098122e8846940c3b4806f5b01f2b3e929da59b2"}
{"id": "train_025402", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "fb68ca49b2cab60fcd1054dd7185a4b9a755f88b"}
{"id": "train_025403", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "dd13de73b164d2422542540f669751027a17c3fc"}
{"id": "train_025404", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "60f5906cef2bc006e32fcf38458e8cf497ee8d01"}
{"id": "train_025405", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "e9dce1ebbab75b51a441d58eb4c59eabf0cdac0c"}
{"id": "train_025406", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "62caa9e90e28e776a3ec84f065aa7d54999bd916"}
{"id": "train_025407", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "7f6cb8f43ea8562573f4e77571db7ada16e64dd9"}
{"id": "train_025408", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "f6cfe9debd72dc194989bb3084f1a88e1e4fd886"}
{"id": "train_025409", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ebeee5977408824882e8458815de347644f5d310"}
{"id": "train_025410", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d514e697d93c8ea193b9efca686c608658972cb7"}
{"id": "train_025411", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2881a5b2975c4a0fc461b97aaf9e588d584921d3"}
{"id": "train_025412", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "e259007daaeff209655608019e1727b7ceca2de8"}
{"id": "train_025413", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "514863eb39f4a903709471a5ea12a8e5b4a315e9"}
{"id": "train_025414", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "d0d328e5071eb73e44d7573a542ab5180d003337"}
{"id": "train_025415", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "1817141f3c45d93176b76a86913932060cebfa85"}
{"id": "train_025416", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "9beafa1a25c8ff678b98409cca3b642f30cc12de"}
{"id": "train_025417", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "22ce7ef0ead00525f591d26081b2bab1fe470da7"}
{"id": "train_025418", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "76f76f2d04aae1f78aee391ff5cab8de08c4cf95"}
{"id": "train_025419", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "af59f67bfb38ec0cc4ce73a47d51b33b2e56b90d"}
{"id": "train_025420", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "76a71a3ccd62e488fb7c39866bc3b62b2c87f85b"}
{"id": "train_025421", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "42f0759037448f70fcbe256ce804f904c61c0a69"}
{"id": "train_025422", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "c94133fa045653ba30ed2ef4ad026c41fede1878"}
{"id": "train_025423", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "075574f79581bd67d4e2c059876e872898fdaf33"}
{"id": "train_025424", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "17e6f90bd1e38b9a6bff00b22a6a518e0b13d03f"}
{"id": "train_025425", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "cbc2354740168569b8dbb154035bed52ae35e3d8"}
{"id": "train_025426", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "9f213a78b1898c6654837260184ae81b67be4b57"}
{"id": "train_025427", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "e8be2768f29d76ec9e3d95483155a5e6ecfa7541"}
{"id": "train_025428", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "eb7c007408e28ea54992e7c8213f032784d82f57"}
{"id": "train_025429", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "4199d82309083f65d92076bb827b3c24ded98f5a"}
{"id": "train_025430", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "3f5fda589bf383a7942d0be4101bd72db4f84326"}
{"id": "train_025431", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "3b88eabaa75baca13cb178768fc73b7a4b7cdb98"}
{"id": "train_025432", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "da11d52c1ce84a91ab3a665025e7a73cec978862"}
{"id": "train_025433", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "3c509973f1fb3e193de9787765d6367f5f4a1680"}
{"id": "train_025434", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "f25ea302af469fbb15541f21f50ee461559fcde9"}
{"id": "train_025435", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b288bb79cd1958f250c3e21633e472de7f9249d0"}
{"id": "train_025436", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c055ae94025c932558abd402aa340015142cba0e"}
{"id": "train_025437", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "3d72623e91a7adc17cdb3786f46ee4dd165b09ad"}
{"id": "train_025438", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "52c5f2273c5c6b3c29857afce1e964325f6ec6b2"}
{"id": "train_025439", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "362f2f9f36070cbad93b6f436e090a456c049660"}
{"id": "train_025440", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "bf7762319edb1c86781583525e27621a1426701f"}
{"id": "train_025441", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c94133fa045653ba30ed2ef4ad026c41fede1878"}
{"id": "train_025442", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "bd45506e0195f21bcc4b46fbc4a813b085d9c7b3"}
{"id": "train_025443", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "98e4bc2961ebcfc87b9fbe391db3d7c8496f0404"}
{"id": "train_025444", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["data", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "4421186002c71102f4e7e9c20d01712d21f5f19e"}
{"id": "train_025445", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["research", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "4f195faf2b2d6384b4ef4f9e166e18f367a7e0af"}
{"id": "train_025446", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e0a52c61a0b7261a746368b732fca9062585c27d"}
{"id": "train_025447", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "94ceffaa7f75921e7409206acdc15a7d25c933e3"}
{"id": "train_025448", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "072d64ce27be2199514d04771c5a1caeebdb22ef"}
{"id": "train_025449", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "7acc8fa431e884ae3866d0734198dcf21eb22f63"}
{"id": "train_025450", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "93760f67837b694081dae7232b8a989a06b404af"}
{"id": "train_025451", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "8f3a2d6d882cde7e2f1ba6336bd43e28db8cb967"}
{"id": "train_025452", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "987a7fad340ce22d3d5ab5b4037116f4da54937e"}
{"id": "train_025453", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c3eaf690ccf5a3e3ba7e38b409a344d413e344d3"}
{"id": "train_025454", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "49fd27fd6a9669a94da1e83dbac1fae58d0530d2"}
{"id": "train_025455", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["evaluation", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6a5d3d3a68c30e9769c69ed721613e8fbb488a42"}
{"id": "train_025456", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ca4ee61b5a4947f6e5e00639682f67a43f595f9f"}
{"id": "train_025457", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "8bc5868e376c6003bd3bf1b84b7d7bed1aa9c988"}
{"id": "train_025458", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "9df64d0785c8e1ff3fc0b1d5ab6c4463aae2d314"}
{"id": "train_025459", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "499673cf76040af3879cc8387f765a0c9253807f"}
{"id": "train_025460", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "5c9aa0fa47537b5e90ade65ea79e29e17082577a"}
{"id": "train_025461", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["agent", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "e7633eca5654c241d15c2618053a0ff20c882b9a"}
{"id": "train_025462", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "e1e5fbdb4e562a280b193af10b9edfad4e7cccbe"}
{"id": "train_025463", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d8623bd4303c33853db78c2482be59effe9f255a"}
{"id": "train_025464", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8b136055fd5ab956b2de686c828a931147e63a11"}
{"id": "train_025465", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b6d19ca66b381aa0c5ac0d7f7a6f6560397edcb2"}
{"id": "train_025466", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["coding", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f46ad2ee9d662cbf607db6436e6dc8a827e87216"}
{"id": "train_025467", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "3d72623e91a7adc17cdb3786f46ee4dd165b09ad"}
{"id": "train_025468", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "ca2ee64993dee3f83d0b9eb725ea85a2a7ffe654"}
{"id": "train_025469", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "40041d5621642c2b3e6e702aeefee666be576739"}
{"id": "train_025470", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "c686e9b148bf580f22850f3540b84fe71b25fe95"}
{"id": "train_025471", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "42bca4fac3a21d431a901837124f395c1754f2e8"}
{"id": "train_025472", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["evaluation", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "e6a90ff22ecffa4a0e30e31714ec2b2317e4308e"}
{"id": "train_025473", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "ed8c88f406e9855d9bff81fafe19b61f7cdf8bce"}
{"id": "train_025474", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "b119088bd48271b745cffbf1b556df4325b96f09"}
{"id": "train_025475", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "aeb92c3dd1689b7444bb4b0dbc5bfa86401b23e9"}
{"id": "train_025476", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "6705b3a9422d52aeeb3927caaf3ec5968799ca35"}
{"id": "train_025477", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "3b03b57964c8a5b512dc5cee7e0fbacc298c29ef"}
{"id": "train_025478", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "afa6e9a5970d52d1b1b636ef9f6af777dee133f4"}
{"id": "train_025479", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "52bfc84d92a70824450ccc3f6883be49225d2379"}
{"id": "train_025480", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c4c67e811942317b762b995d80e790ad55711ebf"}
{"id": "train_025481", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "3d79b19abcce8441c1ce427b83e7fece379a0cd9"}
{"id": "train_025482", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "fa2e027783ae57456eaaa833134d2e03cd64cf1c"}
{"id": "train_025483", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "c813f73d90ef8c424e9388f2d3221b8345abae72"}
{"id": "train_025484", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "3ef9dde0b3886a38935bce35879d88143fd70730"}
{"id": "train_025485", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["evaluation", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "7b9f35dd94e5d077cce6798f2d557d111affe23c"}
{"id": "train_025486", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "cb69aa3c4ba85e61726741b98b3ea0689ef3cc6f"}
{"id": "train_025487", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "623f506288e0e8dc95700479b1f826996d89e964"}
{"id": "train_025488", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "1030c0f689be5094937b5df0a31e247c45240e05"}
{"id": "train_025489", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "4c2f1003f0e0d351b9231d58d1b8f469396f7967"}
{"id": "train_025490", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["performance", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c2771b0de851c32d87a8e74548834ffe3d51b303"}
{"id": "train_025491", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "10047dd8634837cdb32784eedfd195a52d73dc32"}
{"id": "train_025492", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "444c87e3f0fe13a5a3d34266be8d7f4dc8c738a4"}
{"id": "train_025493", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["coding", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "7f93e3617c9e556ec9dc72b4ad2d69c7a70485e3"}
{"id": "train_025494", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "b00eb6e1ccf732628eccaf63d7b225dab7a12bc4"}
{"id": "train_025495", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b3a12b4a5f90aac1585f06480ca8dd4ea942c8ec"}
{"id": "train_025496", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "2db6df6df1bbef8578a67d1d393689f9c75a4ed7"}
{"id": "train_025497", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "cd09d4cb42defa12c6f0c61f183de5263d8ae8bc"}
{"id": "train_025498", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "42921215f4517a0c361ec97b9cc9296493c9c08a"}
{"id": "train_025499", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "33a5fc113de43d9270245b7bc7f820f1c6a9d01b"}
{"id": "train_025500", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "a2ce7c08882776fd71aeeab96449e271b4ea9139"}
{"id": "train_025501", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "f7a9c579b586c56f1fbe640c0d8835a5164f31b5"}
{"id": "train_025502", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "6a6fab19ac55154ccacba55832fa1d919c01faf6"}
{"id": "train_025503", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "57ef51cb2d7e1ad9e0e6d2975e4ff1646f7206af"}
{"id": "train_025504", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "42f187eed915b1b9cc461dd71d88dac7d5ab00f8"}
{"id": "train_025505", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "3042a6e82ef2325163d55f2bba475f3f11d8a3c3"}
{"id": "train_025506", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "97a6ceb3908ba12ce9eb0d8495d31ab55142eb21"}
{"id": "train_025507", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["data", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "875d73fce89ea29591dc9cf9d6aed362f759dd6e"}
{"id": "train_025508", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "f6d44dd8131f1f852e57c6220fdb5bedaf2f8fc1"}
{"id": "train_025509", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "2789be09592473d88ab3dce2a6f7d82d05623179"}
{"id": "train_025510", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "95be4481d88bcd0eaa80cf9b30b60eb2354be995"}
{"id": "train_025511", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "0bef4182fed460e535429e5ae4db75b18f01e2f3"}
{"id": "train_025512", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "718ca5eb18deefa1b660a3cb82dadb07e7603f04"}
{"id": "train_025513", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "5773e1f22d2719f6ce9e61f8808ddac24ac51847"}
{"id": "train_025514", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "251795a5b4f918cd362c1f1204b3dab3a65f01b2"}
{"id": "train_025515", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "b11686bcd95217d5838bfe01912f44d67458f177"}
{"id": "train_025516", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0e7a830edf60a600b899c208ef84b2f58d80b290"}
{"id": "train_025517", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "995e77d4c72d982b08c06d69bb6d6e59ce9af3c6"}
{"id": "train_025518", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e6cdec9a5418dceaf8645ebf67dadf8e99f34547"}
{"id": "train_025519", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a19d8cee7eba4a9e275749692d8f7d4466dd9544"}
{"id": "train_025520", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "64eb109fdf98e1d77b8712fe18200cd88e480dad"}
{"id": "train_025521", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "fedb368be49fbec8c62b34c1b605d94ba2394c90"}
{"id": "train_025522", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "53f22e47271dce66d7cc697705d1adfc8f1b90b8"}
{"id": "train_025523", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "29b9a7b719fe6cb38ff480cbaf4d3742e34e2abd"}
{"id": "train_025524", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "4c1ad3a6bff79287f71c27c9b2b44c8885a7c51b"}
{"id": "train_025525", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "1b498197ad22aa078d377605941b6e491b4502bd"}
{"id": "train_025526", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "d39ceaa4ca505cf43b29e68429b2e2d8c8619275"}
{"id": "train_025527", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f9a288ce5ed5a7f1ead3a8dfd6f715e1eff6d58e"}
{"id": "train_025528", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "57d8e43237e8cc92f1a9b551b475f97beb950670"}
{"id": "train_025529", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "d3b3d9d9e4bd2f27afbd2edf8228cafdd3aa9c38"}
{"id": "train_025530", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "5392bb25dc5cc63b10b9233cb500c3a0c157430f"}
{"id": "train_025531", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "e86dee33901b202f40960e1d6361cc1bbf5e3043"}
{"id": "train_025532", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "244b32b108d5f9bb0e7baa73c8e83989a64b81d8"}
{"id": "train_025533", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "1ffd3f484f3748c72ad2689c9afd16b71a9e946f"}
{"id": "train_025534", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "71697016981aaca265eb6792d95f9948f4d64ef3"}
{"id": "train_025535", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b6ac84cf5ffc7c5e9e749eee9176217f5aae09de"}
{"id": "train_025536", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "cb10615217464c0c0b9f8ba584b3383f8b958e63"}
{"id": "train_025537", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "9e417c037c0977cc3b71e1c9007526b9e4050325"}
{"id": "train_025538", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "661f74f3642eba5f273e079ec0bd467372adff76"}
{"id": "train_025539", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "748a6e819da525a75a09e24322ae425f1c8703cf"}
{"id": "train_025540", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "b562b864486ddf6ed3ceb00f36ec4cc5285c123f"}
{"id": "train_025541", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "44d6de83f38a24f227198303cade0839da82e447"}
{"id": "train_025542", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "321334a26e9dfd74e3f7d8f7bfdd28a6aadaf438"}
{"id": "train_025543", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "5ae96faf8e56d9a77d1baea61d14f6bfafd69406"}
{"id": "train_025544", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ac853c0240ebeacfac30f1ba8fe273abf3260a7e"}
{"id": "train_025545", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "62ce6fa891fde29b766ff4b3eada804d2a629df7"}
{"id": "train_025546", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "e035e658af71f381f768789505acee05ab187d4b"}
{"id": "train_025547", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "9eef229fd9a18a1bfa611e272743a82e41fd7a5f"}
{"id": "train_025548", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "a9989c6855eebaaa4163f22ffcdcab20ff29106c"}
{"id": "train_025549", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "e565d5793a4a21c459bfed9d353b035e7a3201e3"}
{"id": "train_025550", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c315ec07bcac88b043c2bbe28bdcab93995cd3c6"}
{"id": "train_025551", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "106be804f4f77cdd9a41bae83d8908183cfd35e3"}
{"id": "train_025552", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "0ad8eac2eb0c73913662f9a039bd6c3221a8b37b"}
{"id": "train_025553", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "a2ad2355a5784dcaff8f781ab5b0666db63756a8"}
{"id": "train_025554", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "40eaaa8124dce1b16e1eb7a1674e373c7445dee5"}
{"id": "train_025555", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "df5ec76d3f76ab131f738166d9b81d9efa7cc478"}
{"id": "train_025556", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "4825df6d9557c2f7fdc73a5f4335144f7248d451"}
{"id": "train_025557", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "e8622b64ca093843f7cc5f76b475867c2daf9d39"}
{"id": "train_025558", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "6292c8c8630e076e7c81691053f455c35281f379"}
{"id": "train_025559", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "ab329474295282371e119dcd69bfe18d46e0b945"}
{"id": "train_025560", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "91475e0c037933345af7eb10756ec80fa579378c"}
{"id": "train_025561", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d64efabe726093192460ae46f0411fec82f73cb1"}
{"id": "train_025562", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "f5ed7fa6ff72b4f00f7763d5675c6e8ca4fff0d7"}
{"id": "train_025563", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "3571c6bf3719dea0e9a01f0f849bc4e337a813ba"}
{"id": "train_025564", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c614bb0223a2a268b5baa6cfb208de5c34ab4f35"}
{"id": "train_025565", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "2a068cd3d248429eb53fa56b72ddc203353a59fc"}
{"id": "train_025566", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ba42b51ff2838fa89d95b08d5c48502cd1c6831c"}
{"id": "train_025567", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "4ae32b39e316f24b98bd27f47d406b348ed9ff4a"}
{"id": "train_025568", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a619cce14d209b7a7009ede5eca46a0777641e13"}
{"id": "train_025569", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d620483ef593c9b2502b358a1f3433e0d12027d9"}
{"id": "train_025570", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["security", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "d0e4f7885824c5f666b8e2d52753ca12111d8368"}
{"id": "train_025571", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "91fbb42b2091872eb47ef6665dc201c4a5b8e801"}
{"id": "train_025572", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "ed4a63031a93f94a3707afac7bf4d3a4503db683"}
{"id": "train_025573", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "160681b7fece773c1aac141625e90f78b035cf58"}
{"id": "train_025574", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "89544422e432509e87e14cb994c65cb09e41e5fb"}
{"id": "train_025575", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "abc717ef845dce29e474a2ccd0decfe0673f790f"}
{"id": "train_025576", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "f2e240182a3644b57479a4c38264bf657b29c63b"}
{"id": "train_025577", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "bfb0107577aea0b30e08f30a5317a3ad94dd35da"}
{"id": "train_025578", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "21d79e6dcfe91961ca8bc99ff7761839afdc4f79"}
{"id": "train_025579", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "802c77246ffcbd44c2a6fcba42e945888f589158"}
{"id": "train_025580", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "ed1227fb306c33498f4aa38a9b7800c136fea0b6"}
{"id": "train_025581", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "d296e7e77079873d186204de0af718ce32131b70"}
{"id": "train_025582", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "4f086522f1d7c9c9ec1bd9270861c8426a31b3ca"}
{"id": "train_025583", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c5380c6b3eb5fee39a44f3e8ea362bc9c93a10c6"}
{"id": "train_025584", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "94615387b4c9262c075ba15aafcd07e4a92693a9"}
{"id": "train_025585", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "b20b0c2f2285741aa1532713764ffc5cab1a9a28"}
{"id": "train_025586", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "2a3a05c6bec3e0562d38632a82ced67dde72fc42"}
{"id": "train_025587", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "6d17c50e66babffb8f9958013db2f2953477d52e"}
{"id": "train_025588", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "02b9005eff73bdf3facb37999aed2c260a018b41"}
{"id": "train_025589", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "46115203369fb54e6fd64d11bb7d11cf18531d55"}
{"id": "train_025590", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "eef57603c76360572a33f239328d13d3fae5b361"}
{"id": "train_025591", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2457d51059d54d5d543829896f42db3a4dfe4a2d"}
{"id": "train_025592", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f1a96f39ba88c1934fe8864528a0fa1f38aa6aaf"}
{"id": "train_025593", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "cc204a9de94a6cda6b31817472ad572941bff2f0"}
{"id": "train_025594", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "67b32b23e86cf645635f4c47c2d789fcb6e2d2ae"}
{"id": "train_025595", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "17a1f92628e23b4fdacabe2e3f95511e91310d4a"}
{"id": "train_025596", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "8b1cab3b85cb692da4a80adb3096c56dcded4f67"}
{"id": "train_025597", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "5a3ea042f4206641c6aa564adcde6b228186ccc9"}
{"id": "train_025598", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "12134c37be342cc912ed7cd46b373195b7c6173c"}
{"id": "train_025599", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "1e3e57c272156fb163e6f28e8bae65b4117a96ce"}
{"id": "train_025600", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "a7fe2e860831af7f65f4cda105ea8c2163f7ee74"}
{"id": "train_025601", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "278923d979910313d5087c5471c392c3bafe7a00"}
{"id": "train_025602", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "5bf081d5b248c1044a64f24710ad97f75fc07a87"}
{"id": "train_025603", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "104fce281f7e264e5bffbda008ff93c2c768bd65"}
{"id": "train_025604", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "cdb8a13e270e290d4803de156dabec8865f1ef2e"}
{"id": "train_025605", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "3f6da69d11c9d427ec30a77c14fb2d554ccca49b"}
{"id": "train_025606", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "b8a6e7fb12c618557b0c5daba02ada644aab32fb"}
{"id": "train_025607", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "c77c727c0fe2da650518beec2814d7ba69029567"}
{"id": "train_025608", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "8c3e85483192848a5012933bb9847016e078f7a4"}
{"id": "train_025609", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "77b897b2259b7f7acf4560a38c8399b7d4030091"}
{"id": "train_025610", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "01b77535e21b720d26e32a1d7e2d7f42dd6a376c"}
{"id": "train_025611", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "73a6f57777c73502af9461604f3a80096e0a442f"}
{"id": "train_025612", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "91475e0c037933345af7eb10756ec80fa579378c"}
{"id": "train_025613", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ac1a0bd1e374b9291ef3189c65d857c86fad1ca9"}
{"id": "train_025614", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "321037c800d208ed6595677da705466abae5e529"}
{"id": "train_025615", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a097d7d42ee89ce25667fca7b37a5b7da0a30a78"}
{"id": "train_025616", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "08120dedd6857a53b5792af228a87c60b5ea7a57"}
{"id": "train_025617", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "50cf7b639eeaf57fde2577f2d38407e144560370"}
{"id": "train_025618", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "bb8067a72bbd7624f7333b40a588b69378abb41e"}
{"id": "train_025619", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "625cff296a36c10f5ac0c88c7498384a35d184e9"}
{"id": "train_025620", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "9940414d3a3ee8650d1004e35f20289a348a9a7c"}
{"id": "train_025621", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f2375d98a4227f7eb9d7f280948c75f47a6e24b3"}
{"id": "train_025622", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "3539e38f7a81a215d5feaca4b95decbfd4c65c38"}
{"id": "train_025623", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "3632d7b945e37744d2a42ddb24eb3820d08b55c8"}
{"id": "train_025624", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "7f0c7b50fd9135ef9f54e84ab12d27eea10c1ed7"}
{"id": "train_025625", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "1179ffb82e2fca3cce6c47eebd448f33b2a6b6bc"}
{"id": "train_025626", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "1974894efa2ce95e5f91c7dfb020e65cea5fc2c8"}
{"id": "train_025627", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "56ba2aafcca66ecde688af3f52af4fb41fd450dc"}
{"id": "train_025628", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "1179667af74c6e0ec07ab33666f399866e95fad6"}
{"id": "train_025629", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "beef8615e262fc86ed46a350180f3a15ccc82e28"}
{"id": "train_025630", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "71f361557195f1b665354c252a1824866c1527ce"}
{"id": "train_025631", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "cec3a677b37e9979b9018ddf99f504443c6f6216"}
{"id": "train_025632", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "941222a16fca0c243eb92a2173ac01375586a930"}
{"id": "train_025633", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "245c29dfdc00fe19cdcc1fa7cdd906ce571fc3c5"}
{"id": "train_025634", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["research", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "75039968b44b8e5ba9d3f5f019c4ccd4be578eb8"}
{"id": "train_025635", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "5919dbab3937f8a7d42d51fedb618c19363c4fa1"}
{"id": "train_025636", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "8dcc289ed62178dc15c142cb57611bcccfcff1d3"}
{"id": "train_025637", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["data", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "47278affa44a869f9af494ba7ee2a15241d01840"}
{"id": "train_025638", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "84d056b31a865a69e894ad8face63721e93f2313"}
{"id": "train_025639", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "c5187488bfd948540b20d3d615cfda058b73fe27"}
{"id": "train_025640", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "0dddb3d260e763bf7de4d14594e5a9143a1a01d2"}
{"id": "train_025641", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "df22b8ff5094805d99b5df632905cbbe04605de9"}
{"id": "train_025642", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "3b5f18064cc4427d3bd3f60168681782717e9d87"}
{"id": "train_025643", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "5f13dfd7f541fb2617f765eb0d8aad66fc091e01"}
{"id": "train_025644", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "71b81014f75723464196bdb55f40c3927f7fc2a2"}
{"id": "train_025645", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["agent", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "fb8c76db5b6f857bc346c29a3d37f0abdcb7334b"}
{"id": "train_025646", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "3ce758b804a18385c6e9066b0d5b9464df8617cd"}
{"id": "train_025647", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "d490072194eecd2f0ce72f360e83761736a8ea1c"}
{"id": "train_025648", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2d8a27234a8488b4d984da7e061bdb8fd07bf8e5"}
{"id": "train_025649", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "02246aeda5e2de81a1b266c71c22c595f3597841"}
{"id": "train_025650", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "0e940b30ec472c8bd2455ea96f1107e4264810e2"}
{"id": "train_025651", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "7adf01949fdb8a508e41dc7c13f8abc9abd9b6ac"}
{"id": "train_025652", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ffad43a4d102b49b74870e426d57bd55b2825125"}
{"id": "train_025653", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "5c019f9c32fccf3f9e452dddcd62bdfdc99e6fbf"}
{"id": "train_025654", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "16fe2128233378535a2c520fb2fcfaf904af4aab"}
{"id": "train_025655", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["governance", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "504a1aff4c3b0c03fb482cd60c0835dc80799299"}
{"id": "train_025656", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "d9019e54144d513edf30a3668c969d8a0742d20e"}
{"id": "train_025657", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "526c3b9f417fb7efb1e66f1dbc6bfd80487bc179"}
{"id": "train_025658", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "716162ab83510e379a132b06f713e31b22b29778"}
{"id": "train_025659", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "6b6e030362916b3bcac1abb79da2f797006967b9"}
{"id": "train_025660", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6e4e72f7dc36e9d63150e2508bfdfe9821de292e"}
{"id": "train_025661", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "0952e66479340f04cba2bef6131a1e28cc6f2162"}
{"id": "train_025662", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "02e26684f01f136b4894da11d8976eb56d8e7201"}
{"id": "train_025663", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "46320e9b208e07cd53bcb781d399554d06a0af42"}
{"id": "train_025664", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "337228020f9a71f02fa34ce1e267c1f64463be5b"}
{"id": "train_025665", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "ae97d4759cd15241b8be8715379c43c547dcd4e7"}
{"id": "train_025666", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "776f47b535f6ac763aae7e4f037fa46f60b4d09d"}
{"id": "train_025667", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "212f8a1aa90684e9de7f3fa3287e3fa75b1d515c"}
{"id": "train_025668", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "8712adbf79c541468c2912e2a591eafb60609076"}
{"id": "train_025669", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ed72ae003db6d1203c9757fd56860031a4741a80"}
{"id": "train_025670", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "e1ab0cd71d20280fbfb7659ce9d52edf93fd6390"}
{"id": "train_025671", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "ae32c63620d5976494abbf6519427f3f8ce8ef48"}
{"id": "train_025672", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "d79fd2ebb4e7c8043a553c1b9c942a7edaac2b87"}
{"id": "train_025673", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "66e784ea369b5a46a0562d705f411bb27f64c78f"}
{"id": "train_025674", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "4296753438debfe11172155ddf2041ded391b202"}
{"id": "train_025675", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "f832b375c4e6287443581f686e46f7f9cc556f21"}
{"id": "train_025676", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "3874b82bf49e9b9f6bcf26f234d734e292676c75"}
{"id": "train_025677", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "e9f395021e43a7542908a4e60379d7e31b5410d0"}
{"id": "train_025678", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "7cc21443b619546dc0fb22f81233ae3e0518b808"}
{"id": "train_025679", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "74eb982984f08378ac4fd79713bfabcf736604b4"}
{"id": "train_025680", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "31ee8fed58856d237e597edc990f69ea0afd21ba"}
{"id": "train_025681", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "93f1b679f5a2ebe7e0e88698728bb6309b98e845"}
{"id": "train_025682", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "301ba6b640c8fb424c7538cf14537d7b37fda008"}
{"id": "train_025683", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "5ee43c6c3d8e11a87163183772b67792ba8df8a4"}
{"id": "train_025684", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0e7321a355b62fe9949721e93ce4d4a21655bef4"}
{"id": "train_025685", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a4a07695316be59b2d47e84f81bd8341c70e9257"}
{"id": "train_025686", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "cb33da329f879e86340edb5dca155bdc67101409"}
{"id": "train_025687", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "06459876234ce5458f3aa1cc9d1eda4b7e120161"}
{"id": "train_025688", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "b7f2be71215347603591791cfe2ff607dcc6261c"}
{"id": "train_025689", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f39f0a2b80a51a3e4d63145da2ba738b191a66b1"}
{"id": "train_025690", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["performance", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "5622653b88ea664d26e487ba15be32d8f0d162bf"}
{"id": "train_025691", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "7783a5fcbbdef49f62f65cf7d69113d9317d815d"}
{"id": "train_025692", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["research", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "93e02ab3fc4696f906a8bd4b2db804fee4e44033"}
{"id": "train_025693", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "72971499ffcdf30cde9eea808f7c81083f27c1e4"}
{"id": "train_025694", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "c07a7b1ba86b9999016bc8c2f7edc0438e9b252b"}
{"id": "train_025695", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b408ceebc2ced89b454ebf7c4183cb0343f61ddc"}
{"id": "train_025696", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "bdd72856597bdab9f988d7172e209fbfbc6da6ab"}
{"id": "train_025697", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "b6b4482644dbbec0272dda8dfa4168245d1c3e17"}
{"id": "train_025698", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "943aaba7a371c1fdb692fb087bde346f888145ac"}
{"id": "train_025699", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "18101423af537a879cb5a3d1ac4fd5eee2d6d436"}
{"id": "train_025700", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "00f350156cb1c8508984e870a62a24d1363b90eb"}
{"id": "train_025701", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "d64efabe726093192460ae46f0411fec82f73cb1"}
{"id": "train_025702", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "cec11222e05d33baf1e563d218c601c3118572ba"}
{"id": "train_025703", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "6ddc6cde4bed8b419354c25b4984847920c4df99"}
{"id": "train_025704", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "7ce85d944f478e46287b4be3146a9bf927169f2d"}
{"id": "train_025705", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "8e5820bd57d0c79c88b21e454961a9afb1d8de24"}
{"id": "train_025706", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "128b26e7038fecdead43aca7cc30d66ba9dbd9fa"}
{"id": "train_025707", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "31040fe6b6291a8876ee035e279fd0834319d1e2"}
{"id": "train_025708", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "3b8735f2a33347a324956873bc68ed71787ba0c5"}
{"id": "train_025709", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "20f656c3d51fa71d36b1c1078c1728917572d20a"}
{"id": "train_025710", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "9ee1be780ae7ce43a7c19007fb57f969f105ede4"}
{"id": "train_025711", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "b0b88cbdb52c6e24b418eca2316d38c7713916e1"}
{"id": "train_025712", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "900bc418a2e20d7a645f200257cf82cea0d4a771"}
{"id": "train_025713", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "c777a18ee69fe9645cbf53f8028c56322968ec30"}
{"id": "train_025714", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "22199454e2363ac3e79ec5e5ec38bcfdb9d7d57a"}
{"id": "train_025715", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "38e620016bf31441c2c69adbd3f1e4d2305eb111"}
{"id": "train_025716", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "298f21daf3b14ed4bf0f754e64651236fe4a765c"}
{"id": "train_025717", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "dceff455640d55deef1ec300b522f5e0df88e680"}
{"id": "train_025718", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "7c1e2b73ad88e5e8d86fa02edc629b071805e643"}
{"id": "train_025719", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "33c4e6a1e0f5da92192d9b1a1ee81d45a191f8a9"}
{"id": "train_025720", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "6cd9134c39c03db3c3ae74c745906ad3f0e6dc88"}
{"id": "train_025721", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "144379d37767c9b0c8b609782e359489f634b7ce"}
{"id": "train_025722", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "aca413a18faa722b2b8beff812b1758572b4c2b7"}
{"id": "train_025723", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "bda3f4868313e2e2043e46b07db1a8c9f1220bde"}
{"id": "train_025724", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "f4052f25b638a350cf5f22d2b0ed598766bd0824"}
{"id": "train_025725", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "85404d24930b2d9677537dfdf61a6f5b1c6377ef"}
{"id": "train_025726", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a28deee1c995f84bd06452add9c2cc46cc344f36"}
{"id": "train_025727", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["coding", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "7b25e8f5a65c7fa5d457e8a6809dddd9d96a6355"}
{"id": "train_025728", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "cf73e2fdd59054fdbf11580511c3c1fe2d7b9544"}
{"id": "train_025729", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "7f0ab794e58fb04df248819d2c16ea6134558146"}
{"id": "train_025730", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "209e977bd40b3392019e183d78948db3f8c07eba"}
{"id": "train_025731", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "6c1bec8140921d2e164b1fe2f6cddffc81ded98d"}
{"id": "train_025732", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "8e0ff9c5e97003f04528f76457ee2eeed863ae38"}
{"id": "train_025733", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "85b8b2f6d2d87d3f6ba1ba6f784389adf534c9d8"}
{"id": "train_025734", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "bd25a7e75aa2f33cfe27d0b0fdaa93fd0c06b0c4"}
{"id": "train_025735", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "279bda2fe42e931345b400e0e49c3307a8683f06"}
{"id": "train_025736", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "50a2151e69cc11854f3447a24af121e54d0678ab"}
{"id": "train_025737", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["research", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "9859cf2f0a98f6b1a3db1f6c508b00087d421307"}
{"id": "train_025738", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "d78c8b01cb4a895730a0b3184b7f3340f3f69ae8"}
{"id": "train_025739", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "645592ced75782e6fc2dfe2b0435ea8f4fc52761"}
{"id": "train_025740", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "5c34e0aa405d194054c29b15650a20698cb87e98"}
{"id": "train_025741", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "76c52bab886c448be03038eaff9a48e37cf6c5a6"}
{"id": "train_025742", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "8bcf3e2ce945b07fd7de016e369e41b2bd020706"}
{"id": "train_025743", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "35db3d223e108e90fd6f8380e6a2a19319e96c4a"}
{"id": "train_025744", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "bd45506e0195f21bcc4b46fbc4a813b085d9c7b3"}
{"id": "train_025745", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "cf11ee703adc03665e7b0827032cadecbc334061"}
{"id": "train_025746", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "536e15c253b0298f0b1f891d6b10fe9acf66eaaa"}
{"id": "train_025747", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "2bc3592679ad57b1ed3a63ad063961ff761696b3"}
{"id": "train_025748", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "26f4b436a972a6a37d2026cd40addc4e42c967e4"}
{"id": "train_025749", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "c2d94d8fa4480945a54208f3d7db3761fd49436d"}
{"id": "train_025750", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "6647d1bf18d03e51864a9ba06d6669955058e2e5"}
{"id": "train_025751", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "8c4932e7e441c98dfea975d4c73206980dc5d988"}
{"id": "train_025752", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "0f0e5071147dc4e5f929f96a0d88ec8b58b30922"}
{"id": "train_025753", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "553f5e7864fefb22e0fba08f75dba8157ad43310"}
{"id": "train_025754", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a5a6f952b6303edaae8e61a4c746dcadbeb1dae4"}
{"id": "train_025755", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "cd1441097d35b05d94a9326d046f5927f07586dc"}
{"id": "train_025756", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "812b4ffeedad3d3207a75c2ab59fd8e683f9c576"}
{"id": "train_025757", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "1ac1999ba49c245ba4bd2a0bd124596ed9cf887a"}
{"id": "train_025758", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "f77f376691405f242ba11c751200822beb76f15d"}
{"id": "train_025759", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c0035cc91d5a4807a81610cb21e54e6ba230d995"}
{"id": "train_025760", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "9cf3c85c78c8b7778a1419e2b3d6aa0136645d20"}
{"id": "train_025761", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "23f92eae9ace48403d6c3d060138b92364074e71"}
{"id": "train_025762", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "49bad9426edc14d00199a2472217540dbcd80f1e"}
{"id": "train_025763", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["research", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "d620483ef593c9b2502b358a1f3433e0d12027d9"}
{"id": "train_025764", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "21d61423ed92fdf147c0a98a1208c85a15d519d2"}
{"id": "train_025765", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["data", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "02253f77f6c5abbb1fc7166ca98d7b0473ff9e9e"}
{"id": "train_025766", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "e2f8ec3466f641bb521aa9b0524be8f40295233c"}
{"id": "train_025767", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["performance", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "42aac573169f422e024fe4cd8a8602a7b145472b"}
{"id": "train_025768", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "b1082620c3da5ebd5af060a322479a2ae8f5e677"}
{"id": "train_025769", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "244b67c0dba4f5aa1877d696a4e69f5b617d047f"}
{"id": "train_025770", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "dcf22b3b82ceb1d86a8ebe8f32b45d5f0e31bf9b"}
{"id": "train_025771", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "00048be3428c3adf6a84b4b83473656d3599cfbf"}
{"id": "train_025772", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "0b5f078931cb0220c1d45cefb0a671ef73aa0e1f"}
{"id": "train_025773", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "13dec38d420f7ddb7759dd437aeefde604f765fc"}
{"id": "train_025774", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c1c0dc043aa38f08e0a5ff4e76604828109d8654"}
{"id": "train_025775", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "a7b319ca448f4cf5e1d4e2d7816d29a7542d2554"}
{"id": "train_025776", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ac4a67329e64ec29087ae7d89ac7d138339895c9"}
{"id": "train_025777", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "8bf6fd888b62f8885993705ae2ad59a761a99c05"}
{"id": "train_025778", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "e46449e4a4df9b7c405455f95c6e1afa4b1461de"}
{"id": "train_025779", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6e26ec81b43d8e382dd96129a61b2c56e432a93b"}
{"id": "train_025780", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "9dd5b0b6af59146721ac358edee1676e04ce6189"}
{"id": "train_025781", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "775f8fd200146cc609ef804426fa075e8c0ef4f6"}
{"id": "train_025782", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "e3932c23ce0057bfe3068b4b45c5699febdec06b"}
{"id": "train_025783", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "ea04fb2cde37e7d28fe0eeeec17aa7cf1e1483f6"}
{"id": "train_025784", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "91435b273192bdfbdff5b07e6a4082f3bb3927c3"}
{"id": "train_025785", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "311f2eede3507e706da5c5b69841daa8432bb219"}
{"id": "train_025786", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "c3ddd0276fbddfb7f277871afbfe6bd65eecd2c4"}
{"id": "train_025787", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "5d404b57c9ec5d3aca241806bda4853ceadf0577"}
{"id": "train_025788", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "90d5860e16aafd0419e992b8e5c4803192923c0b"}
{"id": "train_025789", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "1b872006ec9ed8e4deae0a4b16555422e0659402"}
{"id": "train_025790", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ecc80c7dcf06b1cb4d472bf1126df6732296bbba"}
{"id": "train_025791", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e12a77bb0d33bbba92889d5215f5fbc8290300ef"}
{"id": "train_025792", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "a73359137c890c84dff29765906a3eb96f4d96e6"}
{"id": "train_025793", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "2e285b46fa7d14dab725963aea16cc335447541c"}
{"id": "train_025794", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "465484ec2f5dadf2664989c8debc48a78959d192"}
{"id": "train_025795", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "dfe5c363a1ba32548eda56111f61c78803f8dffb"}
{"id": "train_025796", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "e3f280d017919269114d12fd4c8ca59030e9c8fc"}
{"id": "train_025797", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["evaluation", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "dc1d88eb16ad6c136602191e7352b092ddac55e5"}
{"id": "train_025798", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "994440c933f68eeb9ce82ebd77c4dcc8053f4bc2"}
{"id": "train_025799", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "c095edccfd191e392c70ec4fe01f75825f2e594c"}
{"id": "train_025800", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "440b21f1218208ddbcd48559bce72d250bad4a73"}
{"id": "train_025801", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "c4395d614a2f153a27b2f8d6a5b961faa35fde7c"}
{"id": "train_025802", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "d597258fa4d26bcc2bf47e33c63b33e4198fcfef"}
{"id": "train_025803", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "bc209825a78bb469e93cf48978a9b70ac6d00735"}
{"id": "train_025804", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "d62430c0494ea0a9dd7c28b389314610a32bb7ea"}
{"id": "train_025805", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "8a45c046a31a15052d95e187662198b7fef6ca04"}
{"id": "train_025806", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "695b71e52cd13da996557dcf72f967e7eb874252"}
{"id": "train_025807", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "93e2702106d0c25a6b512756cb6223414f18c7ca"}
{"id": "train_025808", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "a39e80c91db68f9fe6cf87f2acb9d44e8d807906"}
{"id": "train_025809", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "1402faaef7abfbbd0deb493c2f9235379f031fb2"}
{"id": "train_025810", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "b2f9d9e37a0de78dcdbfbeb887c61cc09ae50a1f"}
{"id": "train_025811", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "f269ad566856cf4447a2ffadfdc258677d70170d"}
{"id": "train_025812", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "925d670bb7843f63c34a890bfefa244a3ba393d4"}
{"id": "train_025813", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f6a5221d9ce41bb00653d5d99d37f4e6194e32e2"}
{"id": "train_025814", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "276e9431353bc86104d687df0286e0daedb5ca6d"}
{"id": "train_025815", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "6602cea9d988a5979b957e478d40c81f408cbb78"}
{"id": "train_025816", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "12d304eef62004a677ab49dd07a0dbd936f0ff71"}
{"id": "train_025817", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "140400ff6e683c21eeb30583b3638eb85b3435b6"}
{"id": "train_025818", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["performance", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "67cc6b96d46d53a899872eb16a4da8a37762884c"}
{"id": "train_025819", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "f12d009fd94e67e8eb26e6d0fb9b3b33a2cdb41b"}
{"id": "train_025820", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "748a6e819da525a75a09e24322ae425f1c8703cf"}
{"id": "train_025821", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ebd752c8963e0d38295bba38247aa4b6058f2b5d"}
{"id": "train_025822", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "3093540377e09eebb6ad67a052eedccd9712188c"}
{"id": "train_025823", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "18e368b34ec4300a37513ab8fc59ffa53686bca2"}
{"id": "train_025824", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "350b4e4b0d643944ca504a426b990e90631768c9"}
{"id": "train_025825", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "f8ca4421cec1b2ce6ab4bb93df21929b7b8f005d"}
{"id": "train_025826", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "b99ab6b66c3cd1c43905e181b6682a0185355b55"}
{"id": "train_025827", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "142a10813380819e812e04d4ca5d4de0feb2266f"}
{"id": "train_025828", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "426aea6f117d5ce95b830ba99c54e3769d6b0c87"}
{"id": "train_025829", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "b9e46799f8a183f17a489612745b0c61d3623b20"}
{"id": "train_025830", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "d257fc2b2d6a42023900602b997f6842fefda99a"}
{"id": "train_025831", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "24ca2e2631cc94860c511bb029ab8fafac1a8351"}
{"id": "train_025832", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "261c531ab1889169573af11d69f2cb95559ead2e"}
{"id": "train_025833", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "6e0b74481a44b0379d791cfb05dc5ea85ef9019e"}
{"id": "train_025834", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "e20a785f126039d471ce4d874bbccfb24b70d4f0"}
{"id": "train_025835", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "61ac6a5b200594613214f2e5d45f6b5463f0bb91"}
{"id": "train_025836", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2fdaa94a56ef7e303f9b8166b610219d8e710347"}
{"id": "train_025837", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "33330e1315bd450b9757849902cc365375bcdbdf"}
{"id": "train_025838", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "c6a018fc6a6e0629efc87652337c153cbba0531c"}
{"id": "train_025839", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "81a48c1bf2d9c068f47104dd98f58f9f0730e1c5"}
{"id": "train_025840", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c35de672765fc518928c9af322ac39e6e6a9ee37"}
{"id": "train_025841", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "06767fe65bc66a312bececa78b76a64a4b955791"}
{"id": "train_025842", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "34b421a63096c27d7c6197fb94c5cc84ed06eca8"}
{"id": "train_025843", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "db72f324dcd9fc0f57bb645b8366946cf0845d38"}
{"id": "train_025844", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "57de9c8205788c8369a5e0c40634421385a81fb9"}
{"id": "train_025845", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "8f201a9e83b5c9d0a117e74c275940654eac0a54"}
{"id": "train_025846", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "af2002ff5c86dec75f4a3abdf65a4d362d832842"}
{"id": "train_025847", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "f2ec51c70a7cd8922328a4d84e43e1b2b259267d"}
{"id": "train_025848", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "287780c98a825db39be0aad02c2522dcc0ac3c22"}
{"id": "train_025849", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "da9b942637130ed1d3953cb6b076279c00f8282e"}
{"id": "train_025850", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "7849dd607d1a5376d937a927ec2f7614f1b60f31"}
{"id": "train_025851", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "d6ceabc3f74c25078d862ea1d68f4ac6bdbb1eca"}
{"id": "train_025852", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "df8ea1372eb6fec5af4dc90e95dc4b48f9fa0887"}
{"id": "train_025853", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "bf563d64b3183f69ebda641bd4e39524fdac97af"}
{"id": "train_025854", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "dadac65941a0a6261592fe38c25ac631644bcc20"}
{"id": "train_025855", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "31ee8fed58856d237e597edc990f69ea0afd21ba"}
{"id": "train_025856", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "dc25fe93a776b99d3ef69be28e6473e81f8bda14"}
{"id": "train_025857", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "a4c2f003a8601c4cbff460851e2edb6ffe372096"}
{"id": "train_025858", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a084deca0468794fe11e04c0ec15788dcf99e40d"}
{"id": "train_025859", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "708e5058ca8f3432a6380dbb28a4037ce704eb19"}
{"id": "train_025860", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "7670a380ce6703608c7e3c9c47e9dd5c4e015d74"}
{"id": "train_025861", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "31efc0bf4c971681c7b2bede5ccb169f74d27781"}
{"id": "train_025862", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "67584d26d93bc31c23f8ae34d1bb98ebf68ee096"}
{"id": "train_025863", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["security", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "4607d6d5357135e047fc12666effd1514887df97"}
{"id": "train_025864", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "aa50d448cf60d8960ae89553cbd790e057138114"}
{"id": "train_025865", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["coding", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "835aafe0943c0f44c37cbc67e8ba16c7c23012f8"}
{"id": "train_025866", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c14ab374b8584a723ee18b082e52b7a6cdb91a2d"}
{"id": "train_025867", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "9f61ee9ce3480deeb79d4e9f8351906a10889df5"}
{"id": "train_025868", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "17dd8ebf2cbc82547759a8fd200f5c0eda5a2368"}
{"id": "train_025869", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "ec1ae5e02626e3b3db1687d4ed759ea9fea53e41"}
{"id": "train_025870", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "6f011b2e04fbf6281040b09b4cd3af5614bf77c3"}
{"id": "train_025871", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "a9756f6f9a48cc18206a2aee27f6f998ffd3b35f"}
{"id": "train_025872", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "7f0c7b50fd9135ef9f54e84ab12d27eea10c1ed7"}
{"id": "train_025873", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "1974b0beb7d5c2b7c84887679406e85149fab38c"}
{"id": "train_025874", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "00dbd27d1126acb24fe7742d90bcc3ac509282e4"}
{"id": "train_025875", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["research", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "5261417af02e460388794105affbbc4caf314643"}
{"id": "train_025876", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "2e27eab700c5ac2a50a8f3bf97f7f780657f66ae"}
{"id": "train_025877", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "aca9683b8478ac63da0945e11edc73b07a32ad5e"}
{"id": "train_025878", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "e91d3aaf89dfc3a576f105bc7010e3b1a9d2e506"}
{"id": "train_025879", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "b6c2e5cd0d9943841ba1764253b2f41759f8a496"}
{"id": "train_025880", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "5c3875792a6d83b956a0b893b21e94191b17c578"}
{"id": "train_025881", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "53562430d9d5a3fb65803433d143c3d220ae8bfc"}
{"id": "train_025882", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "44e45d776b520916a5c448e56219ba56c6aabfb2"}
{"id": "train_025883", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "8aa33947622ad4ae1223fc633d99adf29a7f200c"}
{"id": "train_025884", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "0806539fb76f451a1544686d5f751f170072a99b"}
{"id": "train_025885", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e10229675ba06aaad54eaf9f02490ec1368f6fc2"}
{"id": "train_025886", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "a2e372484588aa88be9d8527b806b6100a847848"}
{"id": "train_025887", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "4e1a304be9fcaa4af06ac0f0b9391e446c07bfc4"}
{"id": "train_025888", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "335ea61cb14c7f2b7523dd0033acdc8ce7cc2201"}
{"id": "train_025889", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "7ebe83fa77e4ba1b4469232923fb26685a886a61"}
{"id": "train_025890", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2c668f993ffb7afadee20327f785b34607a3fc58"}
{"id": "train_025891", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "2a2845a4907f3933fc0c47d660d66f1b2b3ed981"}
{"id": "train_025892", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "d32b633f95f1a5af35ea564c5fdee07c39cb997b"}
{"id": "train_025893", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "6e98e6a4a9fb1e1f800b6070f3ac5d39b881e8f5"}
{"id": "train_025894", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "1972789fac2bac1c1852e7c2e561eccaf84b8b06"}
{"id": "train_025895", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c47a59d3d317275b03f0cbd92302c15eeaafdcb5"}
{"id": "train_025896", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c5cba086148a4fb2339a34237ae0e091d7cb388b"}
{"id": "train_025897", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "e41bc62daa58b600ae09e88a37b098bc5445faad"}
{"id": "train_025898", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "abc5f2b11ca7e71cd6a71bfdd8d0af57c814dd9a"}
{"id": "train_025899", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "9d71bb0a6c26cfb83af2f1c5f4dd9d9ab1506adb"}
{"id": "train_025900", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "f64a705b01d1995a5dfb0bde206c119f71723fb0"}
{"id": "train_025901", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "2ab7307fb8385154d4d1089e8d432a0d786a4ea4"}
{"id": "train_025902", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "cb7628838ad56508eabb38bb56cf23cbac19d9ce"}
{"id": "train_025903", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "30af50af5cbc31571a84d9ee328d188211905c96"}
{"id": "train_025904", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "afec8172a1669851b8742a2a612491da6662bd5b"}
{"id": "train_025905", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "2ec6001b62c95241772c67747bd9f4b72e791bfb"}
{"id": "train_025906", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "79a81aa55b040864ce80625bd33f65a1c9f48091"}
{"id": "train_025907", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["coding", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "b703ee05d83e6e77a0bda08515281a8aafdebfe5"}
{"id": "train_025908", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "0048dcbaa676c76d514fab1a4d1a32ec4da9bf3c"}
{"id": "train_025909", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "9f7a94a121f4f3b45afa3ded47a3fa45982c39eb"}
{"id": "train_025910", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "526c3b9f417fb7efb1e66f1dbc6bfd80487bc179"}
{"id": "train_025911", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "e0064d2eb8301fbe85cb858974990918c329450c"}
{"id": "train_025912", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "507eb762b45e89573ab334effb1086c6e6bf64f6"}
{"id": "train_025913", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "568b3f5885667b29848d17a90944f453c63c74b2"}
{"id": "train_025914", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "3d0df50a9591ff3456a47d0afe3a845da4fa83db"}
{"id": "train_025915", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "99fdd164420e5eb2c4cef9b44ada8181ccbc6a3f"}
{"id": "train_025916", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "9203009c6691a89cc380f60cdfdbb4d8b5ad2577"}
{"id": "train_025917", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "6602d483f35ca02cfa76567283db04d5ef5f4a10"}
{"id": "train_025918", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "10e9c08d357287a412cca0acfc8450d0813cf77a"}
{"id": "train_025919", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "ab25c2300ae7e1cf68df21c1c28e718ac00754c8"}
{"id": "train_025920", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "fb68ca49b2cab60fcd1054dd7185a4b9a755f88b"}
{"id": "train_025921", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "e3e4d1740ea617492b6efa9bbd018ee0064001f2"}
{"id": "train_025922", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "814977036ed130e836499d1bf51bb63cebd82a34"}
{"id": "train_025923", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "b2f9d9e37a0de78dcdbfbeb887c61cc09ae50a1f"}
{"id": "train_025924", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "1a72b00c501f18071727356907810c9e7e7ba300"}
{"id": "train_025925", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "779b38eada577d624574e63b566676219ea0e932"}
{"id": "train_025926", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "6b756ecdadccfd91a4b1d462a222ecf42d8cf358"}
{"id": "train_025927", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "79a81aa55b040864ce80625bd33f65a1c9f48091"}
{"id": "train_025928", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "631d5965abb8d840258fc5b03b07bd3558eb2f01"}
{"id": "train_025929", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "80896d05cd132c17c2499426257601218644271a"}
{"id": "train_025930", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "b5ad848b3a467552dc6801d7ba840a5dd3ff90ae"}
{"id": "train_025931", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "7bc7cd908215dac7b522a7c00a8b4312086c0558"}
{"id": "train_025932", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "2c62b8b71bc14231aaf48703579c4cf74c6478b9"}
{"id": "train_025933", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "68a47d804eba2e57b2ff8ba7728ab9ab541be2b5"}
{"id": "train_025934", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "e296146de0ae08e1d87e0663a2fcfa58ec6edd9d"}
{"id": "train_025935", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "febf842552cdd738201ee3433040edd6491242ae"}
{"id": "train_025936", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "06941dc6f66334fd97678ce761009afd852ab9d3"}
{"id": "train_025937", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "fd7443b5cf00ccfbe7cf2189fd7abe1a1c01c3ce"}
{"id": "train_025938", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "8f5b54531539cc413084f17485ba8b3f29e9b65e"}
{"id": "train_025939", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "0f4b7f8c0daacc1d991863b1f8feceaaded2b099"}
{"id": "train_025940", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "887fcde6e63be1e969c9e0467fe29436b1f685a1"}
{"id": "train_025941", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "357a7ee954cc888d18082a1da8a80cc852fbaab7"}
{"id": "train_025942", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "1dda736b886fd752674bb20da514f0c73648d340"}
{"id": "train_025943", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "673c3f86d6194efef79bd883c2370e13ece803e3"}
{"id": "train_025944", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "65d996ce87d918309595e6aeca3ce88a3750ca98"}
{"id": "train_025945", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c9cd0fc4864dbb4e528bca17c6107b148c74b0d6"}
{"id": "train_025946", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "40073463c38fffcd41009a10183f8c8c94572296"}
{"id": "train_025947", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["governance", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "d58b70336f4be172276809b01eb31f7130c2acb9"}
{"id": "train_025948", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "dc7980699929340416d490334df26a8227fab3d1"}
{"id": "train_025949", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "a8627fdcb8419f4cb591220511f977ca3c71b72c"}
{"id": "train_025950", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c665a38a032b7512e714761899d71f36a9fea460"}
{"id": "train_025951", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "74bc04c800824daa1e19ed81b2014a9593a37107"}
{"id": "train_025952", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "89dfddb695d097fffaeabf3b6c8ec7760d4ea7c9"}
{"id": "train_025953", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "b2638c505e2c3a76674bfd2185cc60e5c175a149"}
{"id": "train_025954", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "945c427e4050e6e19fad2ae971f78cbc62ac8f85"}
{"id": "train_025955", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "9eb2e9f73c443683024cabb6ef5dcc703a9b0213"}
{"id": "train_025956", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e835241685c4476491613698d819e3db804edce8"}
{"id": "train_025957", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f2a687cf4f327673eede6e7c7c2e3551b44d928e"}
{"id": "train_025958", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2b99750dac57e46daf7d149c8c33ca443212130a"}
{"id": "train_025959", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "7cfe96af3fe0958b400b986730bb5f05294986bb"}
{"id": "train_025960", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "70009dc43920ba865757e1ccaa391e3e63119b56"}
{"id": "train_025961", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["data", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "dbca0550631aaafe66f4ff19f436e6b2c72f5bdb"}
{"id": "train_025962", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a62b270e25268ebc9fe750cafc38886931d73195"}
{"id": "train_025963", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c58597d7da75faebd1aa9f4a640c2e911667ea4c"}
{"id": "train_025964", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "0ba8324e520d82fcfa67758769c652109bbd584b"}
{"id": "train_025965", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d97db548be99a559938a715f32d03f4a0312e204"}
{"id": "train_025966", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "df1fdc2b58f00d0e8b3527fd0864d244e0878370"}
{"id": "train_025967", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "76980d4128f4fd5c05c7886b30472dfcb2641e19"}
{"id": "train_025968", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "38d10b8fd72453d4569061c95dbf6776b3a3d119"}
{"id": "train_025969", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "1096bceaad3c5046d05d05983c5f7dc686b2be3c"}
{"id": "train_025970", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "05d0cd62b8f3a3e6f7565f2f8d74d82406d8eea4"}
{"id": "train_025971", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "dd9cfa7cac9a43c3e1bf4b2571e5f853ee9d0cbd"}
{"id": "train_025972", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "f1586beb7692dce6c81751bf3353a53627fa3387"}
{"id": "train_025973", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "219d790455e9dd4be179bb87e4c9f2bbb0943e50"}
{"id": "train_025974", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "89c9233d449e27829890261f5c5f6f0475bd12a9"}
{"id": "train_025975", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "a68ea8b4e125dae98067c7491e9d8698a28cdb7b"}
{"id": "train_025976", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c45d88795bec61609cae8fb8685df8f69a06b015"}
{"id": "train_025977", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "884bf890f028ac74011d2cbc3623cc7962b11fb1"}
{"id": "train_025978", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "f28e71735e730c0883a0e41f491c6442326bb766"}
{"id": "train_025979", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "7802b36cf82193e4ff10a50323664448c5c25b9b"}
{"id": "train_025980", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "40c0a1e6717cbde661ec64ec73ceb601cea30841"}
{"id": "train_025981", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "3f416362d75436f9431b94c8c55c9d781af7dd86"}
{"id": "train_025982", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "e9d57402f76d22f159ae801094d75fe3b2249aa9"}
{"id": "train_025983", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "bf6aaae6f5d9c1f6548854ca4b506b7661d4e138"}
{"id": "train_025984", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "22ed4cca7fc47b2fc00bb4e5f918bb69bb1bebb6"}
{"id": "train_025985", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "31629ff20fa04669b4391d9422aa85d81814aed5"}
{"id": "train_025986", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b1fadd6aaea481f521da88d8b72a97ce5d0cac67"}
{"id": "train_025987", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "cafe6cf5c68a69e661801cd9a257ba5e03f59e4b"}
{"id": "train_025988", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "08d1ce89c2be3155fcee00d4f6150ae36cdb113a"}
{"id": "train_025989", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "2f8112607725779979219aa92d0983a96dec53f5"}
{"id": "train_025990", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "7b6c292cb3c36b3305b6eefba8d7687a410aac63"}
{"id": "train_025991", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "8840faa9303680f4181bba10fdf02320cd402277"}
{"id": "train_025992", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "534688190abb99d54e6e3fb13d21be4ae7c47424"}
{"id": "train_025993", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "6b1c173f307e98f55efc9c127ff49f06b1958e80"}
{"id": "train_025994", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "0de28ceaa8dd726bb12a94444ec17a42578a8a28"}
{"id": "train_025995", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "13268ad56644096213cadaaf9073568a02fbef4d"}
{"id": "train_025996", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "da66f6c91b274a0a491a189f4fe7c03a94b85d40"}
{"id": "train_025997", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "09d978dc81dc51502686c3d81935ee688b1417f4"}
{"id": "train_025998", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "11f3c74e0c57a8bb075906c7c6ad33efc4feafc9"}
{"id": "train_025999", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "080b5386091d1916ec0d73575066dd8677760dc6"}
{"id": "train_026000", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "457712f10d7f83a894d956cdaf0cc30789965c51"}
{"id": "train_026001", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "8842ce66d36ad89bf9fe8684430e26b9fe771a17"}
{"id": "train_026002", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f6f2d08b297091401fdc0cebb0ada299c5ab00a8"}
{"id": "train_026003", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f4a8ebfbb07774b592b3380214d0d0e9fbffecba"}
{"id": "train_026004", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "6f2d7ad61bf7df85c51a883b950b80c357bcbccf"}
{"id": "train_026005", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "27a20a62c0d525016bbd8623edd096aea850d4de"}
{"id": "train_026006", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["evaluation", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f91073b8bb1df2ded60db1de238aed47a1eaad9a"}
{"id": "train_026007", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "c4e706d786de232a553a360278941762e04e1404"}
{"id": "train_026008", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2a4c28f357742b6b873a554a993399be62ff68eb"}
{"id": "train_026009", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "3042a6e82ef2325163d55f2bba475f3f11d8a3c3"}
{"id": "train_026010", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "3ea78ef44d7aeb27ccfd5f0e2c1a4ca6b5623c4c"}
{"id": "train_026011", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "54d873c55babe99d5df306864439d3aac968b559"}
{"id": "train_026012", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "40944a382c72ec09d01fea29ae3aa83ff29b723c"}
{"id": "train_026013", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "30d8f1428c15c8a43406b2dc4c726ba33972fa96"}
{"id": "train_026014", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "772cfe0be0769a198b08991003910c338832d5bd"}
{"id": "train_026015", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "8cbdf15b16cd7941e59cb8a2d6296d153bde6999"}
{"id": "train_026016", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e8190521c7d8a7ff6c4b2303d82d6fcc8715d37f"}
{"id": "train_026017", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "d81891dc4d9840e7b7ce8c25c4ddf8b7c7e01483"}
{"id": "train_026018", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "e468d7b18e1ea6764e1e507f0f8b18fb3c1f935d"}
{"id": "train_026019", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["performance", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "01ba1d3e76af053fee10414988f1348b8ea8ef53"}
{"id": "train_026020", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c82e36b0df51f4546b0cc33c0258a1ed734c63bf"}
{"id": "train_026021", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "58787be06985ee26ed65f13e1f9303e03f5a5c34"}
{"id": "train_026022", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "35f71153ed7c3af5f274830312589421b102abeb"}
{"id": "train_026023", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "2d30c286c88f4b5b13baa772a6c7e0441f6ec147"}
{"id": "train_026024", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ce34716cb54cb26a20481625aa36ef42d5436299"}
{"id": "train_026025", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "9825428d21278b1db8c21b46da390ffed8eeffdc"}
{"id": "train_026026", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "450625fef5b2871c0fc6d029938afafce4f323cf"}
{"id": "train_026027", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ddcb456cb3b92ec3dfda5eecdf6694e8c4598c8f"}
{"id": "train_026028", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c90253594cb24d85bc44c64e484dd50f9511078a"}
{"id": "train_026029", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "fe27b85a0426f61c74f84a9225f8581cb0a0dff0"}
{"id": "train_026030", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "b73c0f017b6dad2f9d152a75f49e8e7a3fcf5503"}
{"id": "train_026031", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "67ddfdb60f553e7dd856e0be18dbb6d77fd375a8"}
{"id": "train_026032", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d9a6bb982d0d6bc7508a7d7861d5926ef080e97b"}
{"id": "train_026033", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "45a08087d6aab872f3820350d9da0a696f97e2d6"}
{"id": "train_026034", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "cc8b5557247833cf51ca4cd23db688ef32d66370"}
{"id": "train_026035", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "f213d2bd5cfa8cfc3cc3f68dccab7ef6079dd034"}
{"id": "train_026036", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2fafee6602be8b010930ddb1fedbce125273f278"}
{"id": "train_026037", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "939101ecbdc0e4d18c819460bfe014e4a3965db1"}
{"id": "train_026038", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9435b31d62a85d61db88c7ffbd273e24ccd68f41"}
{"id": "train_026039", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "72b192b937b2e14cc36b84e8a2863ddeb0a4c5d5"}
{"id": "train_026040", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "5d6e26488ff816bf70a9196586fba0a7d17a2d14"}
{"id": "train_026041", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "eae4f490e8252baaba37a26a6c6e6bf8642fbc9d"}
{"id": "train_026042", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6118f74e8032ced0d5f10d1f68dcdbcfad5d6bcf"}
{"id": "train_026043", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "812b4ffeedad3d3207a75c2ab59fd8e683f9c576"}
{"id": "train_026044", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "869fb5b036a8c1ca9f9a8e02d453d1c635ce244d"}
{"id": "train_026045", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "030127799917a8eeb65509c97c01b9fd283acecd"}
{"id": "train_026046", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "b3d46694475e9485d77428e05e75b2b56716e5c1"}
{"id": "train_026047", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "20a4293830374243de6cefc0d63444c6d002b215"}
{"id": "train_026048", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "0500a8687c75c74d99b88db7fff6168409162504"}
{"id": "train_026049", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "228b3e725a0bb2363be9ecc51a5a5469081ef164"}
{"id": "train_026050", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "f6807152e62e6e511a9152b928a1f2cbb3dec0f6"}
{"id": "train_026051", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "ce4676e413b6e7226ae913b281fcb51b6fc05315"}
{"id": "train_026052", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "2809d5345ec2ec27cbbc0ce697f0f1bd7f8b98e4"}
{"id": "train_026053", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "06aef1b00be3b77fed4a4cd2169c7be7e6218bfe"}
{"id": "train_026054", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "c0b5d13086d4a7faa06ce629efe7396d1c5b9610"}
{"id": "train_026055", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "46a6d0f71216c2e7e22301bee3ae81de9706c8b7"}
{"id": "train_026056", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "fb96f43342c8decbc32884e15daaf7e51723caf9"}
{"id": "train_026057", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "44c1848c2b034444edbe104b68021087730c64ad"}
{"id": "train_026058", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "231344d7d4e3a0c7aed09b436137715b135bcf83"}
{"id": "train_026059", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "b99ba282d9eb6073814735337aaa395d339f90bc"}
{"id": "train_026060", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "fbb831d1c9d9cece6ddb83923c6b7f690d489708"}
{"id": "train_026061", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "7603f26bf012ed85e1fa5bede4d5e31dabb58142"}
{"id": "train_026062", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "c83ea10916edd6952610ef405d2afe1082f5767b"}
{"id": "train_026063", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "b187682e1dfaef7ea5bee3ba0b08d0aa417242c6"}
{"id": "train_026064", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "a91a14e4e9230cf215f240f9f48cfb4eb70574d5"}
{"id": "train_026065", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "8bd70aa5cae48bf35968f9254a36f1aaaeda5add"}
{"id": "train_026066", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["performance", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "157a42e04e3833f6cb23adc3190c8ddc67cb12e1"}
{"id": "train_026067", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "faf3a368ec906d01e74752a748f0ff32dd3a961b"}
{"id": "train_026068", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "547fcff9fc4c4cacb79498987f9d3a77bc570749"}
{"id": "train_026069", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "db8d27bd40f8aff2fa541d298e7a4c727a7e249b"}
{"id": "train_026070", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "75391582fabb2f87794f2f305ff8487d5b0fc2a5"}
{"id": "train_026071", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "2a068cd3d248429eb53fa56b72ddc203353a59fc"}
{"id": "train_026072", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "49ad9149c7b2eb3b16711b4ab536cd4156f28ea5"}
{"id": "train_026073", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "da043f4b1cd88f8220c41204355ec5123164b739"}
{"id": "train_026074", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "5ddf61c2795cb0b8d548a70079f25304758b64ba"}
{"id": "train_026075", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "4bfc0f02fb72feefbab9a87d5079257e99822d47"}
{"id": "train_026076", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "3f12152ad5bd1f4146b4c1c83bb4b51361fb153c"}
{"id": "train_026077", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "e355ff449fe017b3a0e28112d13bc0f7e46432c2"}
{"id": "train_026078", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "b43435b1ce1e7f9c0a79dc13c08485ca896d01ad"}
{"id": "train_026079", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "4c76c2c55d95195415a88645541b25c23fc283de"}
{"id": "train_026080", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["research", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "f7e2563f9e2c0210f3b9320e000b1a0b4f21b0d1"}
{"id": "train_026081", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "d3828ac3efc5fb12625de3fa2a78b24751c9509d"}
{"id": "train_026082", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "289a7568d408fece2b28cee89fc7407a2155eda8"}
{"id": "train_026083", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "76ec744f4262c33e4242b3f6c7600cabeb76480a"}
{"id": "train_026084", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "14843338f67a34522a83c0e86c7fcfb700b94d74"}
{"id": "train_026085", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b13ae8cb149a5124d6f71676969e8b73c068ec99"}
{"id": "train_026086", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "063468880cd23cf11bbe41cfdb3d836311d13be3"}
{"id": "train_026087", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "a0b9daa2a00ae52edd8997b1e49928ed2a47fda6"}
{"id": "train_026088", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e798bb5cab8fc8982cd4e49d5ec8bf0cd51b4e56"}
{"id": "train_026089", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ac97864ac30e95f1e6f8e4b97c4fe679c3ac3076"}
{"id": "train_026090", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "13b0951c3a171fbe1f50dc10063e9c25a5cbab66"}
{"id": "train_026091", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "fef23f2b66e70ce7b459a6cb36c8ea5b9ef9d6dd"}
{"id": "train_026092", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "a51bc4fce97d66b17cd9441187745b420c42d0ec"}
{"id": "train_026093", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "eb6c7d714ca3f40bcd274e0430c7871b6ea2f23d"}
{"id": "train_026094", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "175a7e3f0337c3aadafca5f26c0bc93e814b9226"}
{"id": "train_026095", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "22ebe33bfc1989a7bfd452bc24395fc2c4b1d34c"}
{"id": "train_026096", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "01eeb83a8f1c0665e570985fdd79037cab3e8378"}
{"id": "train_026097", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "1af55fd81195c1b1cfbe7a878b142008f8d23f56"}
{"id": "train_026098", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["coding", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "f20e96f7de1349243db4f005e7e2c9711292b023"}
{"id": "train_026099", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "79299b843b91f41f740af0d3e410648c91050a31"}
{"id": "train_026100", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["coding", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "d1d13a07ec3cb1dccc57be1026ffd44a98fff5f1"}
{"id": "train_026101", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b94e19fe9274740df526fbc6e5601bbef8a4192b"}
{"id": "train_026102", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "5743cf4739f98320955030d35a7bea245722b209"}
{"id": "train_026103", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "498602c41a7beebfc7e6716df8eca6fb58772698"}
{"id": "train_026104", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "be92119acecfa3ece4f6750e0c64ba5c80c35da3"}
{"id": "train_026105", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "1d14477e066aaf2ab4539f32c5476a9a5df3f70d"}
{"id": "train_026106", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["evaluation", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "137725ab89ca99d324236d73ad67f21674fa002c"}
{"id": "train_026107", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "491df64363bbf7f7986bcf547df9bb73c8af62ea"}
{"id": "train_026108", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "94ffff0149b130eeaf320a107fd848b17060bcc1"}
{"id": "train_026109", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["research", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "c626434ab965d054f6f9cff7827ef026de840252"}
{"id": "train_026110", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "1a6dcbd505b180c3084b3f003677e8a660f6aaa7"}
{"id": "train_026111", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["data", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "d6ef3a50c67895e0e191b72bd840f2255c8a9cbb"}
{"id": "train_026112", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "7568f2217e87e3cacbb1a70e8b6409838641e3f7"}
{"id": "train_026113", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "9165749821f61069b109518a4e2a5368cc67a612"}
{"id": "train_026114", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "50e7eab806549a338c41a475698fcc90fe8e1038"}
{"id": "train_026115", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["evaluation", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "40eaaa8124dce1b16e1eb7a1674e373c7445dee5"}
{"id": "train_026116", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "b2ee02064ee55d478f2c6508b31f7cca3e0490ba"}
{"id": "train_026117", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "016ee965bf128ab58386bd98e14d6761f58de5e7"}
{"id": "train_026118", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "7f791ac4684c1092d0fc6365c79d0f4d97ec6de8"}
{"id": "train_026119", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "0be5b0dc7cc1a6cd431a1269f219e15fdee35dbe"}
{"id": "train_026120", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "22dd52b7f156adb3abc3a742809e6a5c187c3e37"}
{"id": "train_026121", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "6f3a42dc052612965f2503d5eb1ce1fbd9afd5fc"}
{"id": "train_026122", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "d296e7e77079873d186204de0af718ce32131b70"}
{"id": "train_026123", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "93691cbf3cf750f24629f60fffec29fbff5d50e0"}
{"id": "train_026124", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "041c6c4ff8d37a8fabd03be8439828baaf2f5950"}
{"id": "train_026125", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "e44b50090d7a38d2d6b013c78646539f23623b83"}
{"id": "train_026126", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e84101f2c99839ee5b4ba05e1dfd426954e2e009"}
{"id": "train_026127", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "5ee43c6c3d8e11a87163183772b67792ba8df8a4"}
{"id": "train_026128", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "7aaf4e8f3ba23d8c46d0bb23414fe7f9711a295f"}
{"id": "train_026129", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "0eb64a143e854459490e1cdfb6c17f3c1744c2f9"}
{"id": "train_026130", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "147421f8723ce46fe230158a9501fe8bcf960db4"}
{"id": "train_026131", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "debd69453fbbdf52670df3a228aec55c07ea7753"}
{"id": "train_026132", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "64edf701665f3e331212ca85024a8f17e822ff3f"}
{"id": "train_026133", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f77f376691405f242ba11c751200822beb76f15d"}
{"id": "train_026134", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "4686de4778771f1d7cfd229a1bc6727905a982fd"}
{"id": "train_026135", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "8466bcb1862021d0359bd6b01b2bb12ff0bbc886"}
{"id": "train_026136", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "5518115bc101544c765e45444ab8c5916ff65b20"}
{"id": "train_026137", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a099657211ff03839238f78ba0d8928f40829010"}
{"id": "train_026138", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "0ddc60fc61864d3962bcdec4e8c88e263fc8adc1"}
{"id": "train_026139", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "d954c4825a77eed1d844687767de9deb37c9d798"}
{"id": "train_026140", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "60420fbcd8b9d3982401faaf6dd5af0ddd8ca6bd"}
{"id": "train_026141", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "cadae7023a5315166b7e18d744542c58152e544a"}
{"id": "train_026142", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "63bb526bed5c6bd91bc9b20b0d5d5b0175849f63"}
{"id": "train_026143", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "17666e78699bc3744be0b12916fab126c0fd06e9"}
{"id": "train_026144", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a0770877c2334a374e08c7830627700460e0d25d"}
{"id": "train_026145", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d45b185cf407f185d54c53a957f98f937452150f"}
{"id": "train_026146", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "5c6b21a7177794f3c1b9c23d99dfbc15f9725d96"}
{"id": "train_026147", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "82313282548f199e875aafa581297d1d4ed3cf51"}
{"id": "train_026148", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "b5aed132fc7a9dab81cdc3bd1b7d0ab1ecb71905"}
{"id": "train_026149", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "abb2b5742755f0dd5261ac8ff1bd9ec4b032c0e3"}
{"id": "train_026150", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "f1b66576aca3f64714b46a4049c1b485e7f61087"}
{"id": "train_026151", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "9291d0677ea8111dc7ccbf0456e495f48f52171b"}
{"id": "train_026152", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "8b6597c79140a6471e75fbbe27a507201ad2ba25"}
{"id": "train_026153", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "0899d484a49a1d52fcd139ff2050981c0d475c62"}
{"id": "train_026154", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["security", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "35f71153ed7c3af5f274830312589421b102abeb"}
{"id": "train_026155", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "13bc956f09da7808699ac088471cba9ed877c12b"}
{"id": "train_026156", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "654f4a7bc536b3402a8351a9b3cfd7a8ad042f90"}
{"id": "train_026157", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a6b30229a4bb377a5c8a4dc02ea6b2cde0cfcbfa"}
{"id": "train_026158", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "9e11bbfd7f6173ef4893b23a43009fcbd3fe7ed2"}
{"id": "train_026159", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "14470e25565b0c22822ae35c7101b39ef2ef0ad1"}
{"id": "train_026160", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "55af69e75ad79b56a4281ed24be21dbd3bfa4956"}
{"id": "train_026161", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "27b6a9bd88508a4c06841d29d7ea69aaeded467e"}
{"id": "train_026162", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9e254152214d1b7f2bab0b82e9c1207d0d9f9a23"}
{"id": "train_026163", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "9b1cdcdd4a3a1379755c8524910e59b00915c086"}
{"id": "train_026164", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "abc5e3b413f45018d8e29dc476c2112727b2382a"}
{"id": "train_026165", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f2ce35a24e1774c0de7863bc09fec714c7f2e51c"}
{"id": "train_026166", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "c79170fa06cedd25a201bca12d14e3ed85389bab"}
{"id": "train_026167", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6ac49a89a411ba72bf9185bb6c055de9ef228354"}
{"id": "train_026168", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "db6c51d366085c7344c26a9ebfd12d9b6e392df6"}
{"id": "train_026169", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "913ce9edaa51ae379d37d7ba710b412e682e5fc6"}
{"id": "train_026170", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "39aa4351685d1611e1bda3739aea99ea63898d01"}
{"id": "train_026171", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "0f887a4eaf20ee703ff88e6192bebc08644a94b1"}
{"id": "train_026172", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a9fbfd5be32368b7343444e840cb7ac2240f658d"}
{"id": "train_026173", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "9f04cf1b92638645dd958abe236c342cb4a76b1b"}
{"id": "train_026174", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f34e7fa68a57e7eb286b35519d7625b43212ef81"}
{"id": "train_026175", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "dfab1422c1e0dd36b75c344dcd4190685fc8427e"}
{"id": "train_026176", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "16fe2128233378535a2c520fb2fcfaf904af4aab"}
{"id": "train_026177", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "55162cdecfb1a624fb80ebf3b1f76cff1c053cef"}
{"id": "train_026178", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "d0fac72958b9437ad3c87f8cf75a4e66044c4eea"}
{"id": "train_026179", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2e03368b4e497eac3e7713c6d2d9a0d17e28c0f6"}
{"id": "train_026180", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "0d7d9386b2175b90174ecfd435b0f5492d0cea1e"}
{"id": "train_026181", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "c6524afa1c68875d5ee91d143bb2cc119d636694"}
{"id": "train_026182", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "35d8aac1cdb2981c6728a99cec0ad9f5556848df"}
{"id": "train_026183", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "a1a028e9eeccac8893bc7cfe0a4bade88edaff83"}
{"id": "train_026184", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "31284d41f55af50b458e47a816ae779e33ddf06e"}
{"id": "train_026185", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "2d8d97c57eb21b71a06a9b35087dcc9e6f6db6a6"}
{"id": "train_026186", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "f319cf15d361342b6d41a17a6a6feb2d57525436"}
{"id": "train_026187", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e3798ffff1addacbf4d969261f8d1393b3a73cb0"}
{"id": "train_026188", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "85011aa0fec75823fab78aac1f37a8577de9b354"}
{"id": "train_026189", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "f8df67c70609fc135d1b2f4e3cebcc459cb5e1ed"}
{"id": "train_026190", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "9398efad2508dec954078fc940287838c59eadc3"}
{"id": "train_026191", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "17566ccb8c94208bc8fbc5512fbd76131d3e7472"}
{"id": "train_026192", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "d1a11bedfea6827f284bdd72767036746e0200f5"}
{"id": "train_026193", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "7b17dfe8535df15bc842d9e7a21a68eb61840317"}
{"id": "train_026194", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "4f52aa9617cfed9af8f72a6bdba0760b305b596e"}
{"id": "train_026195", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "996f29722e1bb35dad9be3fa9a99375d27cbbd70"}
{"id": "train_026196", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "d56ae463123bf559122d8d8ed254bf63e21fed74"}
{"id": "train_026197", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "8622291081792f8d3abf1bd8c779abc478c264f2"}
{"id": "train_026198", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "3ea78ef44d7aeb27ccfd5f0e2c1a4ca6b5623c4c"}
{"id": "train_026199", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "9aa8e92de14b890db38dd8d940f39773b3044a94"}
{"id": "train_026200", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ffe2137e3fa0a78adc1df8423379946d045328a4"}
{"id": "train_026201", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "5a4af986fdd3fc689eca31554f2087c7f4356e73"}
{"id": "train_026202", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "8ccf1df8534f32b30a51a62d8dc92b6547f52e9d"}
{"id": "train_026203", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "268772bd07c136944550319eaf16d90f8b429bae"}
{"id": "train_026204", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "b5ad848b3a467552dc6801d7ba840a5dd3ff90ae"}
{"id": "train_026205", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "9ec7f5a9874e7dc853dc4a3bb0c6daf39fa1ebcd"}
{"id": "train_026206", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "62cf44faa50d5daa79fce65485039fde428d45dd"}
{"id": "train_026207", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "1405ac8baf6dcfb87b75620d0054a48febf33131"}
{"id": "train_026208", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ac018b34fdce58c50671e2014e9fbeb0cbb4c364"}
{"id": "train_026209", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "b3c91a5ee79dbb6706140a748aefea1fe8fc300d"}
{"id": "train_026210", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "20ec8279c2497dcbb3bb01591d41d754dbba41c4"}
{"id": "train_026211", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["security", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "3298760228bbe4f94eabb0a212b34992166bfafa"}
{"id": "train_026212", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "424565d7e8db39562040afb125af8f2731eff37e"}
{"id": "train_026213", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "fd39399cbd9b05cbca45d2c35fab02aabcc00064"}
{"id": "train_026214", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "8d3034c3b0ebcaa6b5f3e5a124b4057d26668813"}
{"id": "train_026215", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "960482ff383554339b76c91e571996e0a97e3ef6"}
{"id": "train_026216", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "61efd778fff7ada7ec406a565efbe1ac6d5e4a9a"}
{"id": "train_026217", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "3909eda411a808a15b316b322daa6c6709feae4f"}
{"id": "train_026218", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["performance", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "68e6866fffc444012eefa7487a86c3a3c282e3a9"}
{"id": "train_026219", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2865ea40b081e2cd1491e45ea076c8d3fff59594"}
{"id": "train_026220", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b04f0e2ce54ddeae387648b144d4491a4b3e1f9d"}
{"id": "train_026221", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "617aeb955700a2fcad99d00a6033e21d02d5030c"}
{"id": "train_026222", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "74dfab6e55e80c24fa784de70a256617452535ca"}
{"id": "train_026223", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "e355ff449fe017b3a0e28112d13bc0f7e46432c2"}
{"id": "train_026224", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "bf54c5a88bda7485bf25359740107ce11d308478"}
{"id": "train_026225", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["agent", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "6707da8b1615dc499398a7fc9bc2b139d238acba"}
{"id": "train_026226", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c980e436cf5902bc6b957019cb9cad3575808098"}
{"id": "train_026227", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "ceac9490b66c1623e33cce497d89f9e5661e3ce7"}
{"id": "train_026228", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "5a3262f9f2ee4132a2ec312ecb25e3cb9e3535b0"}
{"id": "train_026229", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "388614f4961b121c93fa17eacf13775ddef98351"}
{"id": "train_026230", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "fe38c1639a08e32893002f34f6f913f3f0327b32"}
{"id": "train_026231", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "fa60158e5fd2dd17741b311d2f08db858bf43fed"}
{"id": "train_026232", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "d9abc08499c288a937c32918ec03f35bbb6ab84a"}
{"id": "train_026233", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "11ceab9fef1b2e416e164b3ee700b9569d425ec4"}
{"id": "train_026234", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "fe40a5ccc84c09bf38e0f85edc5d4ba5ac20cc7d"}
{"id": "train_026235", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "df5d5fe45b52edca6bb7f709b4d4a6f03d27e76d"}
{"id": "train_026236", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c5e99986c97471316b1ef3b735e2ebf7eff4021f"}
{"id": "train_026237", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "92ea11129f0bbe21db2f788aee99dd4e62151788"}
{"id": "train_026238", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "e0d9f5d742b75cc4e852c67f8a890cbd714a5a56"}
{"id": "train_026239", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "6a6fab19ac55154ccacba55832fa1d919c01faf6"}
{"id": "train_026240", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "0ed89a760652a938a4d470e89c71a0177dae13f5"}
{"id": "train_026241", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["coding", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "636f0a07f55b30ee8082c7612aa6857761fbaa04"}
{"id": "train_026242", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "949166241e53b7ed6f91f03f717c3c4b3e552906"}
{"id": "train_026243", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "66cdc52a9d3f21fc39227b1f5b5074719cc86875"}
{"id": "train_026244", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "26a43579ba358cc2760ce64271ebdc4716479cc4"}
{"id": "train_026245", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["data", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "b1443d496473e621b0be8bf49d7e8403a9a25041"}
{"id": "train_026246", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "bd1bdc736389d468077cda22988808b0894dabed"}
{"id": "train_026247", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "93019e1f95751743ee659aede41d825e237befb2"}
{"id": "train_026248", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "27aaad0cd7a8f0819b3e5842a388798c5fff5c39"}
{"id": "train_026249", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "b0e109b49a8229f1e507fb37c2072cbeebd7d2ca"}
{"id": "train_026250", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c2356e1ce058812530a620fac8cc1f3989e0d42c"}
{"id": "train_026251", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ac781bdc139f201ebdb53da417154ffb1dea4201"}
{"id": "train_026252", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f5142426e5c5e06ed9b4c2af4c6c344470f21042"}
{"id": "train_026253", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "66f2e95e61b41ef8d683cc9604fb0633d4f03a97"}
{"id": "train_026254", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "311ec4643a3814992dd076900d8e28e85df57abd"}
{"id": "train_026255", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "c9c91261cb067ed30a4107a7faff75b229e45a44"}
{"id": "train_026256", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "718a384b40825883b6e5e862de3ef6cdb9f1004e"}
{"id": "train_026257", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "f2632477376871daf22d0423d40eaae903ab5ae8"}
{"id": "train_026258", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "4ed210ed71766a866581f518c872c64a2e863f39"}
{"id": "train_026259", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "15d8b612644badfc9f7f6e22d1cf0494b42345ee"}
{"id": "train_026260", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "a7fe2e860831af7f65f4cda105ea8c2163f7ee74"}
{"id": "train_026261", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "322b7e4f8b8d86e9a15368c39bb260243a6ba0db"}
{"id": "train_026262", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "9c959812936dd0b622aa47821a2fb86559453071"}
{"id": "train_026263", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "20c22698ce278328b6fefe6bb06ad10c195ed2cc"}
{"id": "train_026264", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "017d58f5b7d5650f655aed2998c2d66ffceb48ca"}
{"id": "train_026265", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "31394e369d6351d18d32c563279f8721a77a9e78"}
{"id": "train_026266", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "74b0d5e397d5464da1fb2244539aa0c9e3f3d6c5"}
{"id": "train_026267", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "5451324b514c92101126595dafac1c7ba50ef8fe"}
{"id": "train_026268", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "fb7cfbd18173f36c57d7b9c6512722a55a57af31"}
{"id": "train_026269", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "8ec7a4fa68a64ba9c65aaaa82a4d399f078c572d"}
{"id": "train_026270", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "003fded837a95248f4ac57c64ec4a579ef47ba4b"}
{"id": "train_026271", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "07082ecfe2f4d8799d292e44e47503155d3f28fc"}
{"id": "train_026272", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "dfa184f2b2c9f518c0d9a1e7bfddea1481756bdf"}
{"id": "train_026273", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a10f823d023cb39fd5d48cd770d54d72ebaf8309"}
{"id": "train_026274", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "e4b5f4ba831f22caa16135baa79354ed14693276"}
{"id": "train_026275", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a6a2aba39c59fb2b1fa4e54ad5d027845135edbf"}
{"id": "train_026276", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "852c0c1adc0bda9811a56f045ed7d40dcc0171be"}
{"id": "train_026277", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "50d0e2a0a4fca380203099e82c04651d203a66ca"}
{"id": "train_026278", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e079fb5ec96bb8d0ee46857150344cbe693b96a6"}
{"id": "train_026279", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "23937516580575c994ef58f7b06d240b0f3d54ba"}
{"id": "train_026280", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "9a139b69781492d5c578454416726234ff4ec358"}
{"id": "train_026281", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a6abfb8f4225a3ac57f32bbda1a802280588ad90"}
{"id": "train_026282", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "1b27101f1d0b50e7faa0c400d966ec01ea3de149"}
{"id": "train_026283", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "3bb3a2316527668e0984c0f9e4aa6dca750040aa"}
{"id": "train_026284", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "cb2136ccdce9db0aec891e05aca32004627c3c94"}
{"id": "train_026285", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "7421cf5b0ea1f60edaffaf168c5ae80715be7c1b"}
{"id": "train_026286", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "14e8ee24e95d8ac9d9f320d4a4e12740da3f86f1"}
{"id": "train_026287", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9c7ee9a176c4f0c477928e6592500935cf0795dd"}
{"id": "train_026288", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "43b3425bb3762346e13d1d6307551be0f8799529"}
{"id": "train_026289", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "246f3378872c930b7f08b77db47ad33b837b6857"}
{"id": "train_026290", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "384cb83d6350a86fde6a5eff93d85dff8387f76d"}
{"id": "train_026291", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "3859c616366311506501dafb14c198c26ed0f4d0"}
{"id": "train_026292", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "08fd6adb2530ece689e2e3b501aec81239ed10c8"}
{"id": "train_026293", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "b35cd65f325742a3744a181c2a22b594e714477c"}
{"id": "train_026294", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "956f9ccdd45312c30f2bacbee54498667a9bb939"}
{"id": "train_026295", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "99b8a43ea48b39e0c648e3a65bbe67832d3f5054"}
{"id": "train_026296", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a928122813cfc5d7961e013083954adbfafbe788"}
{"id": "train_026297", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a2d0565068fa1cc53d755abea691ce478c176664"}
{"id": "train_026298", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "8271eac9c4961ec5faf991bebe773e838e95f012"}
{"id": "train_026299", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "144aef08461a7497efd5e3ca42cf5969b59e921b"}
{"id": "train_026300", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2d3c0f3a12eea90b0d3317010a9cd7b9c1eaca76"}
{"id": "train_026301", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c307f8cd58211a9cbf16888790c436dad45c9692"}
{"id": "train_026302", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "034cdbb3095c421dff1eaf9dd706bdbd83f147e0"}
{"id": "train_026303", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "3ae1403d3751381ffb898850f9165189357e064d"}
{"id": "train_026304", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "f5087718a0555aba840fe9da344a86634c4fa2fb"}
{"id": "train_026305", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["performance", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "d419a48049665c062e2b219b1bf18199aba9e900"}
{"id": "train_026306", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "1c20916063afd738e4a11e7e970dfc41e47a40ab"}
{"id": "train_026307", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "c55b33bc972f2c7e08aefc5c1e7512e655441759"}
{"id": "train_026308", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "cc29d191909b1333fae5e465fc51b00487dbf7aa"}
{"id": "train_026309", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "e8190521c7d8a7ff6c4b2303d82d6fcc8715d37f"}
{"id": "train_026310", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "28722df3ce46b86c17fa65f020abaa5c1a4797e3"}
{"id": "train_026311", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0e00a7b47b0cd4cf6f5b16b99251ea0009d34856"}
{"id": "train_026312", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "21e7aba4505317acd94c2910d36e1aa15419907d"}
{"id": "train_026313", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "b7ef183b50381f5085d45d921762e8a695b4f1f3"}
{"id": "train_026314", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "3d4af73174a066046fa2f3f67a856571c7a3eec7"}
{"id": "train_026315", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "e1814be69088ea1fed092e28b16c8b8a7b523924"}
{"id": "train_026316", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "bd1fd3cdbc932d0863bb4c120fa47e315c371237"}
{"id": "train_026317", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c7455b8e188acb0cc050c446eb8429aeebcf39ac"}
{"id": "train_026318", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "049bc3ad2fe5bed8f10acc973603dda6f57b80f6"}
{"id": "train_026319", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "52fccb3048d2118dccfefe42e3b74afa9e5fa749"}
{"id": "train_026320", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "c73d3f6690c87001afb5b0c01cca4e2f6bd07b10"}
{"id": "train_026321", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "33824aa12a42e85084efa85df31653d7013a8fe1"}
{"id": "train_026322", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "91bdca75b617653dc21ecb741c4971008c993ecd"}
{"id": "train_026323", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "a617c329e2695030284c6bc23492ed1ba84a5d88"}
{"id": "train_026324", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "a7d1933a796cb1cc36f3228181ee9fa291d4d157"}
{"id": "train_026325", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "f30cdbf1278246c6a063a43937f8328ef28abf0c"}
{"id": "train_026326", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "4873664f0584aaf12ff94a35be7c68f714cde37c"}
{"id": "train_026327", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "65ffbf2208afe9e075075d7b33fc7bb504d85644"}
{"id": "train_026328", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "433f61ccd1f90b9a01257c706d04ae99d7ac3a09"}
{"id": "train_026329", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "5c6be5751c96ee0af8df7d5ab1a902c957f975ae"}
{"id": "train_026330", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "fd7443b5cf00ccfbe7cf2189fd7abe1a1c01c3ce"}
{"id": "train_026331", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "3332cb2f6d470424f2dda572b42edb68b7f2f983"}
{"id": "train_026332", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ecd3dff5b82684eee6f1855b80dd3697074bd2a8"}
{"id": "train_026333", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a882b703f279444012322c978df12bf692dc6a92"}
{"id": "train_026334", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "40eaaa8124dce1b16e1eb7a1674e373c7445dee5"}
{"id": "train_026335", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "1e82f406f46b5adc61ca7cd152b6173b4ee9f31e"}
{"id": "train_026336", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "87fafe2f333eb050ebf1f8c622ee7d108054fe67"}
{"id": "train_026337", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "fa7224ec1b6d22ab2d0beca29047040ff3f0597e"}
{"id": "train_026338", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "13eb48940bf070b5874dc50e88714e520385590f"}
{"id": "train_026339", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "8d2329cbab7a2df71e19b6f69fb898a0dfd8fd80"}
{"id": "train_026340", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "83ce991d5dd8baf597741a5d1ed22089433bb271"}
{"id": "train_026341", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "860d90e4e27b1eafc914266c02f1718836a0eba7"}
{"id": "train_026342", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "fd5ffdd9bddb0d6253803cf5a170059b96a75cab"}
{"id": "train_026343", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "8fc7894bd124b5e5dc36563a1f97f33904dc4896"}
{"id": "train_026344", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["security", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "4af3701aa3a208bd5e7349876ec93ae203d32c6f"}
{"id": "train_026345", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "d45bb2d5ba4dd3eec842589cdfe0cf7fa9dc7747"}
{"id": "train_026346", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "d224da9c218ee788e8f57181c122e65aadf0b388"}
{"id": "train_026347", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ed4a63031a93f94a3707afac7bf4d3a4503db683"}
{"id": "train_026348", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "704c9d0f524080900706c70441861c65a7fbf2a6"}
{"id": "train_026349", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "1ed57bb3c68061b2db03d9a29a6a8e30558f60e3"}
{"id": "train_026350", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "7ae569fda34761fadca28b35b1378140e819d81f"}
{"id": "train_026351", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "df549922e062460e57d49e5e316be49a24609f15"}
{"id": "train_026352", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "00ca8b239c567cba5a4a84e9d43e894450b5ecf5"}
{"id": "train_026353", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ff5250bacd3fc2ae826be931e46a098b2971f9e8"}
{"id": "train_026354", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2fcfdec5defb224ef4803b640577fb2e3e24c610"}
{"id": "train_026355", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "35f71153ed7c3af5f274830312589421b102abeb"}
{"id": "train_026356", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "cc8259d1daacb2061225720d59405ebd70e5d3f9"}
{"id": "train_026357", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "78960b90006de8a9c624e87769cd1eb42eee88be"}
{"id": "train_026358", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "8b705a0d5e913bf9bea8dd670af4ae9567f7a1a6"}
{"id": "train_026359", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "672ce8e02145200afdfb8b396dc4eeb87b6b10dc"}
{"id": "train_026360", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a88e02120226abcbbd7f466b3f588420fbd6e3e4"}
{"id": "train_026361", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b2caa2662a60d3cc55580b789644340519743a4d"}
{"id": "train_026362", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "dafb9860f53df9fb25daa37b4fd4c47594708559"}
{"id": "train_026363", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ebbcc6a5102de01d8256f16b19352a4f2736d9f8"}
{"id": "train_026364", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "9edc4562c69fbb1b5e94a0b9941a282c4bb7f110"}
{"id": "train_026365", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2a82776090fc58a28d5e89b775696f3b102fd4da"}
{"id": "train_026366", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f5d0fd588dcfd2bcab65e4b8c474e419d21d0242"}
{"id": "train_026367", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["security", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "d8c6568cb75791c100fa3664c7f5050a444ec298"}
{"id": "train_026368", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "2965ca68c3f68dcf625470282224a588b4ce6a4c"}
{"id": "train_026369", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ee47fb252944a86e6fac0ffcb441b4ae2027cffc"}
{"id": "train_026370", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "b0ff43a76af0ae35d583434edf65931423bcce4f"}
{"id": "train_026371", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "182870542e3ffe2cadba5c75e0c7895de77389b8"}
{"id": "train_026372", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "01a8d061a65d3a5e63ea994ad16c07cb3a7279c8"}
{"id": "train_026373", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "f3ff892ff4288854f9e4c63406ca90a1aece356a"}
{"id": "train_026374", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "f0cf8fb495499c280b95fe37546ed37f0121edb8"}
{"id": "train_026375", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "b2f0aa84d7591c1e07e6359361e80209b4820e99"}
{"id": "train_026376", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "016ee965bf128ab58386bd98e14d6761f58de5e7"}
{"id": "train_026377", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "14228746434456c3e54883cad34d17e834c54afc"}
{"id": "train_026378", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c3ddd0276fbddfb7f277871afbfe6bd65eecd2c4"}
{"id": "train_026379", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "8b8cd1988faad266d98544f041a01942b48256bb"}
{"id": "train_026380", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "c040765a90fa7a1c34e3f431e1b70d4515d2859c"}
{"id": "train_026381", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c5adab3803cde572bf8af38ec8f3b45d995e64ff"}
{"id": "train_026382", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "e41073f8aed79cb45120dfa600c2876ba405c316"}
{"id": "train_026383", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "cb871a6c1cd10d3b712ea89865f731e2c8710289"}
{"id": "train_026384", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b6148328ab4fcf577c80cbd8eb1aea9b4e89ee9d"}
{"id": "train_026385", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "534688190abb99d54e6e3fb13d21be4ae7c47424"}
{"id": "train_026386", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "3970192374e22a757cf6cac98e3f2ad181950905"}
{"id": "train_026387", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "d738b1c35a8cc1effc0d6f7bb98f68649a201f6e"}
{"id": "train_026388", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "aeb1cc71027a6ba4bafdac8a8dd612a681c5a7ca"}
{"id": "train_026389", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "65a5e77374b984001aa1ed2abbb976e3915bcde4"}
{"id": "train_026390", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "0904d0dd93e05c6e9ea3b330c092892efff45c95"}
{"id": "train_026391", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["performance", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "213b2114d5dda24bf593136e2a93678c805ec8a2"}
{"id": "train_026392", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "642785893275558cc404eaca4e3d8d7d1c22875d"}
{"id": "train_026393", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "499ba53451836b9b223282826ef793f794674167"}
{"id": "train_026394", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "54776fd26b1c36a4e9f9142abfdbb8f224453a7f"}
{"id": "train_026395", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "7a864550d4d4b2ccb9bdb7510a7cc409ffeef473"}
{"id": "train_026396", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "4effa13ee32873c08999e3a0da18785e4600155a"}
{"id": "train_026397", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "6e833dd627bea24432d3a3424a7003456a4dafb2"}
{"id": "train_026398", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "cad50e911630bfbb66cb7ddbbe5ba4cb7d05589a"}
{"id": "train_026399", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b0057c44454895dc5af0a6b785b6f06d46283411"}
{"id": "train_026400", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "093f0a478260c6565167948f26ae503912764c7c"}
{"id": "train_026401", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "2cfd709b5f0b7878635daedd111e50fa63a2bc2a"}
{"id": "train_026402", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "a9b83a21a5cdd4a66d839a95aa00f28d9be21d6e"}
{"id": "train_026403", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "3f23d879a492914a5a3ec74e3b262b8be17fe652"}
{"id": "train_026404", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "d370904f0781fc7671152cebd8cfe2bfa5c2a058"}
{"id": "train_026405", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "3c2affab7449849157261abfa1ebcfd2e80f4ac7"}
{"id": "train_026406", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "b93433af064315b64b1107126888e16a5d186b71"}
{"id": "train_026407", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "df318718884de167237443723bdb41235f580925"}
{"id": "train_026408", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "4e86e9aa89212b1e2bc812b475fbf73ac8b90a4a"}
{"id": "train_026409", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "7bf52d94371285a85fe6382294975b3d15265e27"}
{"id": "train_026410", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "8f659b2af442a3881e99c3449e2e84ccebd62540"}
{"id": "train_026411", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "de4fdb96200699cc97e92b969075d24a3dd8cd4e"}
{"id": "train_026412", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "3a12fbd4449182acbfad259166c48ccfbdd19942"}
{"id": "train_026413", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["performance", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "ef8687768d9c687f765d5fc3458c430b4856b701"}
{"id": "train_026414", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "2ed845100c01ddc480b90dd93644c999e565af58"}
{"id": "train_026415", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "afdc7638971e624aa309c3e640fdc64c9cbf8574"}
{"id": "train_026416", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9c567b5e7942d76fd0960765a53f9a2b2f749051"}
{"id": "train_026417", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "7690fd2710fc42dbdcc8dfb00780d028ca0edb98"}
{"id": "train_026418", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "4217c6fa93b954cdf8d0a69227c54b04db1e2524"}
{"id": "train_026419", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "093f0a478260c6565167948f26ae503912764c7c"}
{"id": "train_026420", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "b0207863162fce570dada3ffe71e66984da9937d"}
{"id": "train_026421", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["security", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "624a45b07227ba6753988ddcc50d18204c883492"}
{"id": "train_026422", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "9edc4562c69fbb1b5e94a0b9941a282c4bb7f110"}
{"id": "train_026423", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "879b2549c8fcb3ddb66758150cd5f209dd69a3b0"}
{"id": "train_026424", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "f0cf8fb495499c280b95fe37546ed37f0121edb8"}
{"id": "train_026425", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "860c7a53c5be46e2a853c550e6ffcbc986dd8b05"}
{"id": "train_026426", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "14be10890e09b5e81cf79c99447878a1a0b9631b"}
{"id": "train_026427", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "bea77e664ba73df4b1e03c3b0b420d082ab00ca0"}
{"id": "train_026428", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "085bd2728c90dcb30f606143c1938ab1436392b5"}
{"id": "train_026429", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "38ee0bfc97452c605701ab0a69b6c4b3578c080c"}
{"id": "train_026430", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "e7120e84ee797756972c1057a33b67940b2cc9f3"}
{"id": "train_026431", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "cd00b9a27e67455f45f5ba813678807cd1f7b214"}
{"id": "train_026432", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "f6dd3de8a48858d78e290946950cffd1db2ec1a0"}
{"id": "train_026433", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "8a31780c7cd8ed20421f8e918e0ebb9630338a9f"}
{"id": "train_026434", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "1e12d2041d8805e9476dc6d1f7ec050c486b72ef"}
{"id": "train_026435", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "4386d72f48fe50b72d10e5af82ec663f54bb21cc"}
{"id": "train_026436", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "4360d5f5020345bd9bfeb8442c4623d4c849ccca"}
{"id": "train_026437", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "32f1f29a7a96f7eb93a143d215ec82d5ebcff2df"}
{"id": "train_026438", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "10a5262879c1957864999aa9776cd8253b520366"}
{"id": "train_026439", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "63b84c08317b79900b1edda4de2fc92122b7af74"}
{"id": "train_026440", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "27b6a9bd88508a4c06841d29d7ea69aaeded467e"}
{"id": "train_026441", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "526c3b9f417fb7efb1e66f1dbc6bfd80487bc179"}
{"id": "train_026442", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "4e67f1816aba76564c540a691543a6667f12734f"}
{"id": "train_026443", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a5fb58669cc0bf696456702a6761bc1a94d51682"}
{"id": "train_026444", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "488d9bb80273342eab6ac9cf5999f25d4926c1a4"}
{"id": "train_026445", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "e4bdf19152541a61346c1354bc6c157d8d98f579"}
{"id": "train_026446", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6b033b809ad11ea0fd288ccc28ce03d3d255fb19"}
{"id": "train_026447", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "9a67cc5c6ed44687337fe2c006e9f2f7ddfbef14"}
{"id": "train_026448", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "5cc86bb599de4afe9cf6bd5599fb27766bda4be4"}
{"id": "train_026449", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "1535c377d3ac89c44dde2d0bca04768b92775911"}
{"id": "train_026450", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f19b3b3c5d68087f81879a81ec706dd0d7d14f63"}
{"id": "train_026451", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "9275c2f7d693c0a44ec86f9d2ae18a7c79a2492e"}
{"id": "train_026452", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "e3c65b5c527c15eceb4e1921bb6ce8fb4236d397"}
{"id": "train_026453", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "9376a8a6389455b5c1274dcd5a7e4e60ade7c89d"}
{"id": "train_026454", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "d3ae880988a39c1465f861cd753adb03dcf024d5"}
{"id": "train_026455", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "2913d213c5042719f0619a0abf679c9aefc2a4c8"}
{"id": "train_026456", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "1fb1890959360671b3ef781cde7dc97a565599da"}
{"id": "train_026457", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "e8dfd42424d11d9e324683a2e04efc6384811bcb"}
{"id": "train_026458", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "7bdeb420ea6dd790e74a38a8ba903f93e25269b6"}
{"id": "train_026459", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "3e8ff9a29af923801a519f1dda58a0023b126c8a"}
{"id": "train_026460", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "7f9648a8d3510366b1ca6467eb8f0f3dce763b3a"}
{"id": "train_026461", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "9812b90823ab73b1ddbc92246d2b2caf4887c3d7"}
{"id": "train_026462", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "96cb4b6bb043447bcb092359cebcf091feeba148"}
{"id": "train_026463", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "5c515f7cf9211171a1d24df23c981b2b198b2d4b"}
{"id": "train_026464", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "001dea9703a0f4ab981057838a5f871b3a187231"}
{"id": "train_026465", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2932b8cc7201af118989604965a51cc78cee46e8"}
{"id": "train_026466", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2b0a299c5cb1f1b7f35635979aafffc4622aef3c"}
{"id": "train_026467", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "403c46c7b04c29554bb62e8fc14f90c3e95e0748"}
{"id": "train_026468", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "228b3e725a0bb2363be9ecc51a5a5469081ef164"}
{"id": "train_026469", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "f3f029f1d487cfb8842bc7d4dca60d2f40fc6efb"}
{"id": "train_026470", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "8a2eccb0cad8c138de9499cc6f659bcac361c60d"}
{"id": "train_026471", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "e6850e89a2cfdca98a009d2205bf583ceb6140ab"}
{"id": "train_026472", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["agent", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "bd62b1b55ada7b279870ab4568a5542c07799797"}
{"id": "train_026473", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "21470691981c1d3a7cdcbfc117406af7f4db4a73"}
{"id": "train_026474", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "c053b6d782185b9d7a2c9a30525e36d21aff9e55"}
{"id": "train_026475", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a936bfb9c82cdef0f198b9cb940549e6f70762d4"}
{"id": "train_026476", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "cfc4db1b593f48c405d0df59d5cf838c2356d702"}
{"id": "train_026477", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6f2d7ad61bf7df85c51a883b950b80c357bcbccf"}
{"id": "train_026478", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "93ce105a00ae2126dc3a9ce4d8cbf4aeb9e60e03"}
{"id": "train_026479", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a9c36c70a9bfb7112245cca72ac177207f861988"}
{"id": "train_026480", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "ab54060901754270eb4930d2f0d9ffc67130772c"}
{"id": "train_026481", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "2a601b60b48730d05dce86c835faec9975849f07"}
{"id": "train_026482", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["security", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "83c9200aaebdedf9e1c54639a748740741b0fb5e"}
{"id": "train_026483", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "787b2b00e935d908bcca36a4fa3179bc9980f592"}
{"id": "train_026484", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2e7616c1a224e78b788fa392c9611603eff3eae2"}
{"id": "train_026485", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "7d4949dcdfc11b7a16c70a25b260e5d0810b5505"}
{"id": "train_026486", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "f03834f21357e481cd8d290c8d91d4b9aaf2a1f1"}
{"id": "train_026487", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "f5dbb675c062aae94045c852a386ffd55b78c3cc"}
{"id": "train_026488", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "39fc14bb3d94d09d053bef941f346da6bf5f5111"}
{"id": "train_026489", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "78071b3ab6b92be5aa7f9d8a1e7ae2b05bfd4256"}
{"id": "train_026490", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["governance", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "989fdcaf9aa8872bef58c3b36bf4ca10312d1e99"}
{"id": "train_026491", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "aea05c7a1ef3a2a72daaeedf7f0ad65c697bfbf9"}
{"id": "train_026492", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "dc9853db06c3e3882da0f397b233eca1cfd69c40"}
{"id": "train_026493", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2406b4334a1c84ce4f236572d48c63d68aefca05"}
{"id": "train_026494", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "66fae0a587584e14e6dad59a2a0abcbe4096c2cc"}
{"id": "train_026495", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "7fc3722ea6d9ca06722a657384ab8826b91d7a56"}
{"id": "train_026496", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "c82bf716fedd3063e446efee17c15f372f4edff9"}
{"id": "train_026497", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "710479b385359edd4a1140097b03cb115ba77551"}
{"id": "train_026498", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "bec18459a62dc511897633145509f8c33d7dfc8b"}
{"id": "train_026499", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "5fb254dc96de9017db59092a2fbff43001baca95"}
{"id": "train_026500", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a4b85ffd4a7437392f80d3dad234c28a046ec328"}
{"id": "train_026501", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "1ba1cb6a0a0ab341833c2f503716c42ae9e85c35"}
{"id": "train_026502", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "fe56adc6b1bcfe520a10d99e85f9e373bff3362d"}
{"id": "train_026503", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f007f2995b7ceac2f73016122d09f279ab015d51"}
{"id": "train_026504", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a05fcedcc8f4f4cba36dd37d4a1919aa92e34ff4"}
{"id": "train_026505", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "169bda21f9bd1f80480dc0e226ffb5b6791265f8"}
{"id": "train_026506", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "97a6ceb3908ba12ce9eb0d8495d31ab55142eb21"}
{"id": "train_026507", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "306b0345898e9249f7262c0b53d037803672c28e"}
{"id": "train_026508", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ac98339a542c716d97f1670d5b66839446aad92f"}
{"id": "train_026509", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "213e550f2f4d0d9ce0e2039396dcbc893fd59842"}
{"id": "train_026510", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "bf5d4d796973dc1cdc39373f1ec4a69d0fdc896e"}
{"id": "train_026511", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e9371d78a44e8e671c34ba862483b26b3a6e7a1a"}
{"id": "train_026512", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "0ac0bf40cae780cf0ff9cc0647c016f6ded5f3b5"}
{"id": "train_026513", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "0532f29da7862e8cca0fc17eed72f6a028466faf"}
{"id": "train_026514", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["security", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ac6bc346b987c32b6dd81b6fe436795ed93c8a12"}
{"id": "train_026515", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "8adc11fe6f981a095e2856d9af151abc2191e0e5"}
{"id": "train_026516", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "870e811b2102d751f8742ac542da0d7aabf4a66d"}
{"id": "train_026517", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "98da5adee67d32c57c66a53afa43ed3f6be67f56"}
{"id": "train_026518", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "1742755342dbcbaedd49f5d5d966e187f914e07b"}
{"id": "train_026519", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "d93de19affaf7495fa3b3354ae2356422890314e"}
{"id": "train_026520", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "3f1fadd34f090cb940819658b1e5e5ebc17ce151"}
{"id": "train_026521", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "44268a3cb5404ccce090e641d2cae35d4ac80baa"}
{"id": "train_026522", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "cfd6faeb75763359cbccca93e62476e5357f41c7"}
{"id": "train_026523", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "27598e1961e5b310d1525f5faf6a95de2b87670c"}
{"id": "train_026524", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "1c7b131f3faa7702c356d5ba938f37b2e42ea432"}
{"id": "train_026525", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "35c486acec29948d7cadb41b813558f6787b8726"}
{"id": "train_026526", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "5d404b57c9ec5d3aca241806bda4853ceadf0577"}
{"id": "train_026527", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "296a50a57650c5fb4ef3c82abdc814fa01208e5d"}
{"id": "train_026528", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "0fae53ea6c60af2a9fc4fd67b40c3a3a53bbd19a"}
{"id": "train_026529", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "6b0d4829439a7303d2a1d0756386388c656dd16c"}
{"id": "train_026530", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "5d1660fe7cab6bc6d4fd8ab54b0a7ce951d4ebe8"}
{"id": "train_026531", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a83fa917a3a09f1614488ee59e7e591ab13ca754"}
{"id": "train_026532", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "ee940308d8d8643a517d6fb4bf9fa30c65d2a276"}
{"id": "train_026533", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "96b0319295a060f4b34adfc8e45f14dec1e19a7b"}
{"id": "train_026534", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "7aee87d911789afe1cdbcdc043539d165ef2260d"}
{"id": "train_026535", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "3882ae2aaa6f360bfbe0b049521b09a478e33dd7"}
{"id": "train_026536", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "a2f3a485e8ddc57284a95288dee0145fffa9d456"}
{"id": "train_026537", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "10992e849df09b51b3688d61b640248ea177f264"}
{"id": "train_026538", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c055ae94025c932558abd402aa340015142cba0e"}
{"id": "train_026539", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "1dc491f79f71c52c0681b62ab2cf890175822521"}
{"id": "train_026540", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "741ab9a9d449426877feb61e9a977c209478519e"}
{"id": "train_026541", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "326ac1a093531264aed3aecaf086e8fcf2f48e45"}
{"id": "train_026542", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "3ac4769efa8572b39a9470484ad407e44b4a4090"}
{"id": "train_026543", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "304457fe369bdf7e58549b2efed20981f96243a6"}
{"id": "train_026544", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "de41d778f72686440cc318248ab64ad90ade3755"}
{"id": "train_026545", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "22290eb88faef093dacd284d70570960e8b93674"}
{"id": "train_026546", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "34fa1c637efc6c5daeebdf73b9dc73398a34cb9b"}
{"id": "train_026547", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "f9145469b2597fe7a8d8fbed24a577e64b7c74b1"}
{"id": "train_026548", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "c6576eec34559a52a3c25cb339616eaabf98cf3a"}
{"id": "train_026549", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "af0e8cf3d68bf0e4f0760e6937b27a3fa3fae4cf"}
{"id": "train_026550", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "981be08bc94713ecff6d7b6978ef95e50aa9069c"}
{"id": "train_026551", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "e88669e8d422787fbe24c33b27cd065367a53d66"}
{"id": "train_026552", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "224f43f3dd065b285f1980b51e2fe9949e51eaf5"}
{"id": "train_026553", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["coding", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "239c7d769aa0298e14ca98efa10c175ed3bec185"}
{"id": "train_026554", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a6d76707e173ba7faedd3fe1ae0c19e245c6e85b"}
{"id": "train_026555", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "0bef4182fed460e535429e5ae4db75b18f01e2f3"}
{"id": "train_026556", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "d8766015e370cfcec7c6c578eb5e227b9a99a7ac"}
{"id": "train_026557", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "3ac648631d39f9b6278e82a1c8d200f3cb906dc5"}
{"id": "train_026558", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c274e2426ee4ae20c594568ca9117374049562ea"}
{"id": "train_026559", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["data", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "f54af3e352455085cfb40fd034048fd9e70b2812"}
{"id": "train_026560", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "c973a2cd0391419e3c55535e1c8e70cafea0ea0b"}
{"id": "train_026561", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "dc10cf2ec3c5aa13a4addd4083f023dc93db1dce"}
{"id": "train_026562", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "381e5d5e05bffb57d9869053516f83f6ec9e8e66"}
{"id": "train_026563", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "3394719c4a32fec74503c8a4b2b41e6bf02743ac"}
{"id": "train_026564", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "3994a9d46030efaf3b94e8651ce09e5c928f8cc3"}
{"id": "train_026565", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "80831c1510c2e02c958667cfad4ffbe77c7dde00"}
{"id": "train_026566", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "d646f74b65d61e099964ae842b4c0a50b568a3d7"}
{"id": "train_026567", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b269a5c3efc95e97ccb22a1d81bfc44ba426da52"}
{"id": "train_026568", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "3e6994ddac16e64c23e9596ebb6d5654278a90bf"}
{"id": "train_026569", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "31c82ffb73e5baeb656bd3d763b754b6d004bc3b"}
{"id": "train_026570", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "1eb8544f087b0c6b16b5255145fab145a97e2479"}
{"id": "train_026571", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c9b241958be6396827ca0feff7969cd5cb2712fd"}
{"id": "train_026572", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "66df9f989e1a5065439e599c7e060b000287c139"}
{"id": "train_026573", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "6a0aec13635bcd75069d97640abaf66d564fd226"}
{"id": "train_026574", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a91a14e4e9230cf215f240f9f48cfb4eb70574d5"}
{"id": "train_026575", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "2ab7307fb8385154d4d1089e8d432a0d786a4ea4"}
{"id": "train_026576", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "cf29b58df72fa4e756ec874f596df644651e8590"}
{"id": "train_026577", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["governance", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d2e184984486965703e61a1af8eebe799fc971f3"}
{"id": "train_026578", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e0e326255456e59c945e7f9f1b17bc0b11b5e3ba"}
{"id": "train_026579", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "2d870b7bd92b18634006ea080c107e6f3fff5974"}
{"id": "train_026580", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "2f6e578b42b5c6f29c15ac5cf692fa5549dac08b"}
{"id": "train_026581", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d3b66bce75852a7136f7b1156ec9fed3a0330e3a"}
{"id": "train_026582", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "e410ed7c52c01fd010341f23a31d74448f401399"}
{"id": "train_026583", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "16fd72d33a52c9c39d502589976b830d3a803299"}
{"id": "train_026584", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "dc95b3d4b216b7f25371614dff49088c869edccc"}
{"id": "train_026585", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "0ab7e034f6cedbc4388f5b68116bd40d519102d7"}
{"id": "train_026586", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "81314c73eef4ea4feabf21c24a25fb29909afcbc"}
{"id": "train_026587", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "cbe8df6960c6e486023f21b0a45304c8dba1fb35"}
{"id": "train_026588", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a32142f6cd6095f6355791483e4e6987e090634e"}
{"id": "train_026589", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "948f7d78c2c6542cb76bef95aa664908119227bc"}
{"id": "train_026590", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "ddb5a4aba9740c8831b435d7313cf7904ee6ac9d"}
{"id": "train_026591", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "72d88483bfdcb36134bb93b5944c57b701c1c068"}
{"id": "train_026592", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "02561af755791af61118326abc59ef9c13e56643"}
{"id": "train_026593", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "17e8212aca3f75b9d3ea8ee8b7a3de26fc3ebc21"}
{"id": "train_026594", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "542d97acc90e4ab4f38c6633a5e91b34e7da8970"}
{"id": "train_026595", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "5ea57a6a4430ce85180995b0fbda2d6e36ee761f"}
{"id": "train_026596", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "e15d5a7cc50bc9e0316a551bb6da16e93862184e"}
{"id": "train_026597", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "2cff2fcaeb7aa1880360794388e3c0aeb179c17a"}
{"id": "train_026598", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "5fca3b36be0d912abcb080e85965d6dcdb93b6ff"}
{"id": "train_026599", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "9697bdb3c594a1d41ef1d43906825764bd03507b"}
{"id": "train_026600", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b9c269d1397fd27fb259c7414b6cbf1a1bca4bcf"}
{"id": "train_026601", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c2b6efefe1e24aab94f959f966f306c57971a661"}
{"id": "train_026602", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "8403da2bc84ca312a5160d411a1ce4326c6ac147"}
{"id": "train_026603", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c1cad9ac1d09a4f065ae4c02cf48e3e03c8cfcc5"}
{"id": "train_026604", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["performance", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "d42e103fd606a8408d4730580023dbc18080cdbe"}
{"id": "train_026605", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "8d5c195eec28177fe64bf5e645a567dbc7c180f9"}
{"id": "train_026606", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "7e93bc51aafbd97ff064305bc65870558919b5a0"}
{"id": "train_026607", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "03daaf7e52f400ebff8481878c7a5312c956bd75"}
{"id": "train_026608", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "15a484160ea1382700567c9b3f738a62b5c860e6"}
{"id": "train_026609", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["security", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "2cdf949044d67140b41c173561ea18bb4bc3e582"}
{"id": "train_026610", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["data", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "0d7d9386b2175b90174ecfd435b0f5492d0cea1e"}
{"id": "train_026611", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "b599cf079314b0f278592a7e9a726cc31934ca64"}
{"id": "train_026612", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "cf0e8525fd29816f1f9c4a9c9789f6e58ca1d686"}
{"id": "train_026613", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6ff2917047d655e7fd47384ca87426e026552214"}
{"id": "train_026614", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "cc6036d68f774d2b3fcbfc28bbe5c2821b9c90c4"}
{"id": "train_026615", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "05c5bf70ebd2243080ab00711a7affbe5b4bb9f1"}
{"id": "train_026616", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "15454d94cd9dc537f4953d80cc80d05dd8530510"}
{"id": "train_026617", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "4a48f95524ade2a2e0ead54d4b5eea0d9d2c65ee"}
{"id": "train_026618", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["data", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "ed6746ddaa8799893aa0ecc9322f0e062330333a"}
{"id": "train_026619", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "5c75cf9474888fb61bb189beb14f555afb1c4137"}
{"id": "train_026620", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "9b89361a0946fc65d93f0e23fc791589f918ec16"}
{"id": "train_026621", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "860c7a53c5be46e2a853c550e6ffcbc986dd8b05"}
{"id": "train_026622", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "eaee2d0fd83cbdde646879e0ee1b2961d2ad2f67"}
{"id": "train_026623", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2ab2b6c5967c4f7a78a787b507af29093f2bd0fc"}
{"id": "train_026624", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "296ae91328a36484b3ac1363e3ea2926872cd483"}
{"id": "train_026625", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "69e539f078b77ca2626a5aea18da9c4d5a7ac791"}
{"id": "train_026626", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f62101a1032528617f0df1ca6acb89da57c48409"}
{"id": "train_026627", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "89f6ec044b894fac7663f99a2738d3ea0ca1211f"}
{"id": "train_026628", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "cbd77bdc347e3df285bd74a53764e6f318a19c85"}
{"id": "train_026629", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "15c5898bd899875a1e04729378a5c6eed7e2832f"}
{"id": "train_026630", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b92ba45710e2a4b63a69d593298e048f7c4cfe7c"}
{"id": "train_026631", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "aedf4ce24834a6e75044b6d12b0bf6cb2cd0fd75"}
{"id": "train_026632", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "1e37e49d8f7c2cf54759bd7c0a0172247e8872d4"}
{"id": "train_026633", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "d514e697d93c8ea193b9efca686c608658972cb7"}
{"id": "train_026634", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "1b94df5cf69d341fc66f42d3988822190a4d943b"}
{"id": "train_026635", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "4f4a7e2cbefb5a3532f36b11397c89b4a5326386"}
{"id": "train_026636", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "d76d125886ebd22b120d21e779dbc625a37efcd8"}
{"id": "train_026637", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "cd92fb80ee3f1abb1fb8fccf6f89f985fb3af129"}
{"id": "train_026638", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "9b1ee3e4e2bc7d14450b5a16735dfe9846d1d696"}
{"id": "train_026639", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "473aa77d11c11b6617f67d5de5deda3d249469ba"}
{"id": "train_026640", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "757f16433ceb821a8bc5c8fb8ddc6512cc74f51d"}
{"id": "train_026641", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "1304514010697fd5b7b27f34323b78bfb37fa793"}
{"id": "train_026642", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "e4f2e6e90b842844879f6c7c013af48f75bee039"}
{"id": "train_026643", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "d490072194eecd2f0ce72f360e83761736a8ea1c"}
{"id": "train_026644", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "b99443d39edc9d8d6e9067341a9e4571e40c53b1"}
{"id": "train_026645", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "de806bda1c216b4799b693703f9753c429d59210"}
{"id": "train_026646", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "195016aeb83247febae021274339a13b669b04c6"}
{"id": "train_026647", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "acf16c77362bc1a65addf854bcd613f61bf3c2c8"}
{"id": "train_026648", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "6c1bec8140921d2e164b1fe2f6cddffc81ded98d"}
{"id": "train_026649", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "c02697e8aec2a71937d2f383f4b7bd97997479de"}
{"id": "train_026650", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "082ba5a9ad965873c51c3fbee972cec33f2b1cc8"}
{"id": "train_026651", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "8491ed2f2328fe17f3ab5af6e1e2e2fa2322f109"}
{"id": "train_026652", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c9e41e37bf313e1e5a7a2e0fce535600f242897f"}
{"id": "train_026653", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "d99352e43216292bde3cbe2c8924dad17c98e1d8"}
{"id": "train_026654", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "750c3caa5b3cc72720965a3c657436249cc3a96c"}
{"id": "train_026655", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a58447e0c4af391e22bf8f19e23da6ab91ea1e10"}
{"id": "train_026656", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "29cbd4deae69394b34eea2d25c010256aad71d75"}
{"id": "train_026657", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "a24a0079d7c7b48da351c67ee247beda5ad56a44"}
{"id": "train_026658", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "ef4ee3c849430de8eefbc430653ffdc3d0cd200d"}
{"id": "train_026659", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b1ac1fc910a3b4f3a98722b55214550c16d42710"}
{"id": "train_026660", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "7fa3a8fa476aa00b0fbbc91b26a9d76eccda9a63"}
{"id": "train_026661", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["research", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "c8568aaf365f3ab5fc7e12c417cb2b6bc1f0f852"}
{"id": "train_026662", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "89b409ac29449fd6cfc4ff8f65bac4425e333ce3"}
{"id": "train_026663", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "af100504f98161778031f87b4d767671dee40db7"}
{"id": "train_026664", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "e4d788888a005e64670912c4f7a42edb962cf6bc"}
{"id": "train_026665", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "2f7f3b4e1c7dbce55b8dd12d95bfca5383dcd053"}
{"id": "train_026666", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ee1415a518e7e9c770c077f3c48e76fb52e6cdd8"}
{"id": "train_026667", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e1a2e3b20fba4d6d2490067283360efbacf1736f"}
{"id": "train_026668", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "9965bd68ce5e8450e4c9ee2183322c8b2a52d8bb"}
{"id": "train_026669", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "42b50fc4714216c3ed6384444d4a13340621e15c"}
{"id": "train_026670", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "1ebdbf509b733aa22349026fdeaba9d939298494"}
{"id": "train_026671", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "c734f75f34f9e098862de6ba9498936c986ba499"}
{"id": "train_026672", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "22980b0fdfd5efe75a15f2a1f88c0cf7ed0eaf94"}
{"id": "train_026673", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "e3831373afd2fdfeda2bf1419035880517ac19fd"}
{"id": "train_026674", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "8842ce66d36ad89bf9fe8684430e26b9fe771a17"}
{"id": "train_026675", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c3de58d72de4ee2b9f6d41faa0a62808076d1ae9"}
{"id": "train_026676", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "9d6efe51fd7f963618821385e842603b0a57b317"}
{"id": "train_026677", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "7a5c56efa811fe02419e37eba91c3d6fe4cbf4df"}
{"id": "train_026678", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8fa7e2988699f1ba25d6cf73bac2c9bf214d3993"}
{"id": "train_026679", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "dcfdece8466f01bd6a94d90c25c889a98f5cff1b"}
{"id": "train_026680", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "a6807096b568ba16005187ea1fa520009a1cd893"}
{"id": "train_026681", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "d7bc156642229da6d4cb400a6fe582bc435431d2"}
{"id": "train_026682", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "588f4546bea4cadd2de462f3763e25a7904c7294"}
{"id": "train_026683", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "a4a07695316be59b2d47e84f81bd8341c70e9257"}
{"id": "train_026684", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["research", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "5207ad9d7df02ac1e0bf5cb7daeecaae031ca41d"}
{"id": "train_026685", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "c1734ca1d9728a761e4735d5c0f0eacc15c6cf15"}
{"id": "train_026686", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "d7c3a2582823794d3e21938ac2861268c8a9855d"}
{"id": "train_026687", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "07553415fa80a7a82e1053504056d6eb1bea7d05"}
{"id": "train_026688", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f87232cd658c0a3cfd8af758248fa96cd5e5cfce"}
{"id": "train_026689", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["security", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "1f6471f1fe30a678cf7991be9c9abce59be20086"}
{"id": "train_026690", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["governance", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "e13b3ea6fd8ea44faa65216f4e0d3d1c30d3801a"}
{"id": "train_026691", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "cd857525a1de2d6ec3dc1a6348eda66328c80331"}
{"id": "train_026692", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["security", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b30c1e70a0fe1c4c54c4bb8178164249bf83df8d"}
{"id": "train_026693", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "e53100faa005b1bab41feb3e7baee78b98cf2aab"}
{"id": "train_026694", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "3959520789d8dee89caee9c042938dfc015f96fe"}
{"id": "train_026695", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c930d1e8383e7b4f1d51501e223fafd88c7521b3"}
{"id": "train_026696", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "2fccb0c5deab1b293e3a067f96e32a212dff9f31"}
{"id": "train_026697", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "fd6c55ff3664b0e9c71581cb31ea49bca90bb311"}
{"id": "train_026698", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "03977ec0e2e52efbdebbdc7de030187bc9786bad"}
{"id": "train_026699", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["data", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "7208d55b78c2d55780a5af071ef259ff7e6a85b0"}
{"id": "train_026700", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "5eb562b1babbcdaa74f6166dac5e8e3a56c06909"}
{"id": "train_026701", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0c9179b60abfa8633eadf1ea7d3dd7e27e414c41"}
{"id": "train_026702", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "e4a59379935fcc045e083c72b8687a8cd6fde889"}
{"id": "train_026703", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f413ffb53cb50cecf9da7d44eaa8dbc5552ff694"}
{"id": "train_026704", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ef6a5c0d40a8decbb654c67ad112317d7b165218"}
{"id": "train_026705", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2f4ceebd2697b921b04d334d57ada19e9de229f6"}
{"id": "train_026706", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "4068a1dd4a6e7f294afa29f2c1be297f8832bd27"}
{"id": "train_026707", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "224e9feccd9d53f9ee52389dab80241db8e322e2"}
{"id": "train_026708", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "b84970d2c2ee333267af591b4fbbc5162c81b545"}
{"id": "train_026709", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "d93de19affaf7495fa3b3354ae2356422890314e"}
{"id": "train_026710", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "96ab87e330ba7454aab8870537676c5ed71bf411"}
{"id": "train_026711", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2752ffc5cc37938a6d9be3aac63b6e135121c756"}
{"id": "train_026712", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "c664fd32416037e33a135585cb20ad57b525ccae"}
{"id": "train_026713", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "db2bfc36a31f333a50dbed2d990f76cb0c14af6d"}
{"id": "train_026714", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "154ae6b5f256e0eab7abb81d8cb82f56392be224"}
{"id": "train_026715", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "586020d4cbac488e23613f28d2efbf1b9b2d24e3"}
{"id": "train_026716", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f6803572862187930deb3bc0b1b67c73abaeb9e8"}
{"id": "train_026717", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2f99146a6baf28ab3d60d35e89a550d56fb966ea"}
{"id": "train_026718", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "9fa8d32500e1a8d5d12cfe21a5ece5b17d2624fa"}
{"id": "train_026719", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "d35e50ebd48feb9d8fcfa059e64ea820b2127538"}
{"id": "train_026720", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "db419701eb9b8819898d52379cbab52e782d18ab"}
{"id": "train_026721", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "75f19f084855cdbe019301ad6dd73ce6e3f9a229"}
{"id": "train_026722", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "4333aef856b9d7614b87b1fe2f2f5e90b9b4108b"}
{"id": "train_026723", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "0302afcbca45d4961b34a940726c0e8f52ae1472"}
{"id": "train_026724", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "708e5058ca8f3432a6380dbb28a4037ce704eb19"}
{"id": "train_026725", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "47dc9abfb3f02aee0787d4d00047f812f70f55fa"}
{"id": "train_026726", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "f19c0b3d3c121a4c482abab72f7b48167b7e552f"}
{"id": "train_026727", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "686f2b4f58111499b3278a85304ad22048eadaec"}
{"id": "train_026728", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0039a6f98bb725bbc0f0c073e8ca8c742f3f02b6"}
{"id": "train_026729", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "b899761e562ac6e55f6f80b22da736c0ddb49a18"}
{"id": "train_026730", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "4712de99f91a508e01b6af995cffc10b1a34240e"}
{"id": "train_026731", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "61eff8cb86d4afcda687c823e6354d1c8434fdd1"}
{"id": "train_026732", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "774bbbeeede71cf886a0eeb873726032777a80a2"}
{"id": "train_026733", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "46a6d0f71216c2e7e22301bee3ae81de9706c8b7"}
{"id": "train_026734", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "345d0a6d2bf87de2dac5e27925e0eae582cacef9"}
{"id": "train_026735", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "dc93c6a219db16cdc2dfd6d95f2079876f70d31f"}
{"id": "train_026736", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "8b136055fd5ab956b2de686c828a931147e63a11"}
{"id": "train_026737", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2865ea40b081e2cd1491e45ea076c8d3fff59594"}
{"id": "train_026738", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "0968c03ea38c45e5e27110cf91ed9f1610a22199"}
{"id": "train_026739", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "afec30f18a9c74767f1860c531172b19f2786134"}
{"id": "train_026740", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "a11b161214d83700c1503279a910f0d9b0fc180b"}
{"id": "train_026741", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "390f6dd48612938dcd9bf831b05401bd6e6814e0"}
{"id": "train_026742", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "cc2acb972300ddb43a23b2cc827ee9d6ec100abf"}
{"id": "train_026743", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "2406b4334a1c84ce4f236572d48c63d68aefca05"}
{"id": "train_026744", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["performance", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "ca19a86ceb074347ec497393f2cf5b058859abda"}
{"id": "train_026745", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "7fa07c9199bf56da5abb0da596a7179fabcde75b"}
{"id": "train_026746", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "eae4f490e8252baaba37a26a6c6e6bf8642fbc9d"}
{"id": "train_026747", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["security", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "a9b83a21a5cdd4a66d839a95aa00f28d9be21d6e"}
{"id": "train_026748", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "5d5595d0e6b1edf0b6cacab9d534d876a72b8513"}
{"id": "train_026749", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "73c5f8046609c7bbe9348dcb6afa247be7089801"}
{"id": "train_026750", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "0dbf4ff665608cf8d54f10f7b6bfce308e7983a0"}
{"id": "train_026751", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "9bc9acd1a1bec79b7542e73ee61f99df36054439"}
{"id": "train_026752", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "d6f2e459424d7d4335d5336ad44250bf5301c52d"}
{"id": "train_026753", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "a47c6959465216b510354b7bd7d419857b76de61"}
{"id": "train_026754", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e4b1f098e79f15c58cb530b74d476860d435e0f5"}
{"id": "train_026755", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "d64efabe726093192460ae46f0411fec82f73cb1"}
{"id": "train_026756", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["security", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "8d1b181c853bc6ed95d0ea42a6bb9c34efe92288"}
{"id": "train_026757", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "d20fc3392dffbcf112b63ef6029b6fe0e6eb7055"}
{"id": "train_026758", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "1e47d14a6772c9b91a12c131d2e49f6c20154e74"}
{"id": "train_026759", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f2983b3902b1c13871da277f75cba40f14943d52"}
{"id": "train_026760", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d426c8a9d9f78c8596620e73ff70596ab1e39c66"}
{"id": "train_026761", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "d9b2e45fd8ecef011ff76c63bf728b2ad0b59e6f"}
{"id": "train_026762", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "8ef6c5bf727738587042669941ad5a41e06f5abc"}
{"id": "train_026763", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "7626ae550cce9691b0934178a2bea179171df907"}
{"id": "train_026764", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2c7663cf70ebb8dd508174b758c64c0f5c3488a5"}
{"id": "train_026765", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "bdc4a1e3eed9ee510c7d099ac5704fa99d925dad"}
{"id": "train_026766", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "98da5adee67d32c57c66a53afa43ed3f6be67f56"}
{"id": "train_026767", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "8657a6d485e0b3f1a5ac7a8f7d8016aadcfc802c"}
{"id": "train_026768", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "9cd44e89796c73b3c48e6208a3c352626bb0f6c1"}
{"id": "train_026769", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "943e954340a048334c7c5ad96d37c201f8093ee0"}
{"id": "train_026770", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "aca90055463714a920406d9c8ee0f7c021a7f0d2"}
{"id": "train_026771", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "beef8615e262fc86ed46a350180f3a15ccc82e28"}
{"id": "train_026772", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "2ab17f50df5ee830d501d849dd9d3f6f43c3a24d"}
{"id": "train_026773", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "9dd5b0b6af59146721ac358edee1676e04ce6189"}
{"id": "train_026774", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "06941dc6f66334fd97678ce761009afd852ab9d3"}
{"id": "train_026775", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a8f3f5ba0ce7c3ef2ba131b1a9aad121ed93e0a6"}
{"id": "train_026776", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "0e124103ee46503b16ef173c5a20ab07b9fdefec"}
{"id": "train_026777", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "eb6e7685725710adc5be99e21abb70d33d041272"}
{"id": "train_026778", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "91b26a41443870e5bc894188171d4e4fe4a0331d"}
{"id": "train_026779", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "7966d56203f18b6cd10238a48797eb2916c86a2e"}
{"id": "train_026780", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["research", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "481179026b261e82d9814873f9179bb36c8c291b"}
{"id": "train_026781", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "701a036a00c5d6f0ca60ccd2f1504e0d8e49c9e2"}
{"id": "train_026782", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "187c367208e489acffbc283f13ba94b6b050efad"}
{"id": "train_026783", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "448b29b035826ea187eabcb57b16751ff11f94f5"}
{"id": "train_026784", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "415c45580a50729e670118925be44d1cb7336c5d"}
{"id": "train_026785", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "3614b71a194f1862537dc0ccf14fe9125083201f"}
{"id": "train_026786", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ec82ecc28984d8c466fa99f243e71a9fccfc507b"}
{"id": "train_026787", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "f42f0c350677cbedba664d6d8cb31d65b43f4eb9"}
{"id": "train_026788", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "030a82835f3535a12d1e444c369db84da74c52f1"}
{"id": "train_026789", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "c99d7cbabc0e954cba2a95fe501c5415c155f30a"}
{"id": "train_026790", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "8165940eebc3729b93f5affc2a88393e91641712"}
{"id": "train_026791", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "d0f94eccf9bbe0a2814d91c40fe914eb668ff43f"}
{"id": "train_026792", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "85ecc5c902fef2b3d6844a172e160e810f207610"}
{"id": "train_026793", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "4a6e843446c41ca468f794d6ef30cf448d60cbca"}
{"id": "train_026794", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "8151508714d125800d068dc7efad85c5147f0df6"}
{"id": "train_026795", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "9804bdbe3f24e7771cf0e28882c8d454424a88bc"}
{"id": "train_026796", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "2e5d14f02a12070c9486d43e4a1b9942b71267ff"}
{"id": "train_026797", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["research", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "47ca2df8133fc1f3b86ecef8b35d7c8cb115f2df"}
{"id": "train_026798", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "643839cde00ebdcdbff3fd28e55c52931562d6f0"}
{"id": "train_026799", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "4b539f53a395b9a47b79f8fcad4de2804a4dca06"}
{"id": "train_026800", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "7e47cbc0fd8eae744867e30f70d0c40fa3558990"}
{"id": "train_026801", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "ee783e2f5d81cb7e49c9b527291b343135c6b3e9"}
{"id": "train_026802", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "d1a2a63853da0441aa8fe09c64944d14ead41d77"}
{"id": "train_026803", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "52117e20acaf9870d35a68eff0c3da6c5bfa27bb"}
{"id": "train_026804", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "03d493f27febdc912abf2442ff337d76c658879d"}
{"id": "train_026805", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "b86c8f172a4cfe604717bc2b7a30993c22117620"}
{"id": "train_026806", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "a78a3a30f993f7160e9f7961e2382a74bd5ce9f4"}
{"id": "train_026807", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "745d9654cbabdb6a0455c0bf9c88476458f8998b"}
{"id": "train_026808", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "319a6ae90ca1ebcc56a7a7b715bc2c90bfac7f67"}
{"id": "train_026809", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "63e8c951ca7f512054d911c3e02f66c2d3049a54"}
{"id": "train_026810", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c304778b8614498cdcef550419cb9af80b6502cc"}
{"id": "train_026811", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "fa3051e5bf2ae5857616afa0c2363522df4287bd"}
{"id": "train_026812", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0ad38ea133e64d531f78ba1e0127e79de43adb58"}
{"id": "train_026813", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "76980d4128f4fd5c05c7886b30472dfcb2641e19"}
{"id": "train_026814", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "e0eebe7440afbe70d64be2912a41478962b60c9b"}
{"id": "train_026815", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "9d916793f091bc28e162b332c0ce112ff15dcf95"}
{"id": "train_026816", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "19f19c1fb746aa966385e88d834b2984854dbf85"}
{"id": "train_026817", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "ca5e26e305f5f592331e429b9f585d59e2a786dc"}
{"id": "train_026818", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "00f6b72a17fd711dd23ec429e0ae5244348400e4"}
{"id": "train_026819", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "101231df9d3a763bc8aae35115d1e3fc6c3be120"}
{"id": "train_026820", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "a408ff9bed6ba566a1309924e42f2b362f6bde39"}
{"id": "train_026821", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "a4f7971647fdb6cdbfd0526666ec2efb1576b053"}
{"id": "train_026822", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "0ddc4a836cc93606cb37fd449922157600dd6049"}
{"id": "train_026823", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "3d16f02808a1c2fcba6a17441b97bccde3184667"}
{"id": "train_026824", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f5087718a0555aba840fe9da344a86634c4fa2fb"}
{"id": "train_026825", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "b9449ae0ddbc54981932010a5e13b1de95460359"}
{"id": "train_026826", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "84874f278db1f6e98b6b36b314061161cd437f97"}
{"id": "train_026827", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "e3e8a5f407a7ad056c68a602504338d1a36d73d6"}
{"id": "train_026828", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "fb391519d5591ff058c5c378a13e893658fa37c6"}
{"id": "train_026829", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "4f321550d2872fc933d3ab0cca410f82e64fbabd"}
{"id": "train_026830", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "5fb254dc96de9017db59092a2fbff43001baca95"}
{"id": "train_026831", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "080a1350ad95ce284931f076cb0d860300090459"}
{"id": "train_026832", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a6d0c8bddc5b7f9ffa6cb9a0b0cdc90320520a32"}
{"id": "train_026833", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ec77b5a3f064de960ecdb887f54e967162af1072"}
{"id": "train_026834", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "776120aa58f7614dd4f1765ddc58eb89c25730ca"}
{"id": "train_026835", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "fa5db579de9b347779b242c3dca2ac6769688c74"}
{"id": "train_026836", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "ae259190a5cee6cb7c0be08fb8a5b6378268681f"}
{"id": "train_026837", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "714fcbb7f555f85e40bc701d5b73e1c73a4bc23a"}
{"id": "train_026838", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "8ec7587b6ce7c5b02b8ecb41200fd8ba36730979"}
{"id": "train_026839", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "5d99586bc06bd81208c069c417999dc173645e1d"}
{"id": "train_026840", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a161f9a7edaf0d02332690ac08fae2be51de4370"}
{"id": "train_026841", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "6cdbf07ab2690393bff11fb172a2d9d5783e08fc"}
{"id": "train_026842", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "3e633bd793dd03a4d202f90638b001cf264bb8c7"}
{"id": "train_026843", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "cbc72d363eb41b38e595e01e0348adbe47f223fe"}
{"id": "train_026844", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ed125307f0f99b71c5479f8f290fd5804f3f0b60"}
{"id": "train_026845", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c88967ea78642e624d04dda07ab246c77c884338"}
{"id": "train_026846", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a856ff1f9afb2af163e85972e11efba953074312"}
{"id": "train_026847", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "5d7b63df8ec7e4a3004ae6ba1c3792daa4594258"}
{"id": "train_026848", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "4d9edea0bdd51f845e1d711913f641fa9b57ea2e"}
{"id": "train_026849", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "8c20e726b5d5218b04a3d00982ca83cd817b356d"}
{"id": "train_026850", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "272b034f1a7f9b1043bcab12d82ddc340132dd26"}
{"id": "train_026851", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "76f9178df81388077cd1d5e759c4ba19d795c628"}
{"id": "train_026852", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "8c5b1bc2fbcdfc51dcea89d0818423d4bd89775c"}
{"id": "train_026853", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "e2483a66dcb7f1059b662f2d065f2ebeea042930"}
{"id": "train_026854", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "57d5c02c0fbc19915dd90f0e17d570446b95249b"}
{"id": "train_026855", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "214afcb5decaf5c710cf7e246ebfd558c45a7b96"}
{"id": "train_026856", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "f6355e42e40f25185bb943a784f26c5dd8544ec1"}
{"id": "train_026857", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "2232a921f4146b6995008ef87456ebdd9ca812f9"}
{"id": "train_026858", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "ccd4f9b70fdfc256096cc48f37b3407c626b95d4"}
{"id": "train_026859", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "79b3548f7283e95f3e5272ee433fb303d800cf67"}
{"id": "train_026860", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "f6418197f72cb026be1264e827eda679718138bb"}
{"id": "train_026861", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "1be7f05da5ad74fc21ce557f5dd5201c21464d44"}
{"id": "train_026862", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "8a2fbde8cb6d3e4a13d2a06c37c8d59509dde24d"}
{"id": "train_026863", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "3aac82a2f3cd5e9624bed581e9c69929fb25ea02"}
{"id": "train_026864", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "7f6b9011630e51d2b326432650dd12a37666660d"}
{"id": "train_026865", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "566445605b78bc782ba57ceafc6b696bf6af2f4d"}
{"id": "train_026866", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a5e3e865c9e832293ee15e5cb56136c1aadc3318"}
{"id": "train_026867", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "966e1d6fa34d85ec62f3e39b50978eb426034b01"}
{"id": "train_026868", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "5b633c1c9451d60faeb2541ffdbcd1dfc59d459d"}
{"id": "train_026869", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ef79b89d9ee03b218457d5504fc44dcc974514a3"}
{"id": "train_026870", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "fb967125194353c30b118e3f56e92cd1f2d8e74a"}
{"id": "train_026871", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "197856f9e730ef98b10faa42255cffcbc2a3819d"}
{"id": "train_026872", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "bec18459a62dc511897633145509f8c33d7dfc8b"}
{"id": "train_026873", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "45c4b36bb3d5f702e102108b7bf299f89e51b19e"}
{"id": "train_026874", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c734f75f34f9e098862de6ba9498936c986ba499"}
{"id": "train_026875", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "1538eda49604d8df967dd8d6d07615ee431bf8c8"}
{"id": "train_026876", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "b7c4483be539c67957fdd9ba64dfc84f683e7299"}
{"id": "train_026877", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "8557e875708c2f65dbc53be278594c1ba1c718f2"}
{"id": "train_026878", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "cda448400c2a343fc2b37a1c852c5fdcbb56d557"}
{"id": "train_026879", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "e578865f54a1ab8857234e8733a3e0190ade8f40"}
{"id": "train_026880", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "334af81e571b2058d9cf8463827272ba49ffaa37"}
{"id": "train_026881", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "c4869bca68f9bc4c03a03c0415a17f49368a718f"}
{"id": "train_026882", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "b9fc8c635b671f7d3bc9df6bdd1f2ca7e43dc871"}
{"id": "train_026883", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "229b45a2e078c56812f8f360593652cb4294b2f3"}
{"id": "train_026884", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "9d2eda413bcada2da67c3ad8b76eb9c07f235fbc"}
{"id": "train_026885", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "2ba02266219edfe39b4160b5d789701f84811c35"}
{"id": "train_026886", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "14da01080daf8025c57a661a26da2de12d9167e5"}
{"id": "train_026887", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "fff46a187bffc8a1b623f140c8d91b4873b18e96"}
{"id": "train_026888", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "92edf949c7f22cbc0fc3dd64d6bc7f5198b0c12b"}
{"id": "train_026889", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "6fce26b63d1f85b43df7ccc5e6f40f4419a92025"}
{"id": "train_026890", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "dc34b2378f9b796e02cc1d6e1e780cb657b1224d"}
{"id": "train_026891", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "f165a8a0099a6be3242f88305dffb16d67b80c79"}
{"id": "train_026892", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["security", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "1797990e0a26fde36fa40a719c2cdc77b820ec59"}
{"id": "train_026893", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ce3f80022d47ab1023b3d7e005c069de44850206"}
{"id": "train_026894", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c2e84a620aad7a0483be51d6c1738257b08f826a"}
{"id": "train_026895", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "66e784ea369b5a46a0562d705f411bb27f64c78f"}
{"id": "train_026896", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "26b97e07197070440750377051b2dd945b89f2d4"}
{"id": "train_026897", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "c7455b8e188acb0cc050c446eb8429aeebcf39ac"}
{"id": "train_026898", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "f2cd6b3d65da1eaaf8c6446e072a20cdc3bfb7e4"}
{"id": "train_026899", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "dc20d822792c9fe4eb7a0ea4ef6e68362c8aea19"}
{"id": "train_026900", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "220a45166b53f686f83bedaa300ffb970d7b4e79"}
{"id": "train_026901", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "62d4e874ea9c719204233092e949cac36c9e4a2e"}
{"id": "train_026902", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "1a958eeab662b518fc625a7e96c0b60a5e738520"}
{"id": "train_026903", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "17ad0c70f7e6d9cc808f45aaf77dbe61f7606e24"}
{"id": "train_026904", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "fc3a96dffe1a961c4404818d720eab9020fe0280"}
{"id": "train_026905", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "cb1ce73bef491be806cb3a2aea1c59c95e6aa056"}
{"id": "train_026906", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "557c59259d191b47662216f54eb8c9ea83455216"}
{"id": "train_026907", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "78a6bdfcd1667a64fc7bf488f4ca05f07793171b"}
{"id": "train_026908", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["coding", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f3072a6ea5ceac8fdbea33149e9a59de3c973d07"}
{"id": "train_026909", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "08203c4312e2c36e9981b6135e45279c75811685"}
{"id": "train_026910", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "36b57e37b5fd8f833f818f1e3148b2476f47b821"}
{"id": "train_026911", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "8fd0295b9648940ea6b1a76d4fb2d4f27c996a99"}
{"id": "train_026912", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2957d0db4bfa5861c5560af8bd341a507776f002"}
{"id": "train_026913", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "187ef2462c6b6cda1d739c0f14fd2c9156f78dd1"}
{"id": "train_026914", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "22d3a91fac3eaf53668c4473de391042d9d73909"}
{"id": "train_026915", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["evaluation", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "82a3ffad3d8efed82263f6fcb8ba1d23de3e77a3"}
{"id": "train_026916", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "679121f768dc30b59db9f04a2054422a1212adb8"}
{"id": "train_026917", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "095c9de977249ecea23b0e24729cc9aa331c8746"}
{"id": "train_026918", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "6d840882cb05ddcd28d28bbd076ec96fa5d134c4"}
{"id": "train_026919", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "8a927be25e58bd34dff7c47b3c4a947fbe46881b"}
{"id": "train_026920", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ec86fbc9541c9d32fb80f8ca2ed28b54650c28bb"}
{"id": "train_026921", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "213e550f2f4d0d9ce0e2039396dcbc893fd59842"}
{"id": "train_026922", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["coding", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "a03e075ef356f884479755acea7b31217f2fc8e7"}
{"id": "train_026923", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "7f90fccd86239d4b88d660edbebcd3c20d900b83"}
{"id": "train_026924", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "bdea4e5385fc214fbdfa24a582b207377eab9bfc"}
{"id": "train_026925", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "16cf1267b788fba85505b751de14603523400139"}
{"id": "train_026926", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "de6745ccb479d6d8f158817aa1359734981c095d"}
{"id": "train_026927", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ef1277b93b77849cb02897a44a7e60c09e8719bf"}
{"id": "train_026928", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "cc31d30ca794691d68a5ed10d7cb18fa131c2f7e"}
{"id": "train_026929", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "fcc5b764708540b97955602ab0325e2e316a4543"}
{"id": "train_026930", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "4a9dc9f6082fa526199e79425b8261e8c2a9c5f5"}
{"id": "train_026931", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a588b072b0f7cef8522adfd8a3190460e17010a0"}
{"id": "train_026932", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "76f0cf2609247867d002609af2b4ea8591968d3c"}
{"id": "train_026933", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "b8a369f20d1c4785902350e72b99d62b185cb345"}
{"id": "train_026934", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "5b88b25a754b90243d56f34bbe275d653aa70c6d"}
{"id": "train_026935", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "65fe7380be8622cc6c720f42ec901d387e7b2f5a"}
{"id": "train_026936", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "13f1cd82e94333b0e2e85d589cdd81e24e426c9e"}
{"id": "train_026937", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c88967ea78642e624d04dda07ab246c77c884338"}
{"id": "train_026938", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "925beca824ca739536c5bec0973ee36429432041"}
{"id": "train_026939", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "7c37e065d5c25f70f685d7601795e27ea8bc9cbf"}
{"id": "train_026940", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "83c01b421afb3e5afd7c91da8317242ce3aac573"}
{"id": "train_026941", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "3074050de7f1400c4162c0d75dd7dbf4e59b2e37"}
{"id": "train_026942", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "cb2136ccdce9db0aec891e05aca32004627c3c94"}
{"id": "train_026943", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "2df37e12f2e57d0c980ce8ee65097def3b181ae3"}
{"id": "train_026944", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "300fcdfa222151351cc7b17dfbf0300ab78b90fe"}
{"id": "train_026945", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "19e3f1630ed57b42233893334bede3589107fc49"}
{"id": "train_026946", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "d25515a77dc6791bed56993c644965c8b70c56fe"}
{"id": "train_026947", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "d19f0dda46923abaa3a26640e35655c64f4f165a"}
{"id": "train_026948", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "ea3f60cfa0876478e42f8d9bba85eb24b905efcb"}
{"id": "train_026949", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "bed38597bbc7c9f4b57b6f59134142eb974d1612"}
{"id": "train_026950", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "345d0a6d2bf87de2dac5e27925e0eae582cacef9"}
{"id": "train_026951", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "90153db059339e0f753954bbf57e24e135a0a074"}
{"id": "train_026952", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "142a9469ecfabd2488398f1ca013300abe73d502"}
{"id": "train_026953", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "9877c5dbeb556c1badd34f5b405f2688b21436b9"}
{"id": "train_026954", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "be6c41263244e23a72838bf37ff14a27bb1485ea"}
{"id": "train_026955", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["agent", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "e41bc62daa58b600ae09e88a37b098bc5445faad"}
{"id": "train_026956", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c32a5345c3b9ce5130b39e1c7e9cdc5b2b7abc48"}
{"id": "train_026957", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "fe09aced8d18325ddb4cf1b6e0796e1e52f35d21"}
{"id": "train_026958", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "e2bab2f8163b0b554833952e35fa72793ae22d6b"}
{"id": "train_026959", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "d7804ae9e6ddf61fbcce625cdca702196fc1e2a8"}
{"id": "train_026960", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "eba3182ee98952702b2e24a954e86b32f9359cc3"}
{"id": "train_026961", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "cdc367a5dbcd938e715efdacab8981dae64bd9f5"}
{"id": "train_026962", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "3b86152527e153e1bd92cad0c2e516043f8aa02b"}
{"id": "train_026963", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "fdbde1e1f8923508b301e49b8cd52894e503aa39"}
{"id": "train_026964", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "672cf6a65211822914f0bbe1d03ead5272aedf78"}
{"id": "train_026965", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "41151031f3a7e6633d502a08214322bf64bda25c"}
{"id": "train_026966", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2889cdae22f80fbe919b99157253c84601243ffe"}
{"id": "train_026967", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "3f8b300be99b335f422bea2b57f529c5e0c9b51f"}
{"id": "train_026968", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "9b6cd909d7c19867d12c26066ef5de3763d86666"}
{"id": "train_026969", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "ff455a3164b917f9bac7b3f75f7e3ae1642e1ef1"}
{"id": "train_026970", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "9f3a96a5649e09801b262ab0cd132653d072f3f6"}
{"id": "train_026971", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "690d5733e0c54738a6fad18dad5dea2a248d0365"}
{"id": "train_026972", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "07f188510eea38b69f2d1999c75d35898c4cf8a7"}
{"id": "train_026973", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "aad9d5b5ec22c346aef8536996d2eaa4e8b823ac"}
{"id": "train_026974", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "7434cd007340c5c18fa6e18e63c42e26f7864bf6"}
{"id": "train_026975", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "d3f648e1a42da1bf3fea3e927d1c816c7d84b8a3"}
{"id": "train_026976", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e9fbbf3dbf44aa1af2ef4d160759ab15bd42292a"}
{"id": "train_026977", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "e9cc55aed05a6fb94b12da27e32ae1faf95afd30"}
{"id": "train_026978", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2bd66369e420db151d0ff41493dfeb0e82cc7e03"}
{"id": "train_026979", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "b8e6af3100f315e6077a155873d04287ccff21cf"}
{"id": "train_026980", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "8a0b9e2c6c4a1d9cfb51b2f8c008be1242625128"}
{"id": "train_026981", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "91265d92afdd1ba741684c7686d1e452a7182322"}
{"id": "train_026982", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "75ff9027ed772e00e37dd512698de3b868a6c67e"}
{"id": "train_026983", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "0e24441e4100f858cd6439c685bf71865ca970b9"}
{"id": "train_026984", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "3ce2398c9c92b7d4aa1c431fc9a564df2d328819"}
{"id": "train_026985", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "1687d8308f5426f659d5c20d850d1a9debdd3d16"}
{"id": "train_026986", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "988a0df66d6a24ddcf6f881410e1b2ca9636c7c3"}
{"id": "train_026987", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "6622d0cda615719625bc98e33aa34fe794cf264a"}
{"id": "train_026988", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "ed72ae003db6d1203c9757fd56860031a4741a80"}
{"id": "train_026989", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "926fbfe2af32eb82a6ecfa3e131074af63a52597"}
{"id": "train_026990", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["evaluation", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e519097dca3e6dfc5038dbb8b0394bff9a15f3d9"}
{"id": "train_026991", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "fd0e2e02b23ad66ae419950c2b1bfc70ae64d6fa"}
{"id": "train_026992", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "c6accf54b8dd632a97cfc58d0dbff05225412255"}
{"id": "train_026993", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "70766f104729592af956911c5f6253052d46c1b3"}
{"id": "train_026994", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "73a54de03fa280a6c8313621bc29db0f4114063c"}
{"id": "train_026995", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "19e915d30bef0aff5e28bde3bae05135db8b8d11"}
{"id": "train_026996", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "916f8895aa83c5ebcfed036411ddc8af76ccfe20"}
{"id": "train_026997", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "bfbda63b7a55b6895bdcc18a44d0d14590ba564f"}
{"id": "train_026998", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "0f735dd4649f37931fd31bce87a3f1537234b5d1"}
{"id": "train_026999", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "417711e762c7ccfb82e132e02045285923c36bfe"}
{"id": "train_027000", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "ecae3c7b1da970de3faddb6bd5a933c8ffec25c8"}
{"id": "train_027001", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "486fba8970f07ce23a01ea5c93c1213dd0d70c66"}
{"id": "train_027002", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "9af9562d7879b67adf2f1126ee1c7cb3e7de2c16"}
{"id": "train_027003", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "5de52ddf3194776a11dc0c21238af88edda93100"}
{"id": "train_027004", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "572341656f4fafa753852753584b4ac2bb8f221c"}
{"id": "train_027005", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "cf73d2de04d107b2b36a3796d5dc03f6040c07a7"}
{"id": "train_027006", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "14d06f97f59eeac520d3052fc13f69e9cd1e1305"}
{"id": "train_027007", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "c2eb1ef0389131ad77684588804ad02ca6ff374d"}
{"id": "train_027008", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8e7b13782149e8534cf5edf8eed541080630bb87"}
{"id": "train_027009", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "bc511480d625049aa6b3978cdf6fb725b6cd1a17"}
{"id": "train_027010", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "20c132e3beb119b1d2ab68039f03c8d3cd97a862"}
{"id": "train_027011", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "844e9a5eff98f2e7c02aabf4624f93fc112e521e"}
{"id": "train_027012", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "182481d13560aead7bbc910f549798032110f9ff"}
{"id": "train_027013", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "9e06b8376319825fd8269e2839e644c3160fd9c7"}
{"id": "train_027014", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f5142426e5c5e06ed9b4c2af4c6c344470f21042"}
{"id": "train_027015", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "9ce9639e0339f2bc232e20ad34781fecb34a328b"}
{"id": "train_027016", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "1cd28719665934b55db7b9c835cb3196529dacda"}
{"id": "train_027017", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "e6e5e775fdb3c0f12eb1507e005db5ffebf5f41d"}
{"id": "train_027018", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "fdb9bc84cd73785e32952b514cd554740abc9cc2"}
{"id": "train_027019", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "bc87e8fcbaa17f09f80d06115555830e4735fac1"}
{"id": "train_027020", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "583816e20ba411dd7b31dc86a7dc8460e2bcf685"}
{"id": "train_027021", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "f334495ccc8a732ba320c4bb8cfa9bd0dde56663"}
{"id": "train_027022", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "b5294d4ff15dbeb88cf73d6dc43d04ccdef16389"}
{"id": "train_027023", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "2c59bdf91f4852a1980c59de0976722821c1e1fa"}
{"id": "train_027024", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f8558a672e6dd85df5b38bf5586ff7b81eeb4a47"}
{"id": "train_027025", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "60e47b564d93015a05617c51308475242c2b9616"}
{"id": "train_027026", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "8f0ce6db3cba4e8882d8e0ed594b028bb7c8b139"}
{"id": "train_027027", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "4dc266490b011222cfa4a46eea0fb932fd44fbad"}
{"id": "train_027028", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "41fc592f41159ea277eb4184f52ea6195a9656d3"}
{"id": "train_027029", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "72598a5229596d90d294d61b66263d72672b3391"}
{"id": "train_027030", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "91687e7b2077e1cb13b4eab5ff48f4fdc0316a67"}
{"id": "train_027031", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "0c3f12fe7dc5782ad8cc66b846a49c622314bc03"}
{"id": "train_027032", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "3118d4563dc6523ce7e07ba78a529ae99b643cfc"}
{"id": "train_027033", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["data", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "cd6eed813887c2ae14926383bd7ffb718a916e95"}
{"id": "train_027034", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "6073ee9db4be70b2b39bc05b17b22dafff13938b"}
{"id": "train_027035", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "c2d1d0609fd782ebd074c4c6a086f72da4538890"}
{"id": "train_027036", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c63ac8670c10f1e92492326e68b8016907551939"}
{"id": "train_027037", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "686fba71ceca27e364d3372d6ab536e5846172a4"}
{"id": "train_027038", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "0c01ffe135fc7fb7e93e26461cb4ab52552fc3c5"}
{"id": "train_027039", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ca0c49cae6a1c73b3f1be8c43afdf26375e56173"}
{"id": "train_027040", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "f217ae30da6bdc74c1de3b373e635c9b061bb0a5"}
{"id": "train_027041", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0b0322fc8ccb919922117b56f26ba832afb4e8fc"}
{"id": "train_027042", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "954631db8d44ac7938d77fb9d8248c804ef478eb"}
{"id": "train_027043", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "4143d4dc8e193735b7258354b1dc90935ac4dbaa"}
{"id": "train_027044", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "d738b1c35a8cc1effc0d6f7bb98f68649a201f6e"}
{"id": "train_027045", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "23b2ce3f615c23d2d0394e30903f2532792eb354"}
{"id": "train_027046", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "2debf7ca372b40695213ba61900eeca7fb6fccd7"}
{"id": "train_027047", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "581141e376c77cffd8c2bd0b6758d263d30cd884"}
{"id": "train_027048", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["security", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "1418a893a3df914c8a13695a0680c0004fc411c1"}
{"id": "train_027049", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "7baecb1b5b32aa5991db795689ea84339c638f96"}
{"id": "train_027050", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "4ee394358beb71c07d45a0679917e5fd23cca195"}
{"id": "train_027051", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "c6011458ca9326c58bcbea6a1a156358b0f1b0d9"}
{"id": "train_027052", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "8acbd90e6eec445040a41da39a355aa6f333a434"}
{"id": "train_027053", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "3d584ce2792c8ef57689d87fa8d2e075592dc11e"}
{"id": "train_027054", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ae6319b29c3d21b0d9e25048fded357f17e24b48"}
{"id": "train_027055", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "deafccee74d1b9fb79e9d6937bbbd4eee9d9f821"}
{"id": "train_027056", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "25f1329e12cccf55c6120673f87321a5592124a0"}
{"id": "train_027057", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["data", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "8eaa719891d6ac53f6553db1198035c84d3e6a0d"}
{"id": "train_027058", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "1f158bf72f926baa5f3ef62fc33e8fdc22a4d8e2"}
{"id": "train_027059", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "a40f712e5f446c5ab15a0630bf21b86bd23715f4"}
{"id": "train_027060", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["governance", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "1f368e8c550742776f487286f177c7c2faed8397"}
{"id": "train_027061", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "8c5b1bc2fbcdfc51dcea89d0818423d4bd89775c"}
{"id": "train_027062", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "d87926c7c2f09a9f920e9a0a443d5906bfdfaa29"}
{"id": "train_027063", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["coding", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "6c134b4f3f5d52ca6165193846cf179107216af4"}
{"id": "train_027064", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "ae32c63620d5976494abbf6519427f3f8ce8ef48"}
{"id": "train_027065", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "06ec99375af12dca663a74182b2dfa731d7cbb4a"}
{"id": "train_027066", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ee5977d2af7f8be2dedf5e44500d0b1ed49e65d0"}
{"id": "train_027067", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "2d4770679b800bc593313b5ed7f28d20aa492228"}
{"id": "train_027068", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "16e2614d53ac4707a4ce46a92ebefc665e239f1b"}
{"id": "train_027069", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c3eebf54a00725e8cfc14e0cfa21d42c0ae5347c"}
{"id": "train_027070", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "28d566a3a925b295e0b27f35c4e3adb8f43b18e1"}
{"id": "train_027071", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c4dfbe86124ee357d51a684c77d4ef4d35d70171"}
{"id": "train_027072", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "6a3c280e74bfb1f4c8a14dde4cceca5b6eb8e920"}
{"id": "train_027073", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "b8063988eb80e87be010bf5f897a754fb23b181e"}
{"id": "train_027074", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "3dd90c434186b25af76bcd4f072742f291d3f7b9"}
{"id": "train_027075", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "a8304293f0689e52e8ef15efb1754124be992c17"}
{"id": "train_027076", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "9018a58bea19bbad216caa5a7ffdf837d05ea6d1"}
{"id": "train_027077", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "e42d687bffefb200c01de83abbba7c976ec22778"}
{"id": "train_027078", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "e9dd259f5df1ad71956b7418facb4e1ead214433"}
{"id": "train_027079", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "81ab849d1e608a509c358a854a8c7aa78c8bdd18"}
{"id": "train_027080", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "1f834022f9be9e2c38dbab743050dbf43c7f58ec"}
{"id": "train_027081", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "4c9d566bdb36dc307085dc481f0a7554127b628a"}
{"id": "train_027082", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "94ff321300cb725dc01b60ca1d1648400b0d7113"}
{"id": "train_027083", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "840e943b126ca894fd9305b78e812ead720074c5"}
{"id": "train_027084", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "37f77c4d2da895d11d438231269f962553b013af"}
{"id": "train_027085", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "8cdc1315fd94fa04a679cb34f45bb9f7df8f19ac"}
{"id": "train_027086", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2284b36d4bdb150273498300c74f73ea4539f71b"}
{"id": "train_027087", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "03f1c1c83ef93c4e2bf3bf1e0207e388513e2edc"}
{"id": "train_027088", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["governance", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "fd20a36e827f04c70dde0caa63c4f77e4a237392"}
{"id": "train_027089", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "dcbdd2f8d1a28f874c3f281d7af4869e9bb5005a"}
{"id": "train_027090", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "813b53ffb3f5b8027ffec87ae9089bd57ab4f615"}
{"id": "train_027091", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "705e8c5dab71f7fb4ee57a46492accd411cdfc1c"}
{"id": "train_027092", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "e3b51b21d9970a0d9ac0f11d68d74019606fc920"}
{"id": "train_027093", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "32dfac31fed7b1d63a1c0a52ba6df465e10d74db"}
{"id": "train_027094", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "b7bddb7e983462fda6261c718c726c77d0d6848e"}
{"id": "train_027095", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "819e4932fca50f69c8bfa3cb0d03438cfb45fc07"}
{"id": "train_027096", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c3065e440fc7d39e641f6f7c0bf6a5ad1171669f"}
{"id": "train_027097", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "9f9d9e0248b63441ef5af60f7292ec3842fbb0ed"}
{"id": "train_027098", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f9b3c8c948782f29d9d75e685fdfb3b4f50d64f5"}
{"id": "train_027099", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "64264f4d3dddb49a4ab71d62305a831325df3031"}
{"id": "train_027100", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["security", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "d6a64687c2394691b4225ad7e85a474fa8e66374"}
{"id": "train_027101", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "2187f13fc6852cb14d4a4bcfb011d103bcf0b49c"}
{"id": "train_027102", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "cd6be4723b4d9926eb5db3ee1bb0a75a799cb66e"}
{"id": "train_027103", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "0c2568bf2e5544605ba41aea1c280c38febd170c"}
{"id": "train_027104", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "bdd17d9e97ff1a4bb469e78dbc876482f4351028"}
{"id": "train_027105", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "3909eda411a808a15b316b322daa6c6709feae4f"}
{"id": "train_027106", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f3f029f1d487cfb8842bc7d4dca60d2f40fc6efb"}
{"id": "train_027107", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "373fa3c071a8c5f1167c03bae835186e42b3be09"}
{"id": "train_027108", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["security", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "8d6661bff4a8bf46ba54e6ec976bed4eb7e908fc"}
{"id": "train_027109", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "4b6f513394bc4acd02125949a6e63453ebb4620d"}
{"id": "train_027110", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "55da5c8331ff707b3112dd117a4d3bd0d95a52f3"}
{"id": "train_027111", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ecfc0ea36965999ad6ef502f3b900a0397cab290"}
{"id": "train_027112", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "2af7ef7549e1eb532dbe6d9843b700232b7fc3d4"}
{"id": "train_027113", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "0b0b750b22a0bd5340bf093bf9e80cc409a9b36b"}
{"id": "train_027114", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "816ecf468d299bf6fb2876d055ed465b66809f1a"}
{"id": "train_027115", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "478f372f4e6edefa72ede1565c81d6664de17a75"}
{"id": "train_027116", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["data", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e01a9c2fa42c027bc6b886a1088e7c2344586a78"}
{"id": "train_027117", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "b0207863162fce570dada3ffe71e66984da9937d"}
{"id": "train_027118", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "1316cf7fe2ef7ed8ae1d5d93f1a3e9038d9e85b1"}
{"id": "train_027119", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "c7041f01c74f1d6fd6f260d16be32b5b4f1a0eb8"}
{"id": "train_027120", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "627b6aed36ff72007c36172ae199eeeee1d8e7a8"}
{"id": "train_027121", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "8a914c85113bf4be34d27acfc4970eafe05e0df9"}
{"id": "train_027122", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "f9e2f3d184f2ef1b0c7f38dfbac3fc8d5fe2c232"}
{"id": "train_027123", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "5b68c16d24858fa5d4a95aaa85926fc0d363de6d"}
{"id": "train_027124", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "ddd759fcd5207081ffc4f05798c17b54ddfd5e0d"}
{"id": "train_027125", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "c3575b4d6b302f14cdfff28f4987e99c9bb564e7"}
{"id": "train_027126", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "92677ead46be9f14ba52e56f68e990a6d59662af"}
{"id": "train_027127", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "06538d880d55a63d451fbcb96fb16b755f9d7161"}
{"id": "train_027128", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "5594c63b20d042d4d6e976e3b633c9c16d817601"}
{"id": "train_027129", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "b9449ae0ddbc54981932010a5e13b1de95460359"}
{"id": "train_027130", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "3b68b41d5c15b0a13b248c9d8eabbd52383ea6a0"}
{"id": "train_027131", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "6ad51610e8019d069ccc33357b1360dbfb498887"}
{"id": "train_027132", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "3516c4e8a0f65744ef8e53ada62e372a743ac86e"}
{"id": "train_027133", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "f93eaa471a7e4b0692b1830c206192b339ea08eb"}
{"id": "train_027134", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["evaluation", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "eb6e77e863d3ff5da0a3d0f34d9553a53c466435"}
{"id": "train_027135", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "b4d8eb9d4d992760809e156c2f815ed7c139bda7"}
{"id": "train_027136", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "bb1f05589ed058ae60dfa95712ff07c39763264d"}
{"id": "train_027137", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "138508d159569c103261bee4c1b6c64cec39aae9"}
{"id": "train_027138", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "744ca5b8daf9abedef33be91c6e1c2b5f0c688ae"}
{"id": "train_027139", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "75084af31552f3a8e12cca27f803fcf6a23422b0"}
{"id": "train_027140", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "e4754167016a7a00431316b590c802ded08d1790"}
{"id": "train_027141", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "63d83a0f8698b59a127300966b588fe929fb31bc"}
{"id": "train_027142", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "ae533309dc1bbed3658f0c06c66902619bf0bca3"}
{"id": "train_027143", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2e9a85542eb488e877d48ae104dac626bcad4aac"}
{"id": "train_027144", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2744499457e21f5d3666b4569333ce9226dd2b25"}
{"id": "train_027145", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ffa6f9ba3819d99b7a707b743d20967af7c0e773"}
{"id": "train_027146", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "90cbc8fc120db272ae5eca4838c2586affe0f91a"}
{"id": "train_027147", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "20ff3b94093cde11db94fdd3d088cf9899fef300"}
{"id": "train_027148", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6b3cc378fdcec6ddbd8d2a2d20ba7e52cd9f6254"}
{"id": "train_027149", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "0e8e7659cc9ce219e60db5bd0c7f9bae24f78c0d"}
{"id": "train_027150", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "205e407254a8a5975357f089d04b90e6156d9ff1"}
{"id": "train_027151", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["performance", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ac4a67329e64ec29087ae7d89ac7d138339895c9"}
{"id": "train_027152", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "b1d3404c5bc2a9c9780a4d413d7eb524eb8c4f38"}
{"id": "train_027153", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "f40b3ef4e1ef698e19cba60afde1aa261dd411d1"}
{"id": "train_027154", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "db5e9ba8f950e6ede050ca5e3ecc76842e690572"}
{"id": "train_027155", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "bfb58c3ed101633ab0d908c847cb9ff03de019e9"}
{"id": "train_027156", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "14493869096ddc2694f928fd815e211756724b04"}
{"id": "train_027157", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a7f21763f3f8fdae68f5cc1d1ace3c0fe0acfca6"}
{"id": "train_027158", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "6ecff23d29291126ad8f90df5035b29fca07e247"}
{"id": "train_027159", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "61c2fe71b4f2367ff1cb3a988167f484d46acb88"}
{"id": "train_027160", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9c23957daea291a5351da27c2e7a1e66efc81bd5"}
{"id": "train_027161", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "cb725a3f9203a7ec9766e512fe6a239da55e5100"}
{"id": "train_027162", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "839b2b211adc2c1c9088a047882ac7b9739e13ab"}
{"id": "train_027163", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "e8dd848ed1e8418219c2f88b810875912e3d219e"}
{"id": "train_027164", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0c8f3eae12fdb19967570eb86bb04a112d994c82"}
{"id": "train_027165", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "8433c91b98bef5db47c2d67cc5dbe23aa42e404a"}
{"id": "train_027166", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "2c668f993ffb7afadee20327f785b34607a3fc58"}
{"id": "train_027167", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "a6f9446156a731b951e6bb37557bc6fe867799c1"}
{"id": "train_027168", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "5e49d87e220382d5dbd987f785b9c410629cd0d2"}
{"id": "train_027169", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "bbbc1241dacdeffc06b77fc010200163d9fe16a0"}
{"id": "train_027170", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "e259007daaeff209655608019e1727b7ceca2de8"}
{"id": "train_027171", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "529b8c9e53d3941dd28849856141a3c5d31cd724"}
{"id": "train_027172", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "54cb00c5719865dff8dd3e1820973c1bb3a2eb6d"}
{"id": "train_027173", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["performance", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "54f2bdccdf3e8188a3acef6fdd28fa636a967bf3"}
{"id": "train_027174", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ac6453bd6eac93869cdcf7b4d09cbe8082c168c1"}
{"id": "train_027175", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c4aa657b8e86e122c59cdb56385f3e8fa5a00752"}
{"id": "train_027176", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "9b2c8945e50250e91810116a85b7d20ee63aa7a4"}
{"id": "train_027177", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "add5a4949434063dc3485bc01975e3508005f98a"}
{"id": "train_027178", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "56603b93c2ccb4e2a012f8fb117020de3a6be041"}
{"id": "train_027179", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "5a3262f9f2ee4132a2ec312ecb25e3cb9e3535b0"}
{"id": "train_027180", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "1e668dfa99e36b274b5c1d839aa1a03ccf475ef4"}
{"id": "train_027181", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "fe9f0cb6e697d5a8c4d01cb7e745229f209b0c06"}
{"id": "train_027182", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "b27123724a5339c4f9363f6fc59b5747deaf5433"}
{"id": "train_027183", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "cf55e5d1b3d62ff4acd6c655caf1b7a020c21e9a"}
{"id": "train_027184", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "b0745b6de4e250beb6c060c21b3d4eced0429214"}
{"id": "train_027185", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "5ccd422bb2f96439c9283cab3378e43c2bccaf2e"}
{"id": "train_027186", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8374e7ff2d63e156023fb7e47b282ef9da0a5e94"}
{"id": "train_027187", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "b3e37a7fd8f2840a691664c3ee3283e5be52926e"}
{"id": "train_027188", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "eea2295f86c5c8b8beafbc32d16d02c6fe0bf1f7"}
{"id": "train_027189", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "63d08f372161c1bb94b4af29ac766698349337d3"}
{"id": "train_027190", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "9f855d3f2165952258b989584036cb7b135114bb"}
{"id": "train_027191", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "5f924332e195a492fa5f78de69d2451a15e93a58"}
{"id": "train_027192", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "48f64acfc1b54cc0dada324b6ed40e2b29fdc0ca"}
{"id": "train_027193", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "9c567b5e7942d76fd0960765a53f9a2b2f749051"}
{"id": "train_027194", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ecacf182df2110c12c91280bff67bc53d5238a78"}
{"id": "train_027195", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "826f68390981e9644621dfa643e7d8f30d23af38"}
{"id": "train_027196", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "462184157ae2fefa9a2c74d6141e8496d629e4f2"}
{"id": "train_027197", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "62c34b435968c6a6f91975b2dc83eb16534bd388"}
{"id": "train_027198", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "060cc9c98899bf6a3f4975958b2858c82be27316"}
{"id": "train_027199", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "1689a552b314ac86bb72b5c7b2bfba1fc0c7ac89"}
{"id": "train_027200", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "424e37c8a4c0ef131d37e6f5c3b548cef8b087fd"}
{"id": "train_027201", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "e1cebb42e609ace1a841276d8c690c33e3f87bea"}
{"id": "train_027202", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["coding", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8eb61400801cc43ee6fd8bb782a2c0d4d6174422"}
{"id": "train_027203", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "dfbacb060e1d92732517711cef53dcecad82eedb"}
{"id": "train_027204", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "6925d6167c8f805aa13bc9ce358f4436562d42ba"}
{"id": "train_027205", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "82e26b07cc04b0c2f67bbcf486a2bcd477bd90f7"}
{"id": "train_027206", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "8568ffa69749b6adaa2490970d233c80566eb883"}
{"id": "train_027207", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "c8cd5d52f227c30d79b711e80c4bd8ea5a9fd8cb"}
{"id": "train_027208", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "0b599a330c5dd25511b2e676119c5e0b33e8b5d7"}
{"id": "train_027209", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["research", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "0e3619b537a851d0cd7ca58347151f3d89fb32ea"}
{"id": "train_027210", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["coding", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "2ff5312f4c95332ccff4a35c01d793e45d4a9fc3"}
{"id": "train_027211", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "1ee4a1cb399af302acca7deda09700d5e73d39e3"}
{"id": "train_027212", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "b7ef55b35b51195598c9308d4a5ad9dab9d44976"}
{"id": "train_027213", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "e2430f05c3b2712158dda100db24743c0b72ece2"}
{"id": "train_027214", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "1cc71091134498597c0104778ee74c5256f72e86"}
{"id": "train_027215", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "3393a949294a2f3894fb02be49824ba1163328da"}
{"id": "train_027216", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "e09ebe4f94c453d3e2b02493a9a7c3cd18df8aa5"}
{"id": "train_027217", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "daa601f7aee3258bca0ccd531ec1abb9a63f8a37"}
{"id": "train_027218", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f1902a4fe5738b13fed8b0b86704248af9ce007b"}
{"id": "train_027219", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2e376b493a22c1486a74ab5ab4cd1ed752642aa2"}
{"id": "train_027220", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["governance", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "d275ed68e2b868e59a19533495b5bafa9abb0ff7"}
{"id": "train_027221", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f7f16d436beeb3f51dfb96b425bd0e1ac5685515"}
{"id": "train_027222", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "f6355e42e40f25185bb943a784f26c5dd8544ec1"}
{"id": "train_027223", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "1972789fac2bac1c1852e7c2e561eccaf84b8b06"}
{"id": "train_027224", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "4e341022d381dea5ec1e219e390459b39e02ebb9"}
{"id": "train_027225", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "1c445476a9a3589ec5c98906a909c4475c6e62e1"}
{"id": "train_027226", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "1adae247b6be3d151185333af57f81fde046548d"}
{"id": "train_027227", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "940d788056fdb513c5e4d501d877bc6f99dff983"}
{"id": "train_027228", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "d71deaaa62d71b4144410ea580e5cd5c1c61b90b"}
{"id": "train_027229", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a9989c6855eebaaa4163f22ffcdcab20ff29106c"}
{"id": "train_027230", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "6db9f80ff868a7ab498724ea6b3bdeb95bc01693"}
{"id": "train_027231", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "330d8be50cd12228596a54431c46372009fbfa5d"}
{"id": "train_027232", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "b3010c7e79661f1442f378ce42fb51253845edb6"}
{"id": "train_027233", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "bf03c7d00a0da08d6e7317ccb1419934b49222c1"}
{"id": "train_027234", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "c49b5f938762bade659b12969d4f058cccd8fe59"}
{"id": "train_027235", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "d769c1e13198c3173c6e2c4d8e5d218382faaf9c"}
{"id": "train_027236", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "100a5cede9532ed8925a2140ad7f6dc59f08003b"}
{"id": "train_027237", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "540b759600e63cdad1457c254546d40a9633b979"}
{"id": "train_027238", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "6fe7ee957a7ee564022a04d3099c1ef15dccb3da"}
{"id": "train_027239", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "717dab252f8cae30f68cd9fa3f6f4b50be66b87d"}
{"id": "train_027240", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "f8d226887e88aa48315ed64a82c03863ecc25ffd"}
{"id": "train_027241", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f7c731e37880a7172ba486a969dccab1bd258967"}
{"id": "train_027242", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "8d5523318ca3b8f4e6a3d04bd627045720185fb8"}
{"id": "train_027243", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "26f4b436a972a6a37d2026cd40addc4e42c967e4"}
{"id": "train_027244", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "7b622aac026be1f5b8bb4081ac71d78087b8245a"}
{"id": "train_027245", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "cacea8cb7f073d082c0c6e4b9c12cec445cdd202"}
{"id": "train_027246", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "a4a8a93ee4faf8e4e939013d6fd0d581a49dd5a4"}
{"id": "train_027247", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "9f51547a54c300e78f3576364e88cd5f297b22e5"}
{"id": "train_027248", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9dbe0f691c40a27314f8ee489a4c62a36f04b91e"}
{"id": "train_027249", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "807bc5165677cee765119da2ce449678f1e727e4"}
{"id": "train_027250", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "0b0aa11927d62a3cea7aefeb31eb4fd888cd28c8"}
{"id": "train_027251", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c932fd7a65f0f0fb2aa81d2d434dd5469a3fa8ff"}
{"id": "train_027252", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "28652ad509709a726ad26ef8850e194ae67b8f67"}
{"id": "train_027253", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "d4c37aa62de929224798e50716009712f4ba1227"}
{"id": "train_027254", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "68a47d804eba2e57b2ff8ba7728ab9ab541be2b5"}
{"id": "train_027255", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "4eadef5f2880ae840c7b01910d39d17bec3e551f"}
{"id": "train_027256", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["performance", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "3cffa656bbc8560ad00b210a6f8f8a5a87ac472b"}
{"id": "train_027257", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["data", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c4c3d548a1bbd2bca28f4226d5ed279bf4c8dcbb"}
{"id": "train_027258", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "51b2220a0e38e388e5ac7c37d9bbb9ebc959709a"}
{"id": "train_027259", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "d40830cc3ab34d9256c5792164e20ac71bcfc0f8"}
{"id": "train_027260", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "237ec746f62869cafd6826c4c151c450cb05a498"}
{"id": "train_027261", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "1bbeaf73337885b85907d62fef6d958fa11ed08a"}
{"id": "train_027262", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e343ca677344997a5554018c81cc6b280596fdd8"}
{"id": "train_027263", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["coding", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "fb8c76db5b6f857bc346c29a3d37f0abdcb7334b"}
{"id": "train_027264", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a81bca81c68181875913c69012427f0a12e37824"}
{"id": "train_027265", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "b1e2d7c2361adb53c5803a68ca3031a47fbace0d"}
{"id": "train_027266", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "83fa77442c3f222bbcbfb758953db0232b1eef3d"}
{"id": "train_027267", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "39e0aa0752d6eaa4b4aad49d13aba2c48874b91a"}
{"id": "train_027268", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "2d7b1d4adecd257145a174a3c6524111a4d76c9c"}
{"id": "train_027269", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "956cbe4b69f10b0ac593ba420056c7491a706906"}
{"id": "train_027270", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "86be5bae17af6e78a7b6214941269bdd2782c4a4"}
{"id": "train_027271", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "017a6e353608c9cd045a672016c9d35a48d0c029"}
{"id": "train_027272", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["governance", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "4a52f897e6c3093e4b432ec2450b774f18f4742c"}
{"id": "train_027273", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "516da9d4dfa3767f9e650921a97a31d465e3a94a"}
{"id": "train_027274", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "1c7fe43dd2407a84b191eae91d0d826985482f22"}
{"id": "train_027275", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "97d9dca1088e97a7dcc7714de3f3048a0300f0d5"}
{"id": "train_027276", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "7c1161c7f59118faa01a5e5957fbe2bc88939640"}
{"id": "train_027277", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "dad9765a84dea011fe271d99af64c6b8535a47a9"}
{"id": "train_027278", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "757e6bd621b77b55e1ae9b683def78fb8b904514"}
{"id": "train_027279", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "5a2871df6528728d0c386ea25cbe727b1d0022b4"}
{"id": "train_027280", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "9cb2754210689751d8cc4de2c8ba4cc3941820c8"}
{"id": "train_027281", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "6f60bb6074a9d4e919ff53cd216b3be6a2f9726a"}
{"id": "train_027282", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "83e7e6491137f9138d380e74ac272ef6fde5ab0f"}
{"id": "train_027283", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["governance", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "28e5f0deaac5bd957e55b94296a4587a5570cf12"}
{"id": "train_027284", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "499854534741aac12a431998923f35e7ccd54012"}
{"id": "train_027285", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c46ba4dd44ba3281606c76ea9afb9ed60dd868ba"}
{"id": "train_027286", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c412f667662cc352673dc0752be6d85b4a54a2a9"}
{"id": "train_027287", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "6baeaa0725de6ec29b70fc4674a5e191c908a60e"}
{"id": "train_027288", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "4d9edea0bdd51f845e1d711913f641fa9b57ea2e"}
{"id": "train_027289", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "4ae776ad8e5b652da92f0b9ff69ecf961d4d92fd"}
{"id": "train_027290", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "49eadba1642e8f85aa2a6fd213cd4b0ec25c8e80"}
{"id": "train_027291", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "64ace31d7db26e46d34b6f3fac741dac07689ebf"}
{"id": "train_027292", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "1f302e82f031ffdb9c14bc7a0c53d5cf83de61c4"}
{"id": "train_027293", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8b1a29b8cb0f51e0d2292e50d541a76cbdbd8e46"}
{"id": "train_027294", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "4c75e54212bf9ff7ed670864925f54e2a272bd6e"}
{"id": "train_027295", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "d6983fc40d31d260fe9d77bd5ca6b1721e1caa68"}
{"id": "train_027296", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "b102790df21105a5e3de024495945f1771c1eff8"}
{"id": "train_027297", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "fc655e590701e44d30e92eaa8047eaa46bdd92ed"}
{"id": "train_027298", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "0ffb424c4f21ca16158ce0198466aca834206f5c"}
{"id": "train_027299", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "a48ace88ed6edc1940b457d9a7bf2ec689ef79b1"}
{"id": "train_027300", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "182a7834bead98b6bb51a41749e202e7f992d082"}
{"id": "train_027301", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "495bcce857f839c382865cefafbaa4f9b148f212"}
{"id": "train_027302", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "a467f9405f747aa90cfe19a942e3fb46135e4530"}
{"id": "train_027303", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "7089d9d2abdd75576330d3e9b162fae7e12aeee2"}
{"id": "train_027304", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "88c82611d624a22ed47a5b94a810c1be61538f34"}
{"id": "train_027305", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "25ca219fd813a5658d5c61d2afea3f0a092e4c22"}
{"id": "train_027306", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "14be10890e09b5e81cf79c99447878a1a0b9631b"}
{"id": "train_027307", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f1ae743d374f8df4f2efb517701d8adec67248ed"}
{"id": "train_027308", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "d9f3cea41eba3e92a9cc3f406391828e8ee57051"}
{"id": "train_027309", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f01a99cc412cbd83ce166b1f2e19b38fc5cd4b7c"}
{"id": "train_027310", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "66812bf0ae0dd8a24629595049c3bcadc565db2e"}
{"id": "train_027311", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "3d466372c67a8210216d71ee576239c925ddc60c"}
{"id": "train_027312", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "fda43ba2aef596e19ac241132fd3ac1e33e6da49"}
{"id": "train_027313", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "d61010e6c71899dca870be814e7900e3b666a780"}
{"id": "train_027314", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ce2c8bc662081d5320b60a0b53799c4910f09cf7"}
{"id": "train_027315", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "810dc497ed0809eaecdc07de3ac749c1a4863c21"}
{"id": "train_027316", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "368b5ee330914aa22a200a7a49ed42e0e4468b64"}
{"id": "train_027317", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "384cb83d6350a86fde6a5eff93d85dff8387f76d"}
{"id": "train_027318", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "06fabb38ef8a49bd6e82253cc35381622a596262"}
{"id": "train_027319", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "1f0853ec5796e20dd49c336bba59588eb4bdac02"}
{"id": "train_027320", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "1db65ef5ead2b7eab92ea978afb73ca8349c2d44"}
{"id": "train_027321", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "014bd45d94ad4188af888b44d8548d62906ff61c"}
{"id": "train_027322", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9ee08b3c4c64ecef985eaf02947457cf9823e6ee"}
{"id": "train_027323", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "053db95b72ecc41090c299ffc493489348983a4c"}
{"id": "train_027324", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "153deef42176e5066a8a152fdeb1b89de712697f"}
{"id": "train_027325", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "38c8fa7f6e865ae5ae4b414cfa7504071617d163"}
{"id": "train_027326", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "63cd31b936571754b15af485b618c6a5c879654d"}
{"id": "train_027327", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "b4f191ceb1d35fe0f3ef87902b7ec178f9d7a667"}
{"id": "train_027328", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "32d551228b9589dd60086f1aebf9755669cc9587"}
{"id": "train_027329", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "18be66614870ed50ae3567c0d54f71f287a650ee"}
{"id": "train_027330", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c6cf9a3f3bea5b0532a6dcd1f9f69cd67af92c2b"}
{"id": "train_027331", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "3d12dc83c50c2e3363e806b7f6abca3598e316e9"}
{"id": "train_027332", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "5f602875da7bfcb4b4413b5f2887b7f3ef41aef3"}
{"id": "train_027333", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c6f825994b63d3cca65101aeeb4243338130327b"}
{"id": "train_027334", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "02595cebaedeecef593ac4afe0eae3a598542b99"}
{"id": "train_027335", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "e619f1a65303b4900e364b5f09182dc2c1eb9ee7"}
{"id": "train_027336", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "f2a1c1c7da30344dec20ee847800e2d727e041d6"}
{"id": "train_027337", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "f29fe2eeafde04a37f274cd193bc15392aec420f"}
{"id": "train_027338", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "84b4744bbde730167997ccb39267a2f50c76a006"}
{"id": "train_027339", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "de0ae333860f4edcb615f0a077ff2a5d71cde718"}
{"id": "train_027340", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "08b2f3af640ecff2aa32b6b50aa00b307ddd7ba2"}
{"id": "train_027341", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "1e3e57c272156fb163e6f28e8bae65b4117a96ce"}
{"id": "train_027342", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "39c2c5d3595e05e8c1387e70791ee796c1258df0"}
{"id": "train_027343", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "4068a1dd4a6e7f294afa29f2c1be297f8832bd27"}
{"id": "train_027344", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "705b6acab13a6f661ecbc1857faf0c11edefc517"}
{"id": "train_027345", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "15d8b612644badfc9f7f6e22d1cf0494b42345ee"}
{"id": "train_027346", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "448cb2144506deb3bc90e519b43fa08ee527a041"}
{"id": "train_027347", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "5f0240e205520679ef270f4d4af34e9c74742930"}
{"id": "train_027348", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "9385d035f324a080a8f75345afc4b09b5dc0d0b7"}
{"id": "train_027349", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "76b7324acdd2a4b83a8b89c0fcce11f8b0cc415e"}
{"id": "train_027350", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "ac29ef0f62b5f99f8103e82501719fc254381def"}
{"id": "train_027351", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "eeb1891fd74cf644d052875aee999eb68024fa60"}
{"id": "train_027352", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "1a72b00c501f18071727356907810c9e7e7ba300"}
{"id": "train_027353", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "ad6d0e67e44e578fbfc674b59611bcbd02db4f76"}
{"id": "train_027354", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c9d6dd78e2500010a6871c2da8adbf3cdca2908c"}
{"id": "train_027355", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["evaluation", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "0e84181be71c092d53aab6cf5befa495767bc33d"}
{"id": "train_027356", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "5834a0ce4b38a09439196d8cc0e5f66b33f6900e"}
{"id": "train_027357", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "461e3b03d90d8e6b75d428eaf84aa4f22b048bac"}
{"id": "train_027358", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "dd50c5245af94ba7a2d6e0ea62ae58137d1bc12a"}
{"id": "train_027359", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "e837cf15045d08d1476bf28cb9d322d32e33d1df"}
{"id": "train_027360", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "63ecbf5c2d6e9d0b86cccc104b2a18fe91c950a6"}
{"id": "train_027361", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["governance", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "38a1d0503aded645e94c222b848f42fc11d88efb"}
{"id": "train_027362", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "bbc5fbe8364290d8ffb45ac598172bcc82fd6e84"}
{"id": "train_027363", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "5f841cf58f591ed88247b70fdcddb68f220b22f2"}
{"id": "train_027364", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e2bc21b61367e87e24c64df0ce80b8ab165ff2b3"}
{"id": "train_027365", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c15d22d2eb0a4bbe877f0be52db6e4b5c944f6f3"}
{"id": "train_027366", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "078c354ca08915154a948d5d5a7123ecd8c2f717"}
{"id": "train_027367", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ffc98aee4c97804b014101e24bc3d90805d89e0d"}
{"id": "train_027368", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "e22d0ee9ec118dbabfc1881aa2e9a5550aa7b791"}
{"id": "train_027369", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "da3dbfa184fca3f8f914df0da0dfbbe7c3e3318c"}
{"id": "train_027370", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "955c0969cb623896eca5917ed31015d8663c1877"}
{"id": "train_027371", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "fc264298ab5466508c8f2a648474b06df4f12b9a"}
{"id": "train_027372", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "623ee77b8fd9410604a9fa859cdd6068c56bd61d"}
{"id": "train_027373", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "29efc11ddb1920461ebf441729623591cb574b87"}
{"id": "train_027374", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "72be3472126f9107b8b0a1d0df51bf07501f2ac6"}
{"id": "train_027375", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "7092f91d9ce01453f0b16c311e6f36093442ed7b"}
{"id": "train_027376", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "b0816926d3a12abb764f63e4134e41ea32f170d3"}
{"id": "train_027377", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "d69325c12bbba479d0a1e2b25497cb06fb4af246"}
{"id": "train_027378", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "5aa272da6c06ce0bf9562041bdcdae44661424fa"}
{"id": "train_027379", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "2ba3baa114f1788261f950dca4c271ab47d37a12"}
{"id": "train_027380", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "e7cc43783efc1bdc198d0232fa27ba8ab61fbf92"}
{"id": "train_027381", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "a34e9d0924237472a7dc7a22a47f07e9067404a8"}
{"id": "train_027382", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "9a6994c144e7077a25d1548d8964bef2c05386d9"}
{"id": "train_027383", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "3b20f09d7c8cf84fa6f0ff3129e3f3b08be3d5a5"}
{"id": "train_027384", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "164e205f8b4cc0ad640be202c249839bfcffe59f"}
{"id": "train_027385", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "d349bbee5229bde5ec15600c33f69723c39fac6b"}
{"id": "train_027386", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "29b5c6d9f8de37d1b9eda439e7a795328c0c29b3"}
{"id": "train_027387", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "eda0b3b7196989557bb87bb55c5413919173aacb"}
{"id": "train_027388", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "919f30984f27cdbe9e0e96c26f881fe29539a24b"}
{"id": "train_027389", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["data", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "dcfa688345e3f9fc5ac2df732f0131ea8a344225"}
{"id": "train_027390", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["evaluation", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "20cb57e2efd0c273d88e3f63c6da48b30f4933a1"}
{"id": "train_027391", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "ed1e491c4172f7fbe9ffae9ad43972ecd0d33ca9"}
{"id": "train_027392", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "065b85a51c37d2d2cd083ca34ca8f3c287324d1f"}
{"id": "train_027393", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "0904d0dd93e05c6e9ea3b330c092892efff45c95"}
{"id": "train_027394", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f4cade7d8c8e935fa4cd3712ea8b5a6270c81b58"}
{"id": "train_027395", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "957f0f161d42f2c28637062b52cefbff440d4c68"}
{"id": "train_027396", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "a57ec6cca6327a651182a4c967d6adabcbdfa841"}
{"id": "train_027397", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "2932f612a5a364b603ff87b1be3ba7975d7eb5c7"}
{"id": "train_027398", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "13e390e8ef3f57d405a49687311962b3cee6cc35"}
{"id": "train_027399", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "4e4ebb69d2646a383dcdd998db5e9717061454af"}
{"id": "train_027400", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "59251769ecac3e675a043a2db1f054bfa9257437"}
{"id": "train_027401", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ae259190a5cee6cb7c0be08fb8a5b6378268681f"}
{"id": "train_027402", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["agent", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "624a45b07227ba6753988ddcc50d18204c883492"}
{"id": "train_027403", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "53f22e47271dce66d7cc697705d1adfc8f1b90b8"}
{"id": "train_027404", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "df116e20f019c82807e524f76687cb92d1062b17"}
{"id": "train_027405", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e259007daaeff209655608019e1727b7ceca2de8"}
{"id": "train_027406", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "515d8a54b264eeaa4d8bac52dcf8ec7a9a29c6af"}
{"id": "train_027407", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "e2d925aa0fbcefd9702330d8d9ed064ee50f71d0"}
{"id": "train_027408", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "62f3c279ed80a4235fc378cdc2508c84c3ec518d"}
{"id": "train_027409", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "1a38b76d091c119dbc75408f7abd99850b52dec1"}
{"id": "train_027410", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "692c4f0b6b506603eba711416d5f20c9b276c164"}
{"id": "train_027411", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "3113d90880d21c88698006f5162adc59a30c8e73"}
{"id": "train_027412", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "10db07ed34d2d10bd9030fa50cfe9a5f5704f5e0"}
{"id": "train_027413", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "1664a0a99e8ddd640fa82ae8c55d16b42ffa5a19"}
{"id": "train_027414", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8685b31a4a9fa114200de1a89276ecff9be844a7"}
{"id": "train_027415", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["performance", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "e7b7526900f5696eeecc0a3dfe13e28272204e8b"}
{"id": "train_027416", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "cd4b57f670160adffa95d76f2d1a97f63c172790"}
{"id": "train_027417", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "76b26a2938ee9f394c7a89500838ac27a453c766"}
{"id": "train_027418", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "dda6e81e669c26180c85b0f2204fc2362e87b938"}
{"id": "train_027419", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["evaluation", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "b440ffae5d5009e6eb779a5a135558f2435b47cf"}
{"id": "train_027420", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "c4639f0eab90e303cd908f95b975a8f1744f0593"}
{"id": "train_027421", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2bf8fecb80e7d08dcdef2e7221d66fbde76087be"}
{"id": "train_027422", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "e459e62b1964d2b4c6e9d180836d6f1f6bbe4569"}
{"id": "train_027423", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "b80147833b52369a378f3d7373b2a3f7586d8948"}
{"id": "train_027424", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "02cb638592cf7203f74d0b9271e01df8216a7330"}
{"id": "train_027425", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2f6e578b42b5c6f29c15ac5cf692fa5549dac08b"}
{"id": "train_027426", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "49c34b019122118cb804ce23a2bb1f6109104e12"}
{"id": "train_027427", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "fb7cfbd18173f36c57d7b9c6512722a55a57af31"}
{"id": "train_027428", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "9d07344b3c43dad9d9921eaf2d07171a6afc01ba"}
{"id": "train_027429", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e299ec84837330835f1af5ba3e69a070fb2ea6c3"}
{"id": "train_027430", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "7a3115815b4e4bd72f2da3188df116b00c0f3503"}
{"id": "train_027431", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "525efc0402e506421aa8e8c9284780935ba7c46d"}
{"id": "train_027432", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "5aedcb7b68f7bed05e1087438de5dee47a9e800e"}
{"id": "train_027433", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "01d898efb719c5b71aad10f8a3993f6ec16021fe"}
{"id": "train_027434", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "499673cf76040af3879cc8387f765a0c9253807f"}
{"id": "train_027435", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "c4fb961629c8eab2353757d47b43a67549cf7fb3"}
{"id": "train_027436", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "e9fbbf3dbf44aa1af2ef4d160759ab15bd42292a"}
{"id": "train_027437", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "b3a36af0bdfa325dc74b2646403d04a4fd9f61ac"}
{"id": "train_027438", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "40464f3c203a60753a280e15501258d40b98faf5"}
{"id": "train_027439", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "9ea041343c0f97280391f7e35dace0166f220456"}
{"id": "train_027440", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "f2d900ce1c515ddad5e4667add507dae8bbf4f23"}
{"id": "train_027441", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "57d5c02c0fbc19915dd90f0e17d570446b95249b"}
{"id": "train_027442", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "77a4c791b629cb23ee06a6227b23a1bb533bcd7f"}
{"id": "train_027443", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "8bc5868e376c6003bd3bf1b84b7d7bed1aa9c988"}
{"id": "train_027444", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ed13eeee6b07411e5e195c156a407f431dd033fe"}
{"id": "train_027445", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["performance", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "54e5dc87406c0b07ffe51299ac58e62178817e66"}
{"id": "train_027446", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["agent", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "9db6921c764b0c82a2830aef1c5302155dfaa700"}
{"id": "train_027447", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6c88bcbb846ba50779200231ae98f1b41d6ff450"}
{"id": "train_027448", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "744575831648e55e6d10e6e77d95b3b71c4e348a"}
{"id": "train_027449", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "91f42e55ed2696626b78206911a1ccb006fcc360"}
{"id": "train_027450", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f21bbed606fc8fad5da607da224a51787039ef75"}
{"id": "train_027451", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "ed10ddd8cec195bbad67aef0ac150b8cd1290dc2"}
{"id": "train_027452", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "dbe65a960a63f0856f84af6197a54bc5622f4ca2"}
{"id": "train_027453", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c3eebf54a00725e8cfc14e0cfa21d42c0ae5347c"}
{"id": "train_027454", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "7335312f5104e0945d5afd5e1d199172f9bd2cbf"}
{"id": "train_027455", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c85f6326d644e1cd631a5c1a6e180ad160a24729"}
{"id": "train_027456", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "47d3da19ce702c5c99552914e2ee51ca620f77a9"}
{"id": "train_027457", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "b2f0aa84d7591c1e07e6359361e80209b4820e99"}
{"id": "train_027458", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "1e1883369a89c849d6e9ecc18bb25537405472c4"}
{"id": "train_027459", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "05a298d01cd6dba662a277e89f0c45183c825853"}
{"id": "train_027460", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "540b290d5d9fc8ce317d17905f2e4db11d90eb38"}
{"id": "train_027461", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "d911644f5641877f59d5a5c1393fbb51a519d73d"}
{"id": "train_027462", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "cf40b19bf0caf7f8ae33d64bf2cc546ea4cc2e5e"}
{"id": "train_027463", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "319bce6183eb198233d06b8bb3e3d324dd4cf3c8"}
{"id": "train_027464", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "95209ac1c83a9fe04fb8186bdb816f3b84ff7cb2"}
{"id": "train_027465", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "40a28373bbeaa7c28e4ff65a42f4323c5efaa9b3"}
{"id": "train_027466", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "0493ad517c86b096548eed241df4c3524bf85442"}
{"id": "train_027467", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["security", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "8610b88a510e4d433af73a2f38d49160efb6c4b3"}
{"id": "train_027468", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "0955208172df2756ca593334a40ecb91d37d6a04"}
{"id": "train_027469", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "04d7c488eb0413115ca82cc3efc23380c70caaa8"}
{"id": "train_027470", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "4c5090e4e543aa0759226e269dde615158c13e87"}
{"id": "train_027471", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "85577f715fe9fa6d25d7ddbcb561d6c9f6a5b401"}
{"id": "train_027472", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "bc970c8afaa4f5285b768b28721f4b996ded4900"}
{"id": "train_027473", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "b7ef91aa1ca9591cdea78e42c6703d2971adc8ca"}
{"id": "train_027474", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "96963cc582edfce1bd82e05810d6c12b94e135ca"}
{"id": "train_027475", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "923c604ea8bf0528a43f3a16a597b1ffd7a42b49"}
{"id": "train_027476", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["agent", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "9a6994c144e7077a25d1548d8964bef2c05386d9"}
{"id": "train_027477", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "d6f046cdcc5cb605bed15d371f6c158b4f6ee22d"}
{"id": "train_027478", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2a7ba893b672cf8655d8976c1a626c7cdcf87e65"}
{"id": "train_027479", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "b35046027cb2275c4d681641020b85b1c7216140"}
{"id": "train_027480", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e908c6737c50b36b69511bc5a2aba243a89002c3"}
{"id": "train_027481", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "17dd8ebf2cbc82547759a8fd200f5c0eda5a2368"}
{"id": "train_027482", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "a9be2e25a1c22e35ef5ce123d22b386a14d14769"}
{"id": "train_027483", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "5626fa2fe5e15bea2692261ec4ede8b43f6cbdb8"}
{"id": "train_027484", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "be705a76348c3461ad58c9b69183f5078c6602b4"}
{"id": "train_027485", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["governance", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "2e2755c2b1958b8bb6fb317b73b4de5017ae3a11"}
{"id": "train_027486", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "759aa86a974fa2a7760536581c1ac053fe603dce"}
{"id": "train_027487", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "64264f4d3dddb49a4ab71d62305a831325df3031"}
{"id": "train_027488", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ab1fdaedddad235fa49f1d0d0e016adfc3be3c40"}
{"id": "train_027489", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "56fd7d76ca8af1467b1c8f662b5be061e8c794b9"}
{"id": "train_027490", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "aaf469ffc9128549eac1252d0919f382d42088bb"}
{"id": "train_027491", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "e355ff449fe017b3a0e28112d13bc0f7e46432c2"}
{"id": "train_027492", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "60da77e417a4f60414b00eb23c1a7db0f5ce1771"}
{"id": "train_027493", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "db9be777943a4603488408272b350c574fa16157"}
{"id": "train_027494", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "9c78eb76850d0c3a3dd47737bfc6a6cfcb478b63"}
{"id": "train_027495", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "e6e5e775fdb3c0f12eb1507e005db5ffebf5f41d"}
{"id": "train_027496", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "6b922d0b62c9b19fa410122657de5b928e8e106b"}
{"id": "train_027497", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "1972789fac2bac1c1852e7c2e561eccaf84b8b06"}
{"id": "train_027498", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "43806347a4703b5030c2b966b65363043578fb6f"}
{"id": "train_027499", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "bf782c38327a112de7e1b6abea404f90d84283fe"}
{"id": "train_027500", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "754e47772da9441e532e50d36e69394fc8bffa20"}
{"id": "train_027501", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "acc090d1986295945fbf280d86f0cc842f82b4c8"}
{"id": "train_027502", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "8d360de7f18d067ea53feb326a5a62a4988d5fc9"}
{"id": "train_027503", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["research", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "851c0f1e02ea365ad34d8302b1d05f6163c2cf94"}
{"id": "train_027504", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "00f807037862536f5c271a3c1a619d67ef1d36eb"}
{"id": "train_027505", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "bf782c38327a112de7e1b6abea404f90d84283fe"}
{"id": "train_027506", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "5355e49016bcbdc1720624bbecacef6ac647d32f"}
{"id": "train_027507", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "6e91c974b0ab75ded584bef5c1dac9dbcd867bcc"}
{"id": "train_027508", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "2e0661cebe25580b2c0a39aac4db07185b1d960f"}
{"id": "train_027509", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["security", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "dfaab52e21e66783106b486b099e432eb9142b1b"}
{"id": "train_027510", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "28b6c4aab232933797a96e08995b2e0e6f175766"}
{"id": "train_027511", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "bc8ba07e8f2f4579a9184a08ccd0586a2d0c33a9"}
{"id": "train_027512", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2283abcd73cfc3ef74511d27dd73b783af1e2604"}
{"id": "train_027513", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "ce4f8acd63445092ed2dc4f147a014cf877615d1"}
{"id": "train_027514", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "94c0560ec0e97afc334c3bb0801db4fe9b646036"}
{"id": "train_027515", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "6e91c974b0ab75ded584bef5c1dac9dbcd867bcc"}
{"id": "train_027516", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "8b38da0b7fbedfc6948e90ce6cd35e535ace8d69"}
{"id": "train_027517", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f12811ce14caab05c315e36fa8b4a0c91dcf24ca"}
{"id": "train_027518", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "58086daee3833a2b85c1074c9234a452c6d021e4"}
{"id": "train_027519", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "b0dbab00ffbb7cf157f61d33898d6b74d5099705"}
{"id": "train_027520", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "43c0c220fb4f9f48f42db0ac05871279e2970590"}
{"id": "train_027521", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "bff12f92dfecc73eac2f0af1eccb513b4fba0118"}
{"id": "train_027522", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "1a7b28db92cf4cb87c2fce597e09569fda618de5"}
{"id": "train_027523", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "635f593a73b68fedb04ce7c90d8e7e1070829986"}
{"id": "train_027524", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "c1465ca9dccd88798d53f72e45380ca3b0ecc673"}
{"id": "train_027525", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "4f1141212e555e632fe924eedf289984a8556b53"}
{"id": "train_027526", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "a91a14e4e9230cf215f240f9f48cfb4eb70574d5"}
{"id": "train_027527", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "5b91297cd84ec7ecd71456ace2c326785f8f8f42"}
{"id": "train_027528", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "65f41958e0e8b016bfc1d914ea4612b90b662270"}
{"id": "train_027529", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "d90cfc8842f397998353974b2f770ece667b8549"}
{"id": "train_027530", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "34f9ea6642298ffd41140a3ec34176d6e65cebac"}
{"id": "train_027531", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "56066e03b782ecc71b5e73297b734cd917e7f065"}
{"id": "train_027532", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "e8622b64ca093843f7cc5f76b475867c2daf9d39"}
{"id": "train_027533", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "044251fa6471ed1eecc54bcefb2a5f4ec11262fd"}
{"id": "train_027534", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "8353afc2d168097ef5cb99ef0873f0e186f835c7"}
{"id": "train_027535", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "5d7eab3caa0d3c3d53fff8bacc5a3760ab58a036"}
{"id": "train_027536", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "5bcc51b2554c09e4eb4ed70e5f4de6f5fd97533e"}
{"id": "train_027537", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["evaluation", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "353685efe6ce33f95582aaa8632f736248f0f932"}
{"id": "train_027538", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "cf0e8525fd29816f1f9c4a9c9789f6e58ca1d686"}
{"id": "train_027539", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "4f086522f1d7c9c9ec1bd9270861c8426a31b3ca"}
{"id": "train_027540", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "71a555c6c22b1930d95196515621a8ff988e5236"}
{"id": "train_027541", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "6c50a9daf8e3f849362d13450a5b87be6cbb59b1"}
{"id": "train_027542", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e7cae698e2bedf062a59dcd2b0654158adf2c19a"}
{"id": "train_027543", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "5207ad9d7df02ac1e0bf5cb7daeecaae031ca41d"}
{"id": "train_027544", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "0fb9db30556fa12f004d8ffc7e3596f87aef3fdc"}
{"id": "train_027545", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "25ddfc49f017944a3770a8baa502202df1420796"}
{"id": "train_027546", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["coding", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "e001a2f4c9446195168209942bb0a87e8a4fc39d"}
{"id": "train_027547", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ce98f526c6111622d0faae43749493c144a8a236"}
{"id": "train_027548", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2ed21fcf82398c918afa3ed0e6dbbf0244c8666a"}
{"id": "train_027549", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "642ceae1c63a98c069f907b7044433979b09be43"}
{"id": "train_027550", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "f7f17b0dee8e1992fbbad53d421ceb6bb05ea70a"}
{"id": "train_027551", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "56998fc30b9fb29699160b25707e92964246af5b"}
{"id": "train_027552", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "8fb2f4c076acd9fae727d4267c82844f4475d534"}
{"id": "train_027553", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "62476d2cb2b8276f1e12606067b660029c5c8f87"}
{"id": "train_027554", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "6d85899bf14a27e9cd577fa71d0f6c566033b368"}
{"id": "train_027555", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "2f0d7f7faa142db675e254b0e00a8622337e39f3"}
{"id": "train_027556", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c0efcc317ddd76c7057ca05b9c124ebdc0547bbb"}
{"id": "train_027557", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "55beb3fe79a0bfabbd24d3a0ff9c1e5bba6b2da0"}
{"id": "train_027558", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "2165f5d9e093d4f72a661d283b713a3591525ef1"}
{"id": "train_027559", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["research", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "0eb760edd61303620aeef32515c3383fa4acee1a"}
{"id": "train_027560", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "c9a9dd4f9ad7d54cfb917dfcb21bc9526e43b335"}
{"id": "train_027561", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "3739955e452c383df60ebb8879a0fbc904ae8500"}
{"id": "train_027562", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "eabe69670a2c1c55c04cc9837f93a85befa013e7"}
{"id": "train_027563", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "87dc246719e1145200f83593bf9b89a8607cf4f2"}
{"id": "train_027564", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "d713b201c0dcf6a6585e02d47d72db22ed805b8d"}
{"id": "train_027565", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "29b9a7b719fe6cb38ff480cbaf4d3742e34e2abd"}
{"id": "train_027566", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "abcbbf2c0429ffa3094f394e2589e58dd1c0ab7c"}
{"id": "train_027567", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "6d2da016e62b792f9e808bdec69b021935228cd3"}
{"id": "train_027568", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "f74213aa227a6677f98ab96e53df7e25a5e3b525"}
{"id": "train_027569", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "a833ef6c370ba2868014b667f04d7bda55acb56c"}
{"id": "train_027570", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "cb1ce73bef491be806cb3a2aea1c59c95e6aa056"}
{"id": "train_027571", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "bbbbdc44b64153e44a583085418945f3f8e52a23"}
{"id": "train_027572", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "2f8112607725779979219aa92d0983a96dec53f5"}
{"id": "train_027573", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "9bd02ed242bdb4b71f47dd6d6eeef525489727a4"}
{"id": "train_027574", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "54f578a4daa898aa4bab5615d33ca23bd84e27c0"}
{"id": "train_027575", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "1e2dd60dffbdec5553b1c207ced219a162a6e9e3"}
{"id": "train_027576", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "34e1f8a9cbd01dba445fa9633998d7bc41fcbd9e"}
{"id": "train_027577", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "b93c2eaa45a7597e49eed196107305d5cf9f7500"}
{"id": "train_027578", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8782bf0591e189bf51a5bd5ca240642d165cd744"}
{"id": "train_027579", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "279788254cf318b8fd5749eb754dcdba01625326"}
{"id": "train_027580", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "8dd1dbd7e8f37e6674de015efb2e1eb718f10131"}
{"id": "train_027581", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "55beb3fe79a0bfabbd24d3a0ff9c1e5bba6b2da0"}
{"id": "train_027582", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "addd97025bd6f1427c14d912b18e73421e587424"}
{"id": "train_027583", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "5f14db28b948f90f67f06e80682c1bbd45713077"}
{"id": "train_027584", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "ebbedf6cb272b41b7720e582d0d19b3442820e4a"}
{"id": "train_027585", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "974317bf967f6e246a7e80eb6a4d07b2d9226bc9"}
{"id": "train_027586", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["coding", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "e1ca50fd7242047e5d8be1918b10219237c7d2dd"}
{"id": "train_027587", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "8dc6d929663932b75372b541e444d150a4b2f48d"}
{"id": "train_027588", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ac289dc75f73bfda237cd466378c808da4010546"}
{"id": "train_027589", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "60647cf4a29a0d8438b6d9ef47acbf32915fc2a8"}
{"id": "train_027590", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "8269f02a66cddbe595d1f348a86ba1d171a97460"}
{"id": "train_027591", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "716e4bf9b3369aa512beaf99984b20c81d391c67"}
{"id": "train_027592", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "1806ebe8f8da2aade4a9780c111d4f81f6becaf2"}
{"id": "train_027593", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "740e12552ed276d9501098bc82e9c8c302a992a3"}
{"id": "train_027594", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "f8060fe1571de1392e1d8ed4a058b8fbec34bb8c"}
{"id": "train_027595", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "59750239f0b9fcbaf29965a609377d59bf472f36"}
{"id": "train_027596", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["data", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "84474590857b165e22dd1d76d4ed12f3b2e6557d"}
{"id": "train_027597", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "2285fdcd24ce9b2730cc33843a4124f6634edd3c"}
{"id": "train_027598", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "832a58e63a378528f9f6af2e326a018e8684c547"}
{"id": "train_027599", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "516da9d4dfa3767f9e650921a97a31d465e3a94a"}
{"id": "train_027600", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "51bf39ffb751a2e3a2550e94b6a00d6be38f3b17"}
{"id": "train_027601", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "15ae73004dafbb87ea1cc49fcd3119de8f45d3a5"}
{"id": "train_027602", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ce5d647135b99d86136a9c57c3c9aa1510af2052"}
{"id": "train_027603", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "534b4df7a648e1162342e164439c5288f53224d4"}
{"id": "train_027604", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "efa676a469a878522548fb268817cfb0eb03c6eb"}
{"id": "train_027605", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c567c83b2b0a20077737fe7a17458632b5afc40c"}
{"id": "train_027606", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "76c52bab886c448be03038eaff9a48e37cf6c5a6"}
{"id": "train_027607", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "fca5df3eeb85b48025635f9d8e58154a4ea95654"}
{"id": "train_027608", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "bd6fb5b87806e7ed8eb656db6566c6c0f41ee4bf"}
{"id": "train_027609", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "eb6e7685725710adc5be99e21abb70d33d041272"}
{"id": "train_027610", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "2af146be4e68ef71b815eeb735f7bd8858a74511"}
{"id": "train_027611", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["data", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a626003ad6efee3efc2b53034c134b50390c8fae"}
{"id": "train_027612", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["coding", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a0ab8c86505f60ca1d2cc5ca8221ecb41099670d"}
{"id": "train_027613", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "9ec3b7835d76d1e361a5f00b5940eb2d8d692bbd"}
{"id": "train_027614", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "0b2365bf5abfb01f127245f1fa482fcb7cd27184"}
{"id": "train_027615", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "bdb8767fe0bd8a9a205a0ba88398b8911a404d1d"}
{"id": "train_027616", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "88b84cae9765369fc51b273aec0a29cdd61fdadc"}
{"id": "train_027617", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "8c5df1ed2612b2a755173d3572799e01c7f18753"}
{"id": "train_027618", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "8798f178322fc2f4184fe027ee8bffef5a61d170"}
{"id": "train_027619", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "0afab87473e921d1eb85dec68555c0a3b807f5bd"}
{"id": "train_027620", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "65898ebbaa23aca13d2cadccc6463ea489b1c348"}
{"id": "train_027621", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "1bc9ae1d3cd2e53bafd9091cb32a8e3b6976286a"}
{"id": "train_027622", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "6443d6bae5ce140a2103b990b37c7f6a9f8ff127"}
{"id": "train_027623", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "65f712ee270b488d213b3f8d0a74e66d1881cf52"}
{"id": "train_027624", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "306c0f452ca3e04c5174ad19787b4c49a7e561c8"}
{"id": "train_027625", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["coding", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "9a39eb530c2fb14ded88901254b180d92046bd92"}
{"id": "train_027626", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "94b400a3807f9a8d016e0dbe3bb110529edb634c"}
{"id": "train_027627", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["coding", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "eb680dcee8f15889404e0d3ce9bfc6c9713839e2"}
{"id": "train_027628", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b038dcafb689fe369f36632db9e7dfa883dce6c1"}
{"id": "train_027629", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "f11c4627fbf44e0d9efef93ec961476cc1cbda13"}
{"id": "train_027630", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "7cc69a2c441f9895a038f181efd89c03c445fa9e"}
{"id": "train_027631", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["data", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a314c894c54fd4927c7bda72ed18fb9bc4619237"}
{"id": "train_027632", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "19ba3272700ca45a886d220c06776dc31f8bde70"}
{"id": "train_027633", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "601dec19446cb7cf08209b91eb2bb57e70abf4f7"}
{"id": "train_027634", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "b0a7468bab2a44a946caffa3f788062439368250"}
{"id": "train_027635", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "19c6e025d39a7f8806fefb9d0955b7d749bfc1aa"}
{"id": "train_027636", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "6ed76ce433c58a59d546af59154be1f40dd63d73"}
{"id": "train_027637", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "04d8cd07d19ac0c3114d8d3125c305546faa8bd5"}
{"id": "train_027638", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "2feddf3eebe3f97d2b961b6fe8064f006e347960"}
{"id": "train_027639", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "cc60357b2487aff5b864baab905f7e5cb1ff430e"}
{"id": "train_027640", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "524b0d0e497d0f95b6ff534e164c7e658b8ac428"}
{"id": "train_027641", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "009c86852e0325875b9ca8cd17c849e540160c7f"}
{"id": "train_027642", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["evaluation", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "75bc660e35f6fcaa46730fb1745dc6065b313ad8"}
{"id": "train_027643", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "53562430d9d5a3fb65803433d143c3d220ae8bfc"}
{"id": "train_027644", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "0aa6fc322e5f9e3b9daf6665ade7461a33d8a716"}
{"id": "train_027645", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "0eee0a6defb188677b59ad6e698d92c6c12c1934"}
{"id": "train_027646", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["data", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "ba5711d70f10f238415f2039e62fed7f40680093"}
{"id": "train_027647", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "fefdd458de10bd9e662ed965855c60613e128266"}
{"id": "train_027648", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "979ac33282efde343460ec6295a0d17b2d2181c5"}
{"id": "train_027649", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "5a5ac1489e4f475ac79c145ab35506e94982254d"}
{"id": "train_027650", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["coding", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "295333e7335aa3e5e19b1a54d33763f1552349a8"}
{"id": "train_027651", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "dc53f4aeffc2ea27fe26a643f2e4d986febd7ff6"}
{"id": "train_027652", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "303bd770ea1e0eff94d22008fbc02cb58701362d"}
{"id": "train_027653", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "7af18d63b36745024079c177f34df365ed342fbc"}
{"id": "train_027654", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "4b969ed6a585f56931d21a5ce0f0919cda37d95c"}
{"id": "train_027655", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "0910809a7fd00e74412ceb48ebe861e1afe1109d"}
{"id": "train_027656", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "a2773dea1b42186e6b70778ca93516f563516eef"}
{"id": "train_027657", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a682b9967c8c6c232eb54056a60cb8fc34cdb26a"}
{"id": "train_027658", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "92a93a682a6bdb985723e7ad24c967f96bc3b9d5"}
{"id": "train_027659", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "42a49b1a080770587de92d57312d896996401a5d"}
{"id": "train_027660", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "0e00a7b47b0cd4cf6f5b16b99251ea0009d34856"}
{"id": "train_027661", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "6bf042bb90c9ece488d7e1c8641cfa752c8a53e6"}
{"id": "train_027662", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "334dca623ce42a9f92eccb2e437026a5b43b84c3"}
{"id": "train_027663", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "a5a1327cbae1387934643ae308805164f1a5459e"}
{"id": "train_027664", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["evaluation", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "991e063358904cd6075fdf7c8e6702051114e087"}
{"id": "train_027665", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "2fccb0c5deab1b293e3a067f96e32a212dff9f31"}
{"id": "train_027666", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "67f17f2b6d4237e8efbe8343e80af5329d62f291"}
{"id": "train_027667", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "4369397aef07c4cad1c6eb49058d373f0ecbd391"}
{"id": "train_027668", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "5d7425afd1ecb79ace73595259e6ddf0d6e9273f"}
{"id": "train_027669", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "0544b47dacfe52cc0aeace021d739dce1868ab31"}
{"id": "train_027670", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "e482bb2f836558a80911d13551b6b8306c5bf523"}
{"id": "train_027671", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "244ebfd72762d0f12fc35f61936f2de69e9f8279"}
{"id": "train_027672", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "5dacefd5effc4db10a7aa7c34abcad78f790d236"}
{"id": "train_027673", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "35f71153ed7c3af5f274830312589421b102abeb"}
{"id": "train_027674", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "412cb12f44b0160ee3a237766f23304de63bc3da"}
{"id": "train_027675", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["coding", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "4561caef5acf8d505b9872163d3ce322b60a635f"}
{"id": "train_027676", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "65d06de36c74f3b2cea7a5442b2dd6a66067ebaa"}
{"id": "train_027677", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "20973a2f759110629a7c3c97ae2425bd42529e11"}
{"id": "train_027678", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "6cdc4cf34f3b9e7c403ee590e7f78f106ad11a5c"}
{"id": "train_027679", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "27b61053ecc757d2655a25b85a1a1d96d9c63db3"}
{"id": "train_027680", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "9e21465bcc63ecafb1166a6b5dd65a24d956a02e"}
{"id": "train_027681", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "5c3adb8c117c1f14f31f0f341d5ae4666ce9ec1c"}
{"id": "train_027682", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ac7a608ecde8bcf57135c9d8a8d902694ad48a3d"}
{"id": "train_027683", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "5de9ef2b46d8c2eacc2ebac629d94ae7f8e21004"}
{"id": "train_027684", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "1de55bce2a9c1df00407d033731beae9fac0ee38"}
{"id": "train_027685", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ffad3061c20e95198f0cf38f84b7a83f1523711c"}
{"id": "train_027686", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "08f3a4e9318521860f581b2ce73a56f26c128812"}
{"id": "train_027687", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "66df9f989e1a5065439e599c7e060b000287c139"}
{"id": "train_027688", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "6453c8e0ca15843cac1e3d18ffeddeea8842f887"}
{"id": "train_027689", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "6e07fc9d84c54d913fd51e169d6f9317215541dc"}
{"id": "train_027690", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["coding", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0b3a9067201654a46a6eae189fa0be7622295110"}
{"id": "train_027691", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "397de0a0711d4e9569a151bfad4cc6e32b383f88"}
{"id": "train_027692", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "f72904487ffd39fbda54628a8016cab22e20503d"}
{"id": "train_027693", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "0f79bc154f88f6e91a2d7a6dc66b9b209ca2e895"}
{"id": "train_027694", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c69d709438c534ce9dabec030f6031e2f37c519f"}
{"id": "train_027695", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "173351ec0b8ca3cea781fdddce09fb73aba7a160"}
{"id": "train_027696", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "32758860cdfad6c15ff250df079e271f389ca2bc"}
{"id": "train_027697", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "72c7bb5efe3b3d4f514ab57815fa81939ac03501"}
{"id": "train_027698", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "448fe2b1bac621b161da5cb8aeb9b5ba7992c31a"}
{"id": "train_027699", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "412e883cf7df0cfe5d5491fcf90827868c189672"}
{"id": "train_027700", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "b42134a88b87bae7e85ce1c3a545a05f3d9385be"}
{"id": "train_027701", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "905c267a06042e66dd1f41a2d255186439c56ae4"}
{"id": "train_027702", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ea5566ab3fb701b105695f6c6cdbca4e9cd0a001"}
{"id": "train_027703", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "c23570af4257c1e858da2826bccc828810653731"}
{"id": "train_027704", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a776c6c813064fdb075bcc9f209e13b871cd795b"}
{"id": "train_027705", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a8deeff04311b85b9ba6469f967ca5cf4f9a59fc"}
{"id": "train_027706", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "1c3e85a7d3d9f051393d4b51d5b23f73bf15e9ad"}
{"id": "train_027707", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "7c9730a91441d7b5eb4fb4b76dc992988a6fa276"}
{"id": "train_027708", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "705b8a594ce5de5c845c4f48912abedb4a7cad62"}
{"id": "train_027709", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "652eb236ca982950f5fc598943ee09e9dee8bb36"}
{"id": "train_027710", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c46ba4dd44ba3281606c76ea9afb9ed60dd868ba"}
{"id": "train_027711", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e1336aaceb8d877f92000eb83c7be96ed7a0c8d3"}
{"id": "train_027712", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "e565d5793a4a21c459bfed9d353b035e7a3201e3"}
{"id": "train_027713", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "4fe8c5ff266a3d407622d1b947dc66bd7d3253a0"}
{"id": "train_027714", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "ffcb8faf9beff9d1422eaabf7ac9f9daf994ec13"}
{"id": "train_027715", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f832b375c4e6287443581f686e46f7f9cc556f21"}
{"id": "train_027716", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "5fe97ea47e593d468bba4291d991876bce87a73b"}
{"id": "train_027717", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "5b54854a426e463d45ed4ad9aef7533872b805c3"}
{"id": "train_027718", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "a3d4bb9ae4d2dae1f21e53a4823c5e9531646ee0"}
{"id": "train_027719", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "8199d9f1aa2e1393fc37369e8b1f950a5e97783f"}
{"id": "train_027720", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "cdaa503bcbec9e840582e1394e0c593e6d3c7ff3"}
{"id": "train_027721", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a17b4b64bdd8ddbe1408330b43a3ab59db74dbe0"}
{"id": "train_027722", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0afdaaaed6e45f233c5f65c5c159b3e106bdc24d"}
{"id": "train_027723", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "098bb6a9ccb199dd15ad3b0c7c94ded747cb86e8"}
{"id": "train_027724", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "4f3c1a2582eea8af3c79c9c1f963f8e2e6aa2312"}
{"id": "train_027725", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "99c25c2fd9951aae6eccc56b0a5d045c195cef9c"}
{"id": "train_027726", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a3bcb6cc301ddd15898f5cf492c498a72467ce47"}
{"id": "train_027727", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "9b6fb6b64c09c0753b43378c2643e8ef9e1c323b"}
{"id": "train_027728", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "3e633bd793dd03a4d202f90638b001cf264bb8c7"}
{"id": "train_027729", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "72602667dd04f1b0c8ecd6e2fec4cd01907aad36"}
{"id": "train_027730", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "624ca98d01caf4847eb0831833d125b33fac6385"}
{"id": "train_027731", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "dfbacb060e1d92732517711cef53dcecad82eedb"}
{"id": "train_027732", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "893cf2e61f3fe8f5d88d1e7930d6e6d17b173ad1"}
{"id": "train_027733", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "58c1fa532ea768a661833cfa541744572bce6e70"}
{"id": "train_027734", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "3dd1ed2f298332daab5460758a0b2df6856ece18"}
{"id": "train_027735", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "a2a0ff054cff55eb05503366ba33b89479efa6af"}
{"id": "train_027736", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "92304c8619e170b4061e5fac0309e6251b3aac0f"}
{"id": "train_027737", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "0c7fc467367c957b711343eb47954707b3c258c6"}
{"id": "train_027738", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "2a9cd653846bf30d7124fde6fc8321d23cce63e0"}
{"id": "train_027739", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "b8e78791d50301e92837e4ad63201d32196d39e2"}
{"id": "train_027740", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "6551a1e94492490f234a1adf7fb0f2dd8968c3b9"}
{"id": "train_027741", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "03c87c63440509500341e3d0920a237ddf8bd63b"}
{"id": "train_027742", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "aeaf616d20c95e1aea1ac2a376289071e2ba4eb0"}
{"id": "train_027743", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "84d309e02161ab0b24045c689897dc70c76d4b41"}
{"id": "train_027744", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "cbbe3fabd3d0681bed6ceb3dca0703da33b1fafb"}
{"id": "train_027745", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "aa4ea851545b3039d6cb7accb33e210fc30d72bd"}
{"id": "train_027746", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "19bff1069dd84304089908959a1c2195de1b25e6"}
{"id": "train_027747", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "01a12492d7d2d257ee4e8f87dd327e36f2d7660c"}
{"id": "train_027748", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "134ea907f0a52f228044d04ef1c94687131bdada"}
{"id": "train_027749", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "9acded0aa1d9febc624c55dea674f9fe7f9a698b"}
{"id": "train_027750", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "6ddc33e78b4bf43d07f4c1fb5c9a55efe9089245"}
{"id": "train_027751", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "cd644527648db31004ce1cba5067dc2e36b6ff8e"}
{"id": "train_027752", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "0792983c1c4bafcf36bd9f2e0c4f3fd21c27c677"}
{"id": "train_027753", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "04baacfef37014c4f08bb284c47f517a5570f589"}
{"id": "train_027754", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "c01e2a88423d8c5860ee456e8e64d5bb98c6d3e9"}
{"id": "train_027755", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "537358187388b4279376312536a1befde6b35a41"}
{"id": "train_027756", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "31d411b1777ba2d62968b6396d381d9ed9a2218e"}
{"id": "train_027757", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "69b2914ea6e409476585f7949bf3482d0aa3f188"}
{"id": "train_027758", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "3137be50a0ca59c148a21415872caef86a766e03"}
{"id": "train_027759", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "4cb2d692e0ce6b5b277d42ecb091a5facf172bc0"}
{"id": "train_027760", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "e0e2b9efe4dfe32223cf5667692a640b841a7257"}
{"id": "train_027761", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "4b75b7e76c6f21dac282e6fff409e433463c38a3"}
{"id": "train_027762", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "9d5f23987fc6fe524442fbcddfc2d4244e5c27a6"}
{"id": "train_027763", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "072b41fbcf23aed5786ee3c5b0702ed1f7c988b0"}
{"id": "train_027764", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "569a48afbf4fe97430035faf4959b36b40ba6dfd"}
{"id": "train_027765", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "e7d6a3e70f195ac9fd6e60be2f75aa0dadcb8cb1"}
{"id": "train_027766", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a2225cb3ead123252930f3014d48d6e1e3bc303c"}
{"id": "train_027767", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "c56caab787d383bad24f994f13fd6dda63b16c5a"}
{"id": "train_027768", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "45f9b0c296e1bf4ca0972d9f7d6f894a1c182071"}
{"id": "train_027769", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "68fc7531bace5be46990d38bf062b9e2421a4883"}
{"id": "train_027770", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "8d45ee23d46bcbca05a155d295992ddb3722b1db"}
{"id": "train_027771", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "d1a3b108fb1a68103d6625c69cd83f905f84c2ee"}
{"id": "train_027772", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "d0e940910012aa13d609338db9af2c2686e08dd8"}
{"id": "train_027773", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "6a5d3d3a68c30e9769c69ed721613e8fbb488a42"}
{"id": "train_027774", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "9ea75eb9497f6ae5eacb3077ceb40e7514c73b0e"}
{"id": "train_027775", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "48f8f5ad05ea6be84af85a466b23d972bbfef3cf"}
{"id": "train_027776", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "4cd01657d0e6a03dc6b2447189518adac568dc1a"}
{"id": "train_027777", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "a10191c11a1b08b4512466da8116379aa8a39fb6"}
{"id": "train_027778", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "35bf3a7b957bd784767b37422fc48ca54b6059f7"}
{"id": "train_027779", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "b958b9bd0efd2b33528129db7a1f9cc1640a61d6"}
{"id": "train_027780", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "4ee67a44f71836f5519c211a8dca96c6f7df677c"}
{"id": "train_027781", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "e4f1fd802a4e0dc383658887273a5af4da8f53ea"}
{"id": "train_027782", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "d85a9369eeb988c8e7624334f70cfbaa362697a7"}
{"id": "train_027783", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "dca048a4f56bea4954ebdb4ab2a0e12dddc50353"}
{"id": "train_027784", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["coding", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "1a6dcbd505b180c3084b3f003677e8a660f6aaa7"}
{"id": "train_027785", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["data", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "5cc20219451136e25e60d83a7132284643faf250"}
{"id": "train_027786", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "0c933b349417011c3d5824e5cf9633c06c352ebc"}
{"id": "train_027787", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "a2f2f5a0aeb3c3b9843c23a0ce1f2679f990068d"}
{"id": "train_027788", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["performance", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "51d67b51da9d6b64c0b73f6d33c7dbde144edd7c"}
{"id": "train_027789", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "4e6fa328f1a417265c8ea73a5740e1af22e6c897"}
{"id": "train_027790", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "ca7d1d2a44da87ef9c010d314e0713421b3e6650"}
{"id": "train_027791", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "40efc2026bca54f5e9a79592ca04a16d1f81c5ba"}
{"id": "train_027792", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a83fc5c9139876b0b6214ee215d2421e8a75bc0f"}
{"id": "train_027793", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "109bb682c2369f58de7209e43f1acb9ce2202fff"}
{"id": "train_027794", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "072d64ce27be2199514d04771c5a1caeebdb22ef"}
{"id": "train_027795", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "bfa750570a8dc7458e69715daeaa1d50c17c53c7"}
{"id": "train_027796", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "8c44d7a74dffb6d6f8477dc6badf26ac0372e51c"}
{"id": "train_027797", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["data", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "ef9cc0cffd531b25ddfb41b2a18b9d9b5acc5c31"}
{"id": "train_027798", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "220a45166b53f686f83bedaa300ffb970d7b4e79"}
{"id": "train_027799", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "705b8a594ce5de5c845c4f48912abedb4a7cad62"}
{"id": "train_027800", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "8a4bdb0134261fd06e5e07dd3cacb54ad23049a0"}
{"id": "train_027801", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "3ee8883795960430c3db5421ced93325256d336b"}
{"id": "train_027802", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "32c3284233f56c00b69b27677c41d8eaf6b0b0ab"}
{"id": "train_027803", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "1f972b64a5cc3154af836c72008fc8a6b2d5c2bd"}
{"id": "train_027804", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "7bb171fe8bb47079c1571386ca7ba2a8d2289b39"}
{"id": "train_027805", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "2d9c7b7ee5740331d6ef878a10a819d7685f94cc"}
{"id": "train_027806", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "52c159e20b41db6063ebc81b2e3c2e72da0b8ccf"}
{"id": "train_027807", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "d1139699c31fb48cc6cc299d1e9201c6e9f4df08"}
{"id": "train_027808", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c951e5c2ecb0e21f751fd9bb21331ab52eaf61c9"}
{"id": "train_027809", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["data", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "87c146c3c62e52cc8977647cb320b55adcbe99c8"}
{"id": "train_027810", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "2f99146a6baf28ab3d60d35e89a550d56fb966ea"}
{"id": "train_027811", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "b44f81323d2cdc66e23b1e6c449d416457842f04"}
{"id": "train_027812", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "ceebe7197ca3775b51706f8e16bd31a3cc66b5ef"}
{"id": "train_027813", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "24c845e2723af903dced58c482deca67b6b02e8c"}
{"id": "train_027814", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "22366c088539c6b55577fa988e4ef910194fdf20"}
{"id": "train_027815", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "9c0b9477adc8665fd14b8786893622d14d66b641"}
{"id": "train_027816", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "5908ec8c26b52c257898650cf5402e7154a34c05"}
{"id": "train_027817", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "f4cade7d8c8e935fa4cd3712ea8b5a6270c81b58"}
{"id": "train_027818", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "0b2a49d38647bb429ae6d1567b1383ae5a870128"}
{"id": "train_027819", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "1073c3e2a533755b9f7dc5c7eea93a622cea586b"}
{"id": "train_027820", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "6e74c838e27c32dd5e491115a2b7fc8a5c6b4bcd"}
{"id": "train_027821", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["research", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "44d0cd699276beaafa9280b7a671b584325b9363"}
{"id": "train_027822", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "fe44ed408fd7913c3895fec1c00be4b0f1100fe0"}
{"id": "train_027823", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "6bae441cc17740590b119c80623200895232d517"}
{"id": "train_027824", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "3950f28a140db0a014583a4ad487469d52d37289"}
{"id": "train_027825", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "601582afbb75fe3d919d91d00a9abaf42d24f4ce"}
{"id": "train_027826", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "e7015c89be1af488babc02a1446a1e321f73de5c"}
{"id": "train_027827", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "521cb5b7762d606581adf1bcf0fff355eb8083c7"}
{"id": "train_027828", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "5c7fa5c83030db25736f19e892c794d369a4a3ea"}
{"id": "train_027829", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "cf8e1aa657b389d464ae5388082f99df2dc05b5c"}
{"id": "train_027830", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "626fdab58f2e27525f312243ff24f91c2eebf3bc"}
{"id": "train_027831", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e38e5da78eb195642defd1aa79bd1794031d59d1"}
{"id": "train_027832", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "0667dd2a816890e0994eadb7c5346534f14b70e0"}
{"id": "train_027833", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "b0b6f841753b12d05df358ca85a97fb9d3b45ebe"}
{"id": "train_027834", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["evaluation", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "73fde33c328d83ee1f1113049816ad1085f5ee11"}
{"id": "train_027835", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "013d49b9bc8f67c3ffda6ad23c50dfbd0a9273d1"}
{"id": "train_027836", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["data", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "79158dfdac0a615aa893975f34528b115dea9625"}
{"id": "train_027837", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "b1734322e8869d303d929e785586c3643434ed74"}
{"id": "train_027838", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "727c47fc57b61afc028576e84504a13a4c499bf3"}
{"id": "train_027839", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["coding", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "ec4791bf1adfdea36d4d9557489676082387e6d4"}
{"id": "train_027840", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "92ea11129f0bbe21db2f788aee99dd4e62151788"}
{"id": "train_027841", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "4ff564f542c1b70126a29779b6e487b072a40e66"}
{"id": "train_027842", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "fbf1ed863608b4bcedc9744c6f48763820c70bae"}
{"id": "train_027843", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "00f27b4a2fae26124e42bbfb5bca315de2570638"}
{"id": "train_027844", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "76fedf100e672d216f60656cfacd06e6726b5d45"}
{"id": "train_027845", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "75c79b299ca5fafa81b06f0f92aa81a0635f22c3"}
{"id": "train_027846", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "51936bda196d42c3cc62212e3b07fd3a0482a671"}
{"id": "train_027847", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a39e80c91db68f9fe6cf87f2acb9d44e8d807906"}
{"id": "train_027848", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "fab4987f5beffcc2d20326415c4c6367fabafa1e"}
{"id": "train_027849", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "b527bd368aa106a8320746f63f06f4427a4e1ec3"}
{"id": "train_027850", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f784f2960db226a72b7c5e63c0b9fc84d544d53f"}
{"id": "train_027851", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "6f2fad9ceaf8ee67afe4611f0e9cc6385245d83d"}
{"id": "train_027852", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b4886ddddbecf68d1edb118093dc2891986a9267"}
{"id": "train_027853", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "7cfe96af3fe0958b400b986730bb5f05294986bb"}
{"id": "train_027854", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "3d5d2cabd6ed87040ead19eccd21696c0fc65bbd"}
{"id": "train_027855", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "e828118d25a3a46b0f6008ed7524a737047df43e"}
{"id": "train_027856", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2614397a5b2fb1c694108652d73cd2b33c864fe9"}
{"id": "train_027857", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "03f1c1c83ef93c4e2bf3bf1e0207e388513e2edc"}
{"id": "train_027858", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b08c46236ad736d348662ded719d865342c72a40"}
{"id": "train_027859", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "0e4f8b04147b88dfa54b8a65a63708fef3953428"}
{"id": "train_027860", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8bf2457e21dd9e209fa33c8b472c215362b53819"}
{"id": "train_027861", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "84197167459f5ccc5e814579347770708b3a45c4"}
{"id": "train_027862", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "458f14dcd883e04df43989012b0c43d2322795f1"}
{"id": "train_027863", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "d50577813ce02536a2ff52260efb76c674ef8586"}
{"id": "train_027864", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "9f63d1dffd74bd7b1bd46feced7aada827f8be76"}
{"id": "train_027865", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "58c6669ff477aba863320ec2a040a3fade6ecf22"}
{"id": "train_027866", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "d3a076c6c4eef085734db476c30589bc20cb3c64"}
{"id": "train_027867", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "4cf5a6d31512aa1bb85cb4d473975bdf7f9bc4d6"}
{"id": "train_027868", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "eeb8697d6fe8b6e6e9a218664150194341bd0970"}
{"id": "train_027869", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "2e4476e11ff4e78df74de6be019038392170d945"}
{"id": "train_027870", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "11dc415f77e42160a724167e5d5573b5408e574d"}
{"id": "train_027871", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["governance", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ab501b86987ab74df0aec56a642e227a9d4d943a"}
{"id": "train_027872", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "0d765b7f434abad08798530df51075ad6e0c1ae7"}
{"id": "train_027873", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "e6049562c126c309aa15620b72a050c609589c6f"}
{"id": "train_027874", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "0afab87473e921d1eb85dec68555c0a3b807f5bd"}
{"id": "train_027875", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "063468880cd23cf11bbe41cfdb3d836311d13be3"}
{"id": "train_027876", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "dd50c5245af94ba7a2d6e0ea62ae58137d1bc12a"}
{"id": "train_027877", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "3d466372c67a8210216d71ee576239c925ddc60c"}
{"id": "train_027878", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "714fcbb7f555f85e40bc701d5b73e1c73a4bc23a"}
{"id": "train_027879", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "50e632ccb113e8b114bee92c64987f9420dc58b0"}
{"id": "train_027880", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "d6a9ad0d500116d1a9a1fc2036011f8b323f8545"}
{"id": "train_027881", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "43e6989e3fd3135c0b3f85484051b0adce70d97d"}
{"id": "train_027882", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "86db11b10ed2dc4a5710c776381bd19d6fe4516e"}
{"id": "train_027883", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "18083aa6b4e3df14bc5d5eb60bb56035d92a35fd"}
{"id": "train_027884", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "e4d74d17d3abe8031766bf607feabde9001fd929"}
{"id": "train_027885", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "237caaa61bc411f3a2e0451e666a57e7ff214ff5"}
{"id": "train_027886", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "036f832acac36000175f7d139b6d8f5b2cb4d21e"}
{"id": "train_027887", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "1ab0a890876a7c3193dae367fea1a08a6c5f7629"}
{"id": "train_027888", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a5bdf87aef9b7abc3fac2ad12187d55c6096817f"}
{"id": "train_027889", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "e83e324e95c726a9ef5d48854ddeb740c4ea7c88"}
{"id": "train_027890", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "202f0b2dcd4ed0999d55aca76dafd41a92249d10"}
{"id": "train_027891", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "c9f486f1530f4da70780143f1d91bd6068cc01b4"}
{"id": "train_027892", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d1419f42b8d2edb800de4c9944a4dea3413f41d3"}
{"id": "train_027893", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["data", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "b13f761f87aa77c034b7f961016fafff7f7b20b2"}
{"id": "train_027894", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "fbeb962dc28310f9aa7b094322cb839ba86ec312"}
{"id": "train_027895", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "2a7d7e2e097419834d7697c4f61049a24feaeeba"}
{"id": "train_027896", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "463df88fee31a7ea9662826d5c670996a3d6ba65"}
{"id": "train_027897", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f5d5fca7d9335e2867172619b4a07969ab3e7954"}
{"id": "train_027898", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "0821b4565a8050b5a6efe5a856ee08196c43d7f9"}
{"id": "train_027899", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "97e25ec09b8fc3343d655218fd9181d8089d2a0f"}
{"id": "train_027900", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "7f5f1b83f9632fe5e93f8432b869251de73cc2ab"}
{"id": "train_027901", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "9d6cf952a2c5f52adf99e07350933c8c8813e4b3"}
{"id": "train_027902", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "03c2d3e701f9820553f670c11759de7fd2f6e09d"}
{"id": "train_027903", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "3c0c8b1928a27bfd629ac494d1f37c1f7476b12c"}
{"id": "train_027904", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "3f82644d98dc15d7457cf1cc2e463fc54e26d7a6"}
{"id": "train_027905", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "fe1d1661a5b4a481d206bd2fffa62f9c6ee57656"}
{"id": "train_027906", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2ce4acd63e113825f2f80f4646f57e3b72669e5a"}
{"id": "train_027907", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "15c98b19fba5212469bd08a749b4757964c590bb"}
{"id": "train_027908", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "f413ffb53cb50cecf9da7d44eaa8dbc5552ff694"}
{"id": "train_027909", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "2b202bdaad9ccada990689b9d28ae6ae60ce67aa"}
{"id": "train_027910", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "f5829d6f3560219625772eeed4f0c57de2724908"}
{"id": "train_027911", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "a99062cc419d9b5a0c44f59914cfe70eded945a6"}
{"id": "train_027912", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "41f69f4bb7f1a74f719efa5ec6e33725115398ce"}
{"id": "train_027913", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "76a71a3ccd62e488fb7c39866bc3b62b2c87f85b"}
{"id": "train_027914", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "ec520fc7d8bc64f61c55ebb3ff922e4be2bbac32"}
{"id": "train_027915", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "2dac9a66e6995158a860b0e574fbd38f8b11d2c7"}
{"id": "train_027916", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "9bcb9a8adb7cb2283e0f2f47515584507146b317"}
{"id": "train_027917", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b0c3f48e81c7dda2b99d95f23a989bcb97130d78"}
{"id": "train_027918", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "8f3a0edf9f32735f39bc98bb1b1a8410c792b39c"}
{"id": "train_027919", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ff4bab78e80c9a9d9bd830443d2b123386cce057"}
{"id": "train_027920", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "7bec5ee83a0e7963c363026cf84b1dc2c9b2d6a0"}
{"id": "train_027921", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c07ca4f5c50797cf7d46b0abe389795fb7d79132"}
{"id": "train_027922", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c8b4db0f3d8ba601c4832e18d64a57ab176bf72a"}
{"id": "train_027923", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "3f452d943186318c1d7cb908b538818834e84037"}
{"id": "train_027924", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "225bbff30dcc4de7d5d1217573b380b5a1ed7afc"}
{"id": "train_027925", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "76ada0ff613cc1b8e3222a81b118789ec4035a91"}
{"id": "train_027926", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "b564fe25751975ad6db2cc331618361ea2d3c809"}
{"id": "train_027927", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "027400c74f50707388a83706bfb7c3840d123f8d"}
{"id": "train_027928", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "ab4172cfa0f3b7e4030d0822d8aba74e46e0bd76"}
{"id": "train_027929", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "2284b36d4bdb150273498300c74f73ea4539f71b"}
{"id": "train_027930", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "bfccd51d126e4a97e3c2e82103d04150095b054d"}
{"id": "train_027931", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "930e98bd5c3bb19ed9d6123281f067acee570fe9"}
{"id": "train_027932", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "601fad7380f323cfc2beffac7cc2a5d66358e00e"}
{"id": "train_027933", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["governance", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "cce81ba46fe0ffd603ec76385f4d07a244b219b2"}
{"id": "train_027934", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "6c677604c39e47df42e2c72ec618e3e86fb3f413"}
{"id": "train_027935", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "ce54614d5c66f14554b87e1def72139487145b65"}
{"id": "train_027936", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "f6e624250a005d2bafb9af181c92a3e623e3169c"}
{"id": "train_027937", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["data", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "455150d0e52c57e9f3eace0e8702d0b85b48ff87"}
{"id": "train_027938", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a1e91fb57456720621b799cc80b6c4ef58784f47"}
{"id": "train_027939", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "70e9da4dc29ab8c04ace2c904db4c32a918759ed"}
{"id": "train_027940", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "84da7f8695b1f08444fd226fe76f4db13e3df34d"}
{"id": "train_027941", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d543348000df0380e247721b787b04ffbdaf42e5"}
{"id": "train_027942", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "73e799f8878046aae2fc476e212bf7e002143af1"}
{"id": "train_027943", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "0f8e9b13d4395d91450c3f29031f88cd6fcdd48a"}
{"id": "train_027944", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["data", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "a7019b72bf03640e650472c1db1b928b44389be7"}
{"id": "train_027945", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "470fc57316b9ff975aac8157f62167ba69ed4a18"}
{"id": "train_027946", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "f0e223698f55a3d9c8d29a0c6d98d6ad52371f0a"}
{"id": "train_027947", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "114e2828f742477bce74698408d532a482dcac9b"}
{"id": "train_027948", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "db72f324dcd9fc0f57bb645b8366946cf0845d38"}
{"id": "train_027949", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "1d376aa991177beddf4b633745c8a029c64c2acd"}
{"id": "train_027950", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "3093540377e09eebb6ad67a052eedccd9712188c"}
{"id": "train_027951", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "e98fd2a249a903a93d924659ef7aaf4f1c9d7fbc"}
{"id": "train_027952", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "f93eaa471a7e4b0692b1830c206192b339ea08eb"}
{"id": "train_027953", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "d7bc156642229da6d4cb400a6fe582bc435431d2"}
{"id": "train_027954", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "93f2ec280f4858e2f2f3da1b6edc4ac604a31a02"}
{"id": "train_027955", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "5850d877e97adc2f71c3aa26fcb5c6e0430c33ca"}
{"id": "train_027956", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "361050e77b8aa6e9614e86b51bbd7a5a5ce8ee0e"}
{"id": "train_027957", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2d2f23c35f3d9ca23a257f973d1a19638a2421e6"}
{"id": "train_027958", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "6602cea9d988a5979b957e478d40c81f408cbb78"}
{"id": "train_027959", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "fc91f3e21a61ee78057a852e93c3bfcc93728cbf"}
{"id": "train_027960", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "6611e3e6e5f6c46369223761e087d30eaaada672"}
{"id": "train_027961", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "9d1bd38b0228dc20d74698908f026b35efe5bcbf"}
{"id": "train_027962", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "4b632bc798afdb3f6b3f72742e7d1663f09432e1"}
{"id": "train_027963", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "6c5c9e32d6b927a0d5e391d80fa9067f3de4b763"}
{"id": "train_027964", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "bc913d35583fcad0a3971d26288914a60a1cc7cf"}
{"id": "train_027965", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "157da1e78bf73dad90e3201a50ee037061bfa680"}
{"id": "train_027966", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "01bf26dd733e0192c63bca19d5ff24b1afb841bd"}
{"id": "train_027967", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "b46e1258decae7e87b7d38b092a0691c3cc875e9"}
{"id": "train_027968", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "19f6abf555f095f87c55e17b0ad10dc890dd91c4"}
{"id": "train_027969", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "fafbe48c23e6e9b7b8be0c5fee19783c92fc7a77"}
{"id": "train_027970", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "4642e0e888262c6aca87a29d3fcd7d09cc4f41fa"}
{"id": "train_027971", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "cd4ea64a8902caf584a7d2afd4d134a24dbd5ae8"}
{"id": "train_027972", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "e5b3cf1cf52601a6b57383bec9f37c31a51b98d4"}
{"id": "train_027973", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "4bde4bffc0529330211e0f636cff4c6e84e80929"}
{"id": "train_027974", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "03f5eb1c26b37718c636389418671be0cdc59d5c"}
{"id": "train_027975", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "b5f2638f09830a3248aef21b42df2da625acb051"}
{"id": "train_027976", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "92fdf803903f15c187fca8347fc782ad7cf898a7"}
{"id": "train_027977", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "433c152c8fab0187ea4d5bf7e8c9345c38f78bd3"}
{"id": "train_027978", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2c9e084d04ed7f9cabc27703dc43b4a6a1cb4fed"}
{"id": "train_027979", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "5516ed0ea400204216756f398212ca835c56f0db"}
{"id": "train_027980", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8186e5afb25a9cbbc1f6fc5f9de91921400da116"}
{"id": "train_027981", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "097a054c3e66ba3b74230a252f06549c6df9eb8f"}
{"id": "train_027982", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "fcfe53791b9be96720839478215a94cc3c5c3615"}
{"id": "train_027983", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "884d4c9ed9fc47a4e8e2f52e90595c7d47bb3d15"}
{"id": "train_027984", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "cff406a5946101f9209da1eafeb9b478b4b73225"}
{"id": "train_027985", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "737eddd8623baaa57c97f0800913a35cce4a0921"}
{"id": "train_027986", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "741559f10f38dcc96e18bd25336f0724357fea80"}
{"id": "train_027987", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "dcbb4246032b474761d8961c44f8e89aba8ff304"}
{"id": "train_027988", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "b0dc8b40a07986ba9146abe3fd771d212b8a7424"}
{"id": "train_027989", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "5345c1485354f863aa97106705fa37cf365e3338"}
{"id": "train_027990", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "6b9f5f817311b00d72215e2c6dcbc81d57cd8d2e"}
{"id": "train_027991", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "29e46f9a233e53b81e021d6809056d1e58994734"}
{"id": "train_027992", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "978431e2881c046f35653999bda7e943b7bb0360"}
{"id": "train_027993", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "9fcde85cf9d38c447328b68e5987edfccb36437d"}
{"id": "train_027994", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "344e635a4298aaed74d28f308ae9731e4071f276"}
{"id": "train_027995", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "662376c3992a9f8e92ef97a5fc29606b46a201d6"}
{"id": "train_027996", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "d84ef6aff6ae9118f027e614d7f5ead7f8aaaa19"}
{"id": "train_027997", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["coding", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "197a2637e9591db13ecf97bb90a35d26c11e242b"}
{"id": "train_027998", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["security", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "472e92bf2f14ab32538b4a5c9f3e3de412861ab6"}
{"id": "train_027999", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "8e0ff9c5e97003f04528f76457ee2eeed863ae38"}
{"id": "train_028000", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "64d591c19c3cc678791fc7c237625cbc4707963c"}
{"id": "train_028001", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "49f48e72b8cb895826b159a63061c8dcbdee0996"}
{"id": "train_028002", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "76904e222680e03266466135dd85aa06e2a742e5"}
{"id": "train_028003", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "492aebb8de275712f760d105f23b24aba50c4578"}
{"id": "train_028004", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "1b5e8b8faa00f56fab684a5efa5cfee4c686c623"}
{"id": "train_028005", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c046bbb6963612c8cf49ddbcc3c5aa5d84bab40e"}
{"id": "train_028006", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2ed845100c01ddc480b90dd93644c999e565af58"}
{"id": "train_028007", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "3fda00a5158fcf815ee9fe245ab178c1ff0acebb"}
{"id": "train_028008", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "1c803d7d28c9da7d874c41a1048be3a2bcc280db"}
{"id": "train_028009", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "6b756ecdadccfd91a4b1d462a222ecf42d8cf358"}
{"id": "train_028010", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "11ceab9fef1b2e416e164b3ee700b9569d425ec4"}
{"id": "train_028011", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "770c97b85df23e58a9756b14c9a9be9b22eb16a3"}
{"id": "train_028012", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "fbeb962dc28310f9aa7b094322cb839ba86ec312"}
{"id": "train_028013", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "fed4f2b8b29375dbfc88190848caaa82342b6852"}
{"id": "train_028014", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "a7fb6fb23e639d23cdfe5599b3ba659560901da6"}
{"id": "train_028015", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "13b580eb16f6234c05cb186d2aa2f51ea07e5d6f"}
{"id": "train_028016", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "887eace6ae63f8d7cb40f9815cabd957ab40bfcf"}
{"id": "train_028017", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "44f06c8ab24ea82aa96f7ea1b31db93d930f1f41"}
{"id": "train_028018", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "bc9916c29ab89a0a3d670cdc7996ba050d1dd21e"}
{"id": "train_028019", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a4904a44b1a1cc1f1c4c3f2c7cd3ed6c6bee9bc8"}
{"id": "train_028020", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "cb209d8c096d7e5246e1cd59ee5741c7f843766b"}
{"id": "train_028021", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "72a096216cc73c7e5e1cda1b7b8cf2a778581f77"}
{"id": "train_028022", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "1f368e8c550742776f487286f177c7c2faed8397"}
{"id": "train_028023", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "0dddb3d260e763bf7de4d14594e5a9143a1a01d2"}
{"id": "train_028024", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "b0a4a6b51bf4fbad49e0e1bb02b84c9f3df7e92a"}
{"id": "train_028025", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "84b4744bbde730167997ccb39267a2f50c76a006"}
{"id": "train_028026", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "3f372768e95c7ad6b286b6c803d234bedc782072"}
{"id": "train_028027", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "93d7e033b4c82771ba7d7a333defe14c66ef3337"}
{"id": "train_028028", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "f3e97f36bf66e896275a14db80c947c5c9af011f"}
{"id": "train_028029", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "5f39e7c1ed318266a8f9a65d2208a95c7b240879"}
{"id": "train_028030", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "1ae81f7a69d96c552f33099d6baa09175ec8684c"}
{"id": "train_028031", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "6f3ee6356243362a4eabf327b34cc203f2bcd5f7"}
{"id": "train_028032", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "64f2871ef0aafc9710074b28bca62195808dd6d8"}
{"id": "train_028033", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "38c8fa7f6e865ae5ae4b414cfa7504071617d163"}
{"id": "train_028034", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c602941b3963ba8d5f6f13039e7a45d054cb54b1"}
{"id": "train_028035", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["data", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "d1956555c9b119b96362617bd23b185b23d3ca30"}
{"id": "train_028036", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "04eb598cdeba1986a4127e9c6961de4761ea6929"}
{"id": "train_028037", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "87c4ed615f4d0200235f28529bcd77d113f11c9a"}
{"id": "train_028038", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "11f3c74e0c57a8bb075906c7c6ad33efc4feafc9"}
{"id": "train_028039", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "16ae1594836da5610ec3f2126932c1d6becf850a"}
{"id": "train_028040", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "fa52e1a2d5f28104a1eb422522058b48d14127e2"}
{"id": "train_028041", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "888420f6a746982494ff01503c5b86922bc038c8"}
{"id": "train_028042", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "c6576eec34559a52a3c25cb339616eaabf98cf3a"}
{"id": "train_028043", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "e9f395021e43a7542908a4e60379d7e31b5410d0"}
{"id": "train_028044", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "6d8b938f1a7455947d8aaf2348c7b70354eb06f8"}
{"id": "train_028045", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "e6a4db9ca2ba432fb514627cefeb8b39a8cc68c1"}
{"id": "train_028046", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8e8953b3b64855575688024ba99cd79f37815912"}
{"id": "train_028047", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "eb287ab03c21ab30398607e864e13573b08f1522"}
{"id": "train_028048", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "e9371d78a44e8e671c34ba862483b26b3a6e7a1a"}
{"id": "train_028049", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "b691579cf5790e4223289606cb78d12f7f6c9f65"}
{"id": "train_028050", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "1cbf5a799a7a58db9ece25ac39ffaf46726a2387"}
{"id": "train_028051", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "188a396f3d4be5e08d7b4b68263ab139bf6068e6"}
{"id": "train_028052", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "3224125a421679abb5664020ef6f76d1e150ae5e"}
{"id": "train_028053", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "ea5eb3584a03183b87c8711042666229948520b5"}
{"id": "train_028054", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c1ff2b3e9377f28a98fdbaca5ebe17e24a13491a"}
{"id": "train_028055", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "3c3e3328d51059f6ce5a8f6efee297cd8532fe13"}
{"id": "train_028056", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "542a94e2cd9fb09558bbbee1417d28ceb37ac48e"}
{"id": "train_028057", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c275910d8dc22adf762ab950cf37ed1d90cd6a3a"}
{"id": "train_028058", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "7d332eef9d12ff68a5fd61b69e0f228c50588759"}
{"id": "train_028059", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "b99ab6b66c3cd1c43905e181b6682a0185355b55"}
{"id": "train_028060", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "1530916fc57f7cb418e9f91145e13003639c4d5e"}
{"id": "train_028061", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "29e07740cde3c0786cb2c3db4212b32f1b0a84dc"}
{"id": "train_028062", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "f507938f1a00e277a79cae7aa448bfa3f92f2ae6"}
{"id": "train_028063", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "52ec5173f7d7ae6863e3b9902779695643ed7f38"}
{"id": "train_028064", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "35d957bc34704b1d4ce1ea6dbcee644964dca3c9"}
{"id": "train_028065", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "3ca486007b978d0955ecda3197e72cc376dea17c"}
{"id": "train_028066", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "6843973934e949620fc86fecd326f190f341f455"}
{"id": "train_028067", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "1b7af7cf0a6e30eac12254b145a2875db1540a71"}
{"id": "train_028068", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "3deb6181b533b6eeaa370c9c6c71c8fd068243cf"}
{"id": "train_028069", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["agent", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "32ab6fa094f3c8e97417cddaa1401469149c270f"}
{"id": "train_028070", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "dc3a2d58f7fb0baae7063f1f00cef9ebb4d84a40"}
{"id": "train_028071", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "5c74e3e3abf721fd7308e9f84e4359f937e08ae9"}
{"id": "train_028072", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6b393496ca8fec8334ba2626e0b114c5f1741adf"}
{"id": "train_028073", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "38a0a897c1b5cca0297c87428d2df1cef5949a13"}
{"id": "train_028074", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "7d6f926dee313c357a51e8ecc5e532a91d49c8c8"}
{"id": "train_028075", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "50d0e2a0a4fca380203099e82c04651d203a66ca"}
{"id": "train_028076", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "4e5c3f4b46e0dec070e0983099c609f392f12194"}
{"id": "train_028077", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "03f83a6aba0fc6f0d805ac31819f9de0f47f31e5"}
{"id": "train_028078", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["evaluation", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "3acaab01339372c76fb87c7b243c9c491ea159ce"}
{"id": "train_028079", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "03d493f27febdc912abf2442ff337d76c658879d"}
{"id": "train_028080", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "791cd82dcd0fc44cd5015a432a5f0dda4d973e06"}
{"id": "train_028081", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "e1336aaceb8d877f92000eb83c7be96ed7a0c8d3"}
{"id": "train_028082", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "5de52ddf3194776a11dc0c21238af88edda93100"}
{"id": "train_028083", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ec4a3320cc836c3705f41075d4f60b0b8a15ce3e"}
{"id": "train_028084", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["coding", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "42fed26a2bc16656090e10c7a842be59cdbc7235"}
{"id": "train_028085", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "7b45d91dec99a46a5347c05b94259c93e7420a15"}
{"id": "train_028086", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "fd499a704f15acaae9798eb819997a4746fb0261"}
{"id": "train_028087", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "887fcde6e63be1e969c9e0467fe29436b1f685a1"}
{"id": "train_028088", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "79f985e3f3aca6ed482a17d1e0c4852bcabc9068"}
{"id": "train_028089", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "fea93faa173cf95a2713fcd3281d4a5fbd1c68db"}
{"id": "train_028090", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "d7ef673371e431b49ce9b90048e762a7aabda6ea"}
{"id": "train_028091", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "8f22470a826548c5a81399041c617a9a43c6a366"}
{"id": "train_028092", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "f2bb02baf63497ccf1a715caf6e714d78f05b6a8"}
{"id": "train_028093", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "448b29b035826ea187eabcb57b16751ff11f94f5"}
{"id": "train_028094", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["performance", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "b8a369f20d1c4785902350e72b99d62b185cb345"}
{"id": "train_028095", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "c85f6326d644e1cd631a5c1a6e180ad160a24729"}
{"id": "train_028096", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "06a6ccb24653a797011d51c3d9aca315b5237de0"}
{"id": "train_028097", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "291a62a253c0098699356a09cce1b517ca105ed5"}
{"id": "train_028098", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["research", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "d90cfc8842f397998353974b2f770ece667b8549"}
{"id": "train_028099", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["performance", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "8d8ffb04d14a343ab8983044b8fda0a81d4348d9"}
{"id": "train_028100", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a9a7c7ddceb90519c39ccc37d55606e490d9ccca"}
{"id": "train_028101", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "8df662f2ed01e1e609976348369e57ff037daca0"}
{"id": "train_028102", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "645592ced75782e6fc2dfe2b0435ea8f4fc52761"}
{"id": "train_028103", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "bec18459a62dc511897633145509f8c33d7dfc8b"}
{"id": "train_028104", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "cb9306181045051ccbabce3f3c9f1b6a8993f805"}
{"id": "train_028105", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "ef6c0df6bf76c24f4de879f28e258847d860ebf2"}
{"id": "train_028106", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "a640e72f0f3ae0c0e367461711c1d93546037eb1"}
{"id": "train_028107", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "4cf5a6d31512aa1bb85cb4d473975bdf7f9bc4d6"}
{"id": "train_028108", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "7dc51a48bd87ae4dcedd275851bfa2065e753df2"}
{"id": "train_028109", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "010cdb5863f4442e66c953c8cf42e7447ea5b898"}
{"id": "train_028110", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "00dd069a7b33b1fd230137d7ae6ca735752ae6f9"}
{"id": "train_028111", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8a289ae08e50f7f7957b169d3d7a6b08fb10f93b"}
{"id": "train_028112", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "00b5da9b17a91c4d170e3dffa9104e9628b662d5"}
{"id": "train_028113", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "baa1103dff753dedd73901a67f41d1d4e61af98a"}
{"id": "train_028114", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ffa6f9ba3819d99b7a707b743d20967af7c0e773"}
{"id": "train_028115", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "acf87143a5c9b01f57feca51ec4cee6eaece064f"}
{"id": "train_028116", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["agent", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "bddfb5559431d0e9e2c0eec66804dd06d156a155"}
{"id": "train_028117", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "22d3a91fac3eaf53668c4473de391042d9d73909"}
{"id": "train_028118", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "8f6edb883b07d9bf6b9153240a14a3f8dab4a699"}
{"id": "train_028119", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["security", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2b23455a5129f5a8c83c427a881daa3e008a67b7"}
{"id": "train_028120", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "0fa2a993a2436f6e557e0273d001253588f621d5"}
{"id": "train_028121", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c740ed73a689c3d127e65f6e87050ea1d1d5cfe0"}
{"id": "train_028122", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a2f3a485e8ddc57284a95288dee0145fffa9d456"}
{"id": "train_028123", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "4f55f4d876656593445686ecca258edab0a08817"}
{"id": "train_028124", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "787dd58d6332b6ce8db1adda40d50c67f92e7fe4"}
{"id": "train_028125", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["data", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "cf44506cf60f5bef7d20f8c5321b6ca284b7553a"}
{"id": "train_028126", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "a8a1db1dabfe472fc9e59a223768d258e7487b09"}
{"id": "train_028127", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "1cd08a829f6619f189e3cbb0880b151c8a1ba177"}
{"id": "train_028128", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "e0434091a1629283b9e562324e7ed72471e6cee0"}
{"id": "train_028129", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "33444c12f3ab335c1acc3b137673c1e1b9f20651"}
{"id": "train_028130", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "39efb3252eab5f546143cbc6d93819644235b5eb"}
{"id": "train_028131", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "9f1a90f8b523b3e6ccb4cb9380a044bfaf35747c"}
{"id": "train_028132", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b87230113028bf5b397879a749c5f36ea8dadd50"}
{"id": "train_028133", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2c639cf4364db01cb1eacf5aa506a5ff04fceaaa"}
{"id": "train_028134", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "46adc5ac26e7e1b2578650719d192969471f5b21"}
{"id": "train_028135", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "af45330e06236d5d9be8dfe5a34f0c0c699e0acc"}
{"id": "train_028136", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ff6294c040e7b04a1d62f61fd35bc626bba4f892"}
{"id": "train_028137", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "72c3e5e37df11a2850d034fd6663e6059d00d26a"}
{"id": "train_028138", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "ab3434b2981e0de2ca7ebe1cc1518cfd485898c0"}
{"id": "train_028139", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "810bf92b0c239a4158886d4ba7df591b9a869b39"}
{"id": "train_028140", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6ee8386353f54e2b6342e5e5b6b87a49e259564d"}
{"id": "train_028141", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "24ca5d86c85df8abaafc8ae80ac60eb7b73025f4"}
{"id": "train_028142", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["governance", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "e9ca80367b8d7f8bcfccaf345bad10a0d2e1d62e"}
{"id": "train_028143", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "29e07740cde3c0786cb2c3db4212b32f1b0a84dc"}
{"id": "train_028144", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "0d8adcc24b9a37346b340e2a3c176ad6253d97b4"}
{"id": "train_028145", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "b10157c4f63f5801357cc6721c65bb331d552546"}
{"id": "train_028146", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "a998bc8b814cffc83331a370f4e3dc3ecab7ccb2"}
{"id": "train_028147", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c375d9d8b8cd1bea89d48a9547fd61975d84ba36"}
{"id": "train_028148", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "698736215376c58c6903fa8cd13deeb6d7549aa9"}
{"id": "train_028149", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "268bbcff3fc7e7d43417528030467529d89a5832"}
{"id": "train_028150", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "e4c10beeffb6ac09a34873b45c221e897f96a46f"}
{"id": "train_028151", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "0be9997c572f7a723b7ef611d9c23d998687e798"}
{"id": "train_028152", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "30e2c0eeb29d2806ff60be6325a3adaf8ddc260e"}
{"id": "train_028153", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "60c5e8723ad54ec2a57d08e32061ff7bff6f8c80"}
{"id": "train_028154", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "c9706d3c07e1ed266efdddef9fa3969ee08f6e11"}
{"id": "train_028155", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "d5fb5c9d208f692cfa60d25feb941ed1eb81a7c7"}
{"id": "train_028156", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "11c1d09e71b4877f31d4c13e8a4521886d95c34b"}
{"id": "train_028157", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["coding", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "9a7582e9cc16165c37f5150d27877d25402371d2"}
{"id": "train_028158", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ecadab1cf90b696fb283bcd65c8532e6f509dbbc"}
{"id": "train_028159", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "4f128ace204a8041c6d993b44c42bbe39ec68c97"}
{"id": "train_028160", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "5a9db075a1627acf0a0a5820469452485915484e"}
{"id": "train_028161", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "91bdca75b617653dc21ecb741c4971008c993ecd"}
{"id": "train_028162", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "7aaf4e8f3ba23d8c46d0bb23414fe7f9711a295f"}
{"id": "train_028163", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["agent", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "0369319836a6ce66d48c21129f7f4012fd78464a"}
{"id": "train_028164", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "e38ddaa8a8cadeb54668087a9aea9b84c1b8b0b4"}
{"id": "train_028165", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["agent", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "fef23f2b66e70ce7b459a6cb36c8ea5b9ef9d6dd"}
{"id": "train_028166", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "c41c7abd248b5173a059655a8dba90d18df9cd74"}
{"id": "train_028167", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "62e9299251f123f47c09db9ff8ddc770f1491ae1"}
{"id": "train_028168", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "f338aa46d8a81ef9032bc14a792395847a85520c"}
{"id": "train_028169", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "db5e9ba8f950e6ede050ca5e3ecc76842e690572"}
{"id": "train_028170", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "bb5b38717670d43ab9d781b0a930afc639748eea"}
{"id": "train_028171", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "6aec9960f9292b6dab6d0bf1929e426906a5ddac"}
{"id": "train_028172", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "e649938b8e9ac8c2d0266093800028ee617012b1"}
{"id": "train_028173", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "f91ec38912e014f8d6499bcfd795ccc12397aaed"}
{"id": "train_028174", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["data", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "760981963b09cfe791bff811799f942a1b508718"}
{"id": "train_028175", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "88bfbaa8f00de2ce699d8572e95a1b22d1b6ef7a"}
{"id": "train_028176", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c34ba49513ee4d10a79926a199f2c7b89d0287e5"}
{"id": "train_028177", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a9989c6855eebaaa4163f22ffcdcab20ff29106c"}
{"id": "train_028178", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "dadac65941a0a6261592fe38c25ac631644bcc20"}
{"id": "train_028179", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "00de19cc09770f2579c1c7ed38beca6c22750754"}
{"id": "train_028180", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "757bd0c7503cc5dc9a2904549710ab565c905d42"}
{"id": "train_028181", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "36d1afb3beb8d5a5f273fdb10be6d38898b1143d"}
{"id": "train_028182", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "45868e7a50ca0ef0242888196af73c6c2b603c44"}
{"id": "train_028183", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "3298760228bbe4f94eabb0a212b34992166bfafa"}
{"id": "train_028184", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "b810319e8200a054e8a26329d859fadc2bf13d21"}
{"id": "train_028185", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c1efbe2f92d30167a1f1014516a7b786a6d9a214"}
{"id": "train_028186", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "086a3b93a3b42c7955ce99d50babfd05ca7f2ffe"}
{"id": "train_028187", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "d5c7924d3e6473170898185cdc2844761df1df33"}
{"id": "train_028188", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "01b77535e21b720d26e32a1d7e2d7f42dd6a376c"}
{"id": "train_028189", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "be6dce8fa9d465c97c4a7bb76255c00868bbcb6d"}
{"id": "train_028190", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "58e949f132f21de59a6722da0637f2bb5a771a44"}
{"id": "train_028191", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "f640b90c36bc45914db3165d7a05933d6c97a242"}
{"id": "train_028192", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "6c96ae8693d3194159112dc41997ca011bc9738f"}
{"id": "train_028193", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "142a9469ecfabd2488398f1ca013300abe73d502"}
{"id": "train_028194", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a901ff348b32ab1ad48a048ef2306ff654823b38"}
{"id": "train_028195", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0a67f24bb224a5737b8cd9f0d428145617f6202a"}
{"id": "train_028196", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "40a28373bbeaa7c28e4ff65a42f4323c5efaa9b3"}
{"id": "train_028197", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "afec30f18a9c74767f1860c531172b19f2786134"}
{"id": "train_028198", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c7e72ec558113a3f5036b1568ad35b4a3bb173f2"}
{"id": "train_028199", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "76c4bb021908a4076cfcf83585eefdc6369f2d2c"}
{"id": "train_028200", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "945c427e4050e6e19fad2ae971f78cbc62ac8f85"}
{"id": "train_028201", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "8ca4bd69f4687954163154ccbd7a20eefe431550"}
{"id": "train_028202", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d8d1b65121d8e476c3f77e70484bba43ea505693"}
{"id": "train_028203", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "ab71ce726379402850858163755a8c0d3ae0d504"}
{"id": "train_028204", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "319bce6183eb198233d06b8bb3e3d324dd4cf3c8"}
{"id": "train_028205", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "bd832b72be7a1f956430959e12430a818425e003"}
{"id": "train_028206", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "06e3719080f9908804ebc01594a316a637707011"}
{"id": "train_028207", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "45bb948b381a655dabef03521f4e689d1f04ab12"}
{"id": "train_028208", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "9e0c7f7daeb3203dc6c55d69c3b1238f9ba84c9f"}
{"id": "train_028209", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9997989455e6f9c7aebcaf710ab34dcd2e0e6023"}
{"id": "train_028210", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "871ed461cc539f29d9484b78ab8b96dacc10b45a"}
{"id": "train_028211", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["research", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "7ee534ca9c35ff8e731a4f1276e68d728d8d82bf"}
{"id": "train_028212", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "0347efc6da497ae5bfcea1e2cad81f61bce7c767"}
{"id": "train_028213", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "b1f2f1cd2d75da1863d81342ce0aefa9f0567c7a"}
{"id": "train_028214", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "bc2615067ad98920d7b99055f699fa989ec96dba"}
{"id": "train_028215", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "943e954340a048334c7c5ad96d37c201f8093ee0"}
{"id": "train_028216", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "6b696a97b899f8012d1becdc07ce8a139640a471"}
{"id": "train_028217", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "c83ea10916edd6952610ef405d2afe1082f5767b"}
{"id": "train_028218", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "8157cfeeeb35acc615c43e618242d9ed08ffa6ad"}
{"id": "train_028219", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a29bbacd754bb5404fbc0fd060dab93c2116e166"}
{"id": "train_028220", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "acc090d1986295945fbf280d86f0cc842f82b4c8"}
{"id": "train_028221", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "1251fdc3a169e0126017c07a5a5b6a96bca49034"}
{"id": "train_028222", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "70c70c0cf50020032c0f683c8e9772d154337f76"}
{"id": "train_028223", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["security", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "12a5309b554e15884a3ef84e67871d79e1e17082"}
{"id": "train_028224", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "af6a99e42332b625f573a8801853d3e3794b9869"}
{"id": "train_028225", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "1db80b992d72f70ad76eed57b28688992c114c67"}
{"id": "train_028226", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "4ab057e5ff30e5b9faf743eded7eea42c88c71f2"}
{"id": "train_028227", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "0ac85336095dbd141f9afddf3f16054784799e35"}
{"id": "train_028228", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "b99443d39edc9d8d6e9067341a9e4571e40c53b1"}
{"id": "train_028229", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6b756ecdadccfd91a4b1d462a222ecf42d8cf358"}
{"id": "train_028230", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "566e357c47b861a4ec027fc05152c5c703b5185b"}
{"id": "train_028231", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "17ad0c70f7e6d9cc808f45aaf77dbe61f7606e24"}
{"id": "train_028232", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["evaluation", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "ec657069a37562ac4a663f5aaf44d6029f762950"}
{"id": "train_028233", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "90c1f8cb01aa7b161aff81038acc9d1300c0e3f4"}
{"id": "train_028234", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "1e95a7ad23bdbd53b53122276acd98165beeb676"}
{"id": "train_028235", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "0caaefd06d1f113d6234fd73b8a953e09c738d78"}
{"id": "train_028236", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "77a4c791b629cb23ee06a6227b23a1bb533bcd7f"}
{"id": "train_028237", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["performance", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "3b2b45c81ca333b34f0f13075ad000f78ac388b4"}
{"id": "train_028238", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "b4a22ba91ed04cab73db50c2dae0b7e97528b492"}
{"id": "train_028239", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "bd56204533f53dd200aa1aac29854a5042259327"}
{"id": "train_028240", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "3a0c86c2a75fb51f6d1220fdb2929b4dc586aa0e"}
{"id": "train_028241", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "996f29722e1bb35dad9be3fa9a99375d27cbbd70"}
{"id": "train_028242", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "6ac83ecaa6a787f054e5dc95a7c7a62b1a6f2dce"}
{"id": "train_028243", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c593107b3db5fe3a94776f77f2d7c4eb48fce393"}
{"id": "train_028244", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c01641339297100d1914c7ac8d8644dee0a9d173"}
{"id": "train_028245", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2d688cd647436e0d6b324a449291dd13d1591721"}
{"id": "train_028246", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "24400ad10e1e150fa6e7261bfac5018a582b6e61"}
{"id": "train_028247", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "b6148328ab4fcf577c80cbd8eb1aea9b4e89ee9d"}
{"id": "train_028248", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "7c34ffc1307987364d467d3ceb7af9bc9000e6bc"}
{"id": "train_028249", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["coding", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "1a3c15f54befb62bc06dcf26b5bc5b15f0a6a2f1"}
{"id": "train_028250", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["governance", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "c4e8b8482179e775381d9108c21be514e63f4e1a"}
{"id": "train_028251", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "80a0546f71f32bd57794e14701435dc57e14283e"}
{"id": "train_028252", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b08b39c4d64d6b987472ad280f4bfd84e0b39caf"}
{"id": "train_028253", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "15886ea4b8fbce078762bde40bd2145ef1100a53"}
{"id": "train_028254", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "0ea63b7dd5d06d69ba839fb8817e688f13a0d8a9"}
{"id": "train_028255", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "64285b074e135203b687b3064567815a6d4f5b3a"}
{"id": "train_028256", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "f26c0ca24c76efbba6f891876bfc7cf6fbd1e302"}
{"id": "train_028257", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c4869bca68f9bc4c03a03c0415a17f49368a718f"}
{"id": "train_028258", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "8961118b61d6781ccde48162790ae04671879ee1"}
{"id": "train_028259", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["data", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ba0a9aafb7e6aa88d972e11284e58bdfa7e19128"}
{"id": "train_028260", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "58fbde1fc0495fc3821077f04594741cdda4487f"}
{"id": "train_028261", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "70766f104729592af956911c5f6253052d46c1b3"}
{"id": "train_028262", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ae8a2b51c331a10627ed35cbe3800f999eb95b14"}
{"id": "train_028263", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "64eb56f00f39929dcdcc9d18b7512c090f9d816c"}
{"id": "train_028264", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["research", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "74f84a016393056a0607b0295d184f9bd6d5f511"}
{"id": "train_028265", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "32d76c9b44e5781a70c072f57c489fb4cc454af3"}
{"id": "train_028266", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "4a5382b8f644ae6c75c6c6fc1a5343342a7c76ef"}
{"id": "train_028267", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "5d4461bbde3ad9f54c7d21cbd251f976556e6661"}
{"id": "train_028268", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "974c1db6eeb2e9386dee94b780ae18837af1d69e"}
{"id": "train_028269", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "317627cfec821b874177ec0b5da586602e49bd9a"}
{"id": "train_028270", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "224e9feccd9d53f9ee52389dab80241db8e322e2"}
{"id": "train_028271", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "20fdbf7eecf63a4125899a3f240e8754f1b735bc"}
{"id": "train_028272", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "b13152c6a4e1e1ee03bf9c6064832245afefd900"}
{"id": "train_028273", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "7b6f1dad868b62134a4095822d0fa468f10b137b"}
{"id": "train_028274", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "426dc30f1c4460139ad4a154a4681b4d99f0b1e4"}
{"id": "train_028275", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "11c4278893cfac78edc4eca69ad21425c69be431"}
{"id": "train_028276", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "4273ab17f6fef8dc29bc63710beae7a018325ab8"}
{"id": "train_028277", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f2b6170acb686c791a571573af9b373b118ea091"}
{"id": "train_028278", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a52aebd151756ba4a1f883a8b5c005cbc5ad86d8"}
{"id": "train_028279", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "35d1276f3737bb3a75c8deb4b07cbd82bfe24566"}
{"id": "train_028280", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "69e17cda61281c30604f0e6f49f9beafddee2682"}
{"id": "train_028281", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "aac6658904c8ff196fd9100b038f5d22994748b2"}
{"id": "train_028282", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "27e32673cf36ccc5b627afd77d5de16939095ed6"}
{"id": "train_028283", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "83e7e6491137f9138d380e74ac272ef6fde5ab0f"}
{"id": "train_028284", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "63862e15d340b3417705d7f6af8d5ee762fbf368"}
{"id": "train_028285", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["coding", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "f3e97f36bf66e896275a14db80c947c5c9af011f"}
{"id": "train_028286", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "54e51bdc5d8e9b6f2a40f68ea5d562935ffcf61f"}
{"id": "train_028287", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "bd46bc4f68fb995eaec63cb832307b236b01485d"}
{"id": "train_028288", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "3bf457af5a99e15d56f9becc5d0312aad03f1959"}
{"id": "train_028289", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c7a398fdf6f9aed23351d5ed57935fa01a2abb18"}
{"id": "train_028290", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "6fdb4a24ab4d476dc97bc913a155d0c5eb9e2127"}
{"id": "train_028291", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "665ed4b4e0b8742b231de0407160a2adbf2d4600"}
{"id": "train_028292", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "98e1d30bf96ce212d8466016c71d163f0653a92e"}
{"id": "train_028293", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "dd77de8b5d5cc8bae351375e54f50882c2fdba28"}
{"id": "train_028294", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "a582e6bb4ce2b542d8cfb76b45a96d001b14ecab"}
{"id": "train_028295", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "38ee0bfc97452c605701ab0a69b6c4b3578c080c"}
{"id": "train_028296", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c4639f0eab90e303cd908f95b975a8f1744f0593"}
{"id": "train_028297", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "9edffdab2dfe17bdea443c6b673ad105cb23ce6f"}
{"id": "train_028298", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "82a1f0f7da70e08efc6b60079216f213007a0d18"}
{"id": "train_028299", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "93d477de990f3af6a3cd30126431fc437e8823b3"}
{"id": "train_028300", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "21c7fcfeaa9b40840c6f11932d0a6c0e9816da37"}
{"id": "train_028301", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "47c44788dfe583927f292581df4759f329729a89"}
{"id": "train_028302", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "aadaf342210c5f5289f6f5cfece18660d1f0cf60"}
{"id": "train_028303", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "52c5f2273c5c6b3c29857afce1e964325f6ec6b2"}
{"id": "train_028304", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "9d02d0923e6907586c96aeae99cfdce52929353b"}
{"id": "train_028305", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "a44db2c6d3dbff7ac7d16a6e6751d2d5ba4bd322"}
{"id": "train_028306", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "6551a1e94492490f234a1adf7fb0f2dd8968c3b9"}
{"id": "train_028307", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ccba1212b00654d7230cbf089603bb64256079c5"}
{"id": "train_028308", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "d276faffe954c29108803d62d2ba57e6a1eda206"}
{"id": "train_028309", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "495009c58cac154858dbd4677015c4f9736da708"}
{"id": "train_028310", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "e7b7526900f5696eeecc0a3dfe13e28272204e8b"}
{"id": "train_028311", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "39707bc5e84a98352b84657ff08e85560790141d"}
{"id": "train_028312", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "498a9e7f483ba10db79fb3b6582e9b9be3c5864c"}
{"id": "train_028313", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "20664917d694702cbac088e52e1cbeeb02534b1f"}
{"id": "train_028314", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "014ec0132ac411a5c6b48137f6e084c8487835b9"}
{"id": "train_028315", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "83a68b46bc5d72038d27016d35361e78636764c0"}
{"id": "train_028316", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e50b372d2b3d03ce67e0de51da50f25a1e514041"}
{"id": "train_028317", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "01609e77fc966844d9c0f2a848a39a9252b23cbd"}
{"id": "train_028318", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c9931b6d6448cff41321da71dc75a0616320f9ca"}
{"id": "train_028319", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8c882f81bfe645520d5af3f4cfc79090796315b9"}
{"id": "train_028320", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "29796717a830bb2cee5113f7a8bd8afa2da9df7d"}
{"id": "train_028321", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "3b86152527e153e1bd92cad0c2e516043f8aa02b"}
{"id": "train_028322", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["coding", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "993843a922ccf6152344df669ec67423e3c7db8f"}
{"id": "train_028323", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "d8e5c3334c94fc98d1b6581c1dc0645ae4f66c3d"}
{"id": "train_028324", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "b78924a3ba0390ade31a8ba720edf9a595bb91f1"}
{"id": "train_028325", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "b7bce14fb4c89b7222598f0b7771a167ce595606"}
{"id": "train_028326", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "0342090df4fcbb1e5a8197ddc005de6c841bb4cd"}
{"id": "train_028327", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "27bf7efb4f7e47435294fc3726a69aacf8de654c"}
{"id": "train_028328", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e9d25e02af86d406fff4b10ca7429c772fb676c8"}
{"id": "train_028329", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "cdde60891cd6aa708572f6d1cb1320e6244086ac"}
{"id": "train_028330", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "5c9ad105a27ae3c9facc756a299bf27740fbb8b7"}
{"id": "train_028331", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "afbb44e24c93778d7b5ec566707d614d850a100c"}
{"id": "train_028332", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "50f9d570fe0fb825bdd7372f779d897017c0782b"}
{"id": "train_028333", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "7d9c30cac8f4adeda347ce5c63ec816290068438"}
{"id": "train_028334", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "8f6839fd1fa6bc60e9f5389015b8e081a19414e2"}
{"id": "train_028335", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "764658f543ac8bfbc88cc7d8637a3089c29b563d"}
{"id": "train_028336", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["data", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "af1804ca6bb96fd9966ab3d28dada72984bc622b"}
{"id": "train_028337", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "969533b1d0d500f9604600e382f3a339ed006bad"}
{"id": "train_028338", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "33126bbf81438c22677b22e60085f689d7214bf0"}
{"id": "train_028339", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "9202fefa9d29f3ab6f88eee8f1a4a7c699d51e2a"}
{"id": "train_028340", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "354083d148cc9dd93976843f62ef5d65f586acd2"}
{"id": "train_028341", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["performance", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "604cafc743f59f05210baf59004b520135128fe5"}
{"id": "train_028342", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "4a6203341199981659b24245a26f24f15d89d53c"}
{"id": "train_028343", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "66df9f989e1a5065439e599c7e060b000287c139"}
{"id": "train_028344", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "6ed063afb80d0da179d974c628701f67b2f91836"}
{"id": "train_028345", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "65146c364aac84c845a7576dd439674933146876"}
{"id": "train_028346", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b17217cedb67394c5bbdda6835e874e381df70ad"}
{"id": "train_028347", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "39c757fb9e79aed888e50311497400b05d3145ac"}
{"id": "train_028348", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "d1431c5d5d9bf218aee134dbe777af36796277b3"}
{"id": "train_028349", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "31c82ffb73e5baeb656bd3d763b754b6d004bc3b"}
{"id": "train_028350", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "3890b91520a52d4cb2584eab9859467588058b34"}
{"id": "train_028351", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "8374e7ff2d63e156023fb7e47b282ef9da0a5e94"}
{"id": "train_028352", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "6a028cc6a11c6d265cc9e7880cbec96c69173683"}
{"id": "train_028353", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "2cb4d05a3dd791bc6be4fd22d4285fd0923be01d"}
{"id": "train_028354", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "5baef62f6f9b193037ffd248a2a47bdb39b10ba7"}
{"id": "train_028355", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "41e8db3988372a4b4a55a67549e5e7005db220d9"}
{"id": "train_028356", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "1184e5a36150a07e5508fe418485cebdb32290e9"}
{"id": "train_028357", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "e0bc3472a783db47503a982343de1ca97d4749a4"}
{"id": "train_028358", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "5fe8601c0a4234683b92700ac784b8d4d87542eb"}
{"id": "train_028359", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "7264d42e56becd04ef0d270a2ef8754720ea8ca8"}
{"id": "train_028360", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "9e21d90da0959fd5050eccfa4940ef6323a0c5bd"}
{"id": "train_028361", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "e55e6c1d77db1a73d342fee5f76a96f11f114fee"}
{"id": "train_028362", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "96963cc582edfce1bd82e05810d6c12b94e135ca"}
{"id": "train_028363", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "bab53fb150b427254f9c2267e576925f5f5da88a"}
{"id": "train_028364", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "cc6df557a29efc50aedfb16732b1b4a57e069973"}
{"id": "train_028365", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "d0615cecb840db3dfec6e0ee0b41bdccaed422ae"}
{"id": "train_028366", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "3685e0ca8d7f9785bee43d638f6c4dcd1b6b418b"}
{"id": "train_028367", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "3614b71a194f1862537dc0ccf14fe9125083201f"}
{"id": "train_028368", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "27d70833adb7704ca92c5585c0d37ec9995d720b"}
{"id": "train_028369", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "f83bfaae89c51211b6cd7889edf734bc86e92715"}
{"id": "train_028370", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "716dceef4be76fb257bb3b3e9c007378580300e5"}
{"id": "train_028371", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "737b25f41e45e68d32a5659c05273a0facc5f42b"}
{"id": "train_028372", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "0f3fdd178f4f34ab1dd2d489aa0cfe7c657ae7bb"}
{"id": "train_028373", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "8121877b2b746c80f2605d3057b299ea0c8850dd"}
{"id": "train_028374", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "33c238c4be49a3e301c5a7d8e2fb282c714b696f"}
{"id": "train_028375", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b6f5615fafbf201c532e2e165a57e08c8adb8cf4"}
{"id": "train_028376", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["evaluation", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f7305f02c60485fd549d1be8c0660c451aa1b414"}
{"id": "train_028377", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ca4657b9ea66bda02f712529e5474bf83624f2a3"}
{"id": "train_028378", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "aaca6ef8735993247caf475c8df5fcb1fe434d9d"}
{"id": "train_028379", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "23b8cd57d7e38f1140a2395ce0da2e58ee9261de"}
{"id": "train_028380", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "b731a52997f62ffb43808082561a9eee726d7353"}
{"id": "train_028381", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["research", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "3cffa656bbc8560ad00b210a6f8f8a5a87ac472b"}
{"id": "train_028382", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "38d10b8fd72453d4569061c95dbf6776b3a3d119"}
{"id": "train_028383", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "894e7cb077537d9fa45ee43657be1161bcd4d3ef"}
{"id": "train_028384", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "1974b0beb7d5c2b7c84887679406e85149fab38c"}
{"id": "train_028385", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "847e4030a0d3b7a8cef1bf2034d71ef04671bd28"}
{"id": "train_028386", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "c51896a34bb932466b64124b427f7c067b3cf1ee"}
{"id": "train_028387", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "72e7ca0b393fcf81d7da4ea70d56c99a75459fa9"}
{"id": "train_028388", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f1ef284cc136beba9e639469bf2a95ee21a30e3b"}
{"id": "train_028389", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "ae533309dc1bbed3658f0c06c66902619bf0bca3"}
{"id": "train_028390", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "960dbc67c167704834089341c6b69802f7c806ea"}
{"id": "train_028391", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["performance", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "20ba8fda0e2931928b79254ad08ef3fe45e0f4fd"}
{"id": "train_028392", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "71e8a89e355d431e904ecd4e271c7ec13f938296"}
{"id": "train_028393", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "3fc523633a9f0fa79e7cc17c80a27fbbafb7b01a"}
{"id": "train_028394", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "025ebb27ac2fd8372f0821ef75d1d100b5fc172b"}
{"id": "train_028395", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "5193fd0711b69a1036604066bb04abce80a7afbb"}
{"id": "train_028396", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "aaf2bb1bd6678773333f842ebfd34d851a3ad47e"}
{"id": "train_028397", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "093e8b4a0f49948553a5b303d5d4b759f5ede253"}
{"id": "train_028398", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "c82e36b0df51f4546b0cc33c0258a1ed734c63bf"}
{"id": "train_028399", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "f2119e6f0486d5ff57d0f1719f6812942a324943"}
{"id": "train_028400", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "43e3257527cfdc7a5180a8f7fbdb9f58b772553b"}
{"id": "train_028401", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "f7c30f12e534e04e0ee40df73c7a2ce5ac871eba"}
{"id": "train_028402", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "e6ce236eac5383cd280343aa2ea0bfb5eab1a45f"}
{"id": "train_028403", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "967bc3bb6927b2c1e0d4e9b7dfd8f9adc27193b4"}
{"id": "train_028404", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "04e20fbc1ea1a0a66e3dd3fa17e51172430d7e47"}
{"id": "train_028405", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "bdf0414ec167a3afff51a645c6c8b5806a37cd6f"}
{"id": "train_028406", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "40f7228171eb11d09c8449f70db4029bee6895f9"}
{"id": "train_028407", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "4000ed3f632c4ffa7dd463de0828ceac2b29a361"}
{"id": "train_028408", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "1a75997f6689f1bf1a8af1b5d590ba35465595d8"}
{"id": "train_028409", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "404ea1aadb6c5791f4d69cdb13e66b2a87007106"}
{"id": "train_028410", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "5154ebbccb740857230cfc3e6ab0cb0f47576243"}
{"id": "train_028411", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ac146e156f21f9d96b16fdb1d13828ee3b5fb2c4"}
{"id": "train_028412", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "e72f29b2937b7eb2797676596f4000ef256e3b89"}
{"id": "train_028413", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "c47326ee0aad2508c650afdd0165cf59e260b6cf"}
{"id": "train_028414", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2117c4da758236600aae5bd37542b353887601bf"}
{"id": "train_028415", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["governance", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "05eccc38d997ef65cdd7a06fbc08c94d40de6f95"}
{"id": "train_028416", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "c4440c5096ae8b7d93ade5d35955f9865f456362"}
{"id": "train_028417", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "44ef2a537683213665afbe0f9b8cece796bee960"}
{"id": "train_028418", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["agent", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "51f83762301cb9369dce93a94a07d02a0b950820"}
{"id": "train_028419", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["data", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2bc3cf10648a9df1fa3b36f65b502fe59161431b"}
{"id": "train_028420", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "0844b77584bdb33384ea363176980a80753878d9"}
{"id": "train_028421", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "04b9e68f6a3efd26ae651b9e0620850c190035c2"}
{"id": "train_028422", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "d0998322162333741f6713f13ed7112f8c687c72"}
{"id": "train_028423", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "0964276e70e255b3712878dff8be681fcc1f70e6"}
{"id": "train_028424", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "71b4d3273d604ce16d2b2c4a1b22e9bed86985e0"}
{"id": "train_028425", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "6494a422452f104f3365c02debc942ada60eb0a7"}
{"id": "train_028426", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "816999fbd5acb3cd3ee53be165b1bd02ddd9b015"}
{"id": "train_028427", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "c3ddd0276fbddfb7f277871afbfe6bd65eecd2c4"}
{"id": "train_028428", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f8eb40a2a05c85395626382ceafa5dd3791861dd"}
{"id": "train_028429", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["security", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "24b8d83855fa9eefa45f902569e82ecf836d254e"}
{"id": "train_028430", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "6381b135475d359e89fb72d56a60ca45d0663fd6"}
{"id": "train_028431", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "177bc0df32a53fefe6fba2ba644bed87e73dbd35"}
{"id": "train_028432", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "4ee1e843e2b182046a58b6c842874b2e0fdec8cd"}
{"id": "train_028433", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a16efd61cbbc7a8189f03d4cb314e57a98bc306a"}
{"id": "train_028434", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "02561af755791af61118326abc59ef9c13e56643"}
{"id": "train_028435", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "7ab5d25ceacea14c2bdf4bda47f0a2bd1230d006"}
{"id": "train_028436", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["performance", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "d60e80d5a7e4f718ff47a51255a08ef1273acb1f"}
{"id": "train_028437", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f731a0bb49f2ac109dbc52e9343738bbdfa074ce"}
{"id": "train_028438", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "a2e3d3ee69dab9130f7242b1ef022de9d5cbcb5d"}
{"id": "train_028439", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "f5e645187a7ae1efccca15ea844eb7304c538887"}
{"id": "train_028440", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "3db824e048d7c7356048df5a6f16003407c689cc"}
{"id": "train_028441", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "0d3a26f9813cc27a0eec059b1a48ab657cde7335"}
{"id": "train_028442", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "4087806a7d3d2e26eecd7e62dcb98370cc166b96"}
{"id": "train_028443", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "e0c7fa608564e4718526d6279f5dff455ce91850"}
{"id": "train_028444", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "d023fa0645feb435ee48b7262902e51fc49b52c2"}
{"id": "train_028445", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "3ace4de40664c7cc8935610ce5311d22607a9968"}
{"id": "train_028446", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "b9eb9f4a426bab0e3a368c12b93a99b328b7c901"}
{"id": "train_028447", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "69801fa0859b018452ccb66ebe467c1106f8d645"}
{"id": "train_028448", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "4928e81846125193202cf0840293bf0c58ee82ea"}
{"id": "train_028449", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["evaluation", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "5b0b84538184c6170515ab576f703a4fa27afc11"}
{"id": "train_028450", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "d3a30eaed64ca69d7a4dc6a137551a63ff508024"}
{"id": "train_028451", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2baa9ce0ce3cf93027124b3b46cc87d4a020e3b0"}
{"id": "train_028452", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "3cb08d331ba43fb2f6dbe8461b74fefb7c1e6b1c"}
{"id": "train_028453", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "59441b43478d3cd512504a02bd56ed1c61078a0f"}
{"id": "train_028454", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c81ecfe17c0907b2e620c97b16f01327291ad7f3"}
{"id": "train_028455", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a68ea8b4e125dae98067c7491e9d8698a28cdb7b"}
{"id": "train_028456", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "af6cd5684563e8894dbebb1545ee880f0f4830f9"}
{"id": "train_028457", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "4e0e1fd5cf9419f6d2abdf5efd1a707f9a92bc2d"}
{"id": "train_028458", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "1f9a0021ebf055d1ada8ff8b48dd676dc9c4896d"}
{"id": "train_028459", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "f5142426e5c5e06ed9b4c2af4c6c344470f21042"}
{"id": "train_028460", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "8526d955e0f77a9d8a64108537f0e0e33fdb80c1"}
{"id": "train_028461", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["research", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9a39eb530c2fb14ded88901254b180d92046bd92"}
{"id": "train_028462", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "badf5e120b9a1beb2cd5b3f7c9085182b31c2cb9"}
{"id": "train_028463", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "243471facbe2e3d9712064994a48bc035b430ddd"}
{"id": "train_028464", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "ab852b2db3058cf21b48d51be264e74c06e2bc88"}
{"id": "train_028465", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2e376b493a22c1486a74ab5ab4cd1ed752642aa2"}
{"id": "train_028466", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "122aeff34979aaca88578640582983c02c4705e0"}
{"id": "train_028467", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "1bb2878ba6f9cfc4332fb009ac08c2d92bcaadb8"}
{"id": "train_028468", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "47477a7f93a995972a854aa456ec67f62f49e7c8"}
{"id": "train_028469", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "d8b2e05e5d3b240e4867dd90e44bb6afee4d49fe"}
{"id": "train_028470", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a57326aae59cc870a97b432517de58d32bcb9366"}
{"id": "train_028471", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "9c0b9477adc8665fd14b8786893622d14d66b641"}
{"id": "train_028472", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "d88bb76a2202a015dbcd8b8cfa9d088bc4172105"}
{"id": "train_028473", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2772d223a63a7da4a1a4cbde6efa7b8aeb5374c3"}
{"id": "train_028474", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "ccc1be1ad3905d5422993d0cc3440a18999eeb27"}
{"id": "train_028475", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "00347b40b709eca111ad6e4b65df0c0c8a1b2b81"}
{"id": "train_028476", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "1bc9ae1d3cd2e53bafd9091cb32a8e3b6976286a"}
{"id": "train_028477", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "c45d88795bec61609cae8fb8685df8f69a06b015"}
{"id": "train_028478", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f1e63a327ebab4e8acbc9740ca2317336e32d056"}
{"id": "train_028479", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "92145bd696130031f4d7029bf9be0caf9c95c3bf"}
{"id": "train_028480", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "9c8675dc0448ec7bb0f6f491dd9c6cefef5d76f2"}
{"id": "train_028481", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "534942d316ded56715745d6bf2380d152ce106c8"}
{"id": "train_028482", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "168fd0f33ee8484261c9254ffd0b2e21ea6ef6a4"}
{"id": "train_028483", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "e4d74d17d3abe8031766bf607feabde9001fd929"}
{"id": "train_028484", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "c5d625d4fb703210662d95c43face1b966e2d56b"}
{"id": "train_028485", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "ca8dd382d1e4ddce1f73890f116bc5a4185113d6"}
{"id": "train_028486", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["coding", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "134939947da988224fb61bfe172145afdde90a2c"}
{"id": "train_028487", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "3df936ea4aca6937a51f7a24d43df0a896f82feb"}
{"id": "train_028488", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "3e42cc0d698e30e784a53deea115544312b9c482"}
{"id": "train_028489", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "65a5e77374b984001aa1ed2abbb976e3915bcde4"}
{"id": "train_028490", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "0fa5a7c0ffa455bdc9917aad579cccae3b64c2e8"}
{"id": "train_028491", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "29b25f32e425cec18cb38ad032b85af06c9476b9"}
{"id": "train_028492", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "689e9eaaa3c65ae094f0f7def4fea51451f0a901"}
{"id": "train_028493", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "d1054c3eefa404d2ea0bb90d6b524d0320844a84"}
{"id": "train_028494", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "aaba2fbce369d579210c09d888d226e355c4ba23"}
{"id": "train_028495", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "3bf457af5a99e15d56f9becc5d0312aad03f1959"}
{"id": "train_028496", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d9cbbfa00abbb3a4e7507a0d00376cd58cb32624"}
{"id": "train_028497", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "a5de435469925a2a2a5d0553c8eb84b6cf7168db"}
{"id": "train_028498", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "13995636745d6bf61abfaf134aea8df17f2ea1d0"}
{"id": "train_028499", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "3332cb2f6d470424f2dda572b42edb68b7f2f983"}
{"id": "train_028500", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "25d90f6ba63dabf6eb44d55e836ef6281e09f1c4"}
{"id": "train_028501", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "bbe698404635430fc8d627d33f7ee1767009f3b6"}
{"id": "train_028502", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "cff606ee95bcb4fbaf9bdad46b8177bcfe6a3809"}
{"id": "train_028503", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "6fa1f0717cac5dec9dec056a01b92087f9d22d6c"}
{"id": "train_028504", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "39a9198ea56b9292170825a551455bec3c9c284c"}
{"id": "train_028505", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "1f958ab19a16c953e8fdf5b7c3fc6b3671e68d70"}
{"id": "train_028506", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "59251769ecac3e675a043a2db1f054bfa9257437"}
{"id": "train_028507", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "52bc6dbe3cc587ba4ae69f138edd5f3b671b9a6b"}
{"id": "train_028508", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "7a54ca524c6db25f6be774036e6f7737d20e48c7"}
{"id": "train_028509", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "68ac3113d6cdb75e06a2dec120b5696f5bbe4d68"}
{"id": "train_028510", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["coding", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "7c21f875d4fd6f65bccddbf67ef74b7662c6a050"}
{"id": "train_028511", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "85011aa0fec75823fab78aac1f37a8577de9b354"}
{"id": "train_028512", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "50e0416b62dcab61944c9b00d6ac2d08d20d7735"}
{"id": "train_028513", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "8809ee7a253ed82ee488bc84799d5e5efd10d5ac"}
{"id": "train_028514", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "173e439712a4eaafe7fc553c83fde12224219a39"}
{"id": "train_028515", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "4b2868d4853879aea1373bf58ed1360df3cc384c"}
{"id": "train_028516", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "b35a9c716db24a5a8a50919e10a0c6b731cf5a78"}
{"id": "train_028517", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["evaluation", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "0f27ea8fdf097a0fb4d1da70e8cbf4b53e7a26ea"}
{"id": "train_028518", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "2a28a5bdf4567be3100c8d063092623228611b68"}
{"id": "train_028519", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "249a528c01a25d7071321c88079e6a3cace428c1"}
{"id": "train_028520", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "98d3804af86dda8fd592ead143610577a8eef4dd"}
{"id": "train_028521", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "cb2136ccdce9db0aec891e05aca32004627c3c94"}
{"id": "train_028522", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "1c7388c4977c3c72d07279c2e864eb410ee31318"}
{"id": "train_028523", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "f785e2a1f58bff737d815a1380f492c651a6c1c3"}
{"id": "train_028524", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "8657a6d485e0b3f1a5ac7a8f7d8016aadcfc802c"}
{"id": "train_028525", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "a42afc121f09f2419122a3c5b891e90a0261f22f"}
{"id": "train_028526", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "44f05f944843e3d1cdc3420f1a98fe3dc101ae30"}
{"id": "train_028527", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "178baa7f5a245ca8274660353036422205f21f51"}
{"id": "train_028528", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b88b44301a1b645a58c342a8ca14fcf69a519148"}
{"id": "train_028529", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "3eff1572354ea78aa37096a6825d7c488eedba0f"}
{"id": "train_028530", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "740687816497338ed7edf8a56ad138c736798364"}
{"id": "train_028531", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "91b26a41443870e5bc894188171d4e4fe4a0331d"}
{"id": "train_028532", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "499673cf76040af3879cc8387f765a0c9253807f"}
{"id": "train_028533", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "fd12de49bd960f0cce55344e72bf94393f61c739"}
{"id": "train_028534", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c3bfddf016b2c9f5c2633acbcc5121ce67f13864"}
{"id": "train_028535", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f18410317464e1b58b7d50627cb98f8e52186190"}
{"id": "train_028536", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0ea63b7dd5d06d69ba839fb8817e688f13a0d8a9"}
{"id": "train_028537", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "fa680f76e52bfe2d80ccc0a0cbe973fa30a43a21"}
{"id": "train_028538", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "dcbb4246032b474761d8961c44f8e89aba8ff304"}
{"id": "train_028539", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "0ec788ac0040599ec7aa6192d93a9b9de24127fe"}
{"id": "train_028540", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "3dab7a988ee8c839d5aff3e8f267c663b7d14d55"}
{"id": "train_028541", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a4257b3a02115b07930bf966fe788c037b620011"}
{"id": "train_028542", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["data", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "bee6fc8789f2f2f8a13fc88ff6d5307ca29d5638"}
{"id": "train_028543", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "0ad38ea133e64d531f78ba1e0127e79de43adb58"}
{"id": "train_028544", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ab54060901754270eb4930d2f0d9ffc67130772c"}
{"id": "train_028545", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "5079eaecf65cf14ea534b08ed07eda5738ea30c0"}
{"id": "train_028546", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f526b3e1b4f69e84e86288bf06e12b468df2b4f3"}
{"id": "train_028547", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "cc29d191909b1333fae5e465fc51b00487dbf7aa"}
{"id": "train_028548", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["performance", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "5ee2f63f476091c3698d4882d180f25cbe9f6b15"}
{"id": "train_028549", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "807bc5165677cee765119da2ce449678f1e727e4"}
{"id": "train_028550", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "53a9e95634b976f4f629317af4bc158628ebc366"}
{"id": "train_028551", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "ec01455f842684f3f97f60f43ad7af923548677b"}
{"id": "train_028552", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "a43586314b27251f7cce36b4b7c9b918bbaed1d5"}
{"id": "train_028553", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "4857e3007239aed0f5a76c2e2e34a9cb9dc34cb0"}
{"id": "train_028554", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "910d60007108eab9f9716ae03580f023ff2f06cb"}
{"id": "train_028555", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "90e64ebec8e27849dc0876854f2d87827899bd88"}
{"id": "train_028556", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "a1cd2b18be4ef6ddc79fc5003dd09e9516ae4bc3"}
{"id": "train_028557", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "ba615a90ca688bbd0cca2b78575607a012eddd92"}
{"id": "train_028558", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "6b99761a761e26816caa2f42f011de80a16be14e"}
{"id": "train_028559", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "ea79651e8dcf292f82b80b7a0dfa6295859a8de7"}
{"id": "train_028560", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c8f5385d27ab45ac378e30154585752b62d5f336"}
{"id": "train_028561", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "2d614e6292518537063dcb30c6c829178de21dae"}
{"id": "train_028562", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "63f883661575f9a44aff1bc65c29f09879c8cf3d"}
{"id": "train_028563", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "7da8ca702e988d7ba4ee703b06f4256b367f728b"}
{"id": "train_028564", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "3b3b260f18f31901f21e35c805e72acd1c167f92"}
{"id": "train_028565", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "1a3c6deef942d76df2cb22ce0c4d9afbe8c3dcec"}
{"id": "train_028566", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "bff68e182b3562013e3af90ba90dbf239c044e65"}
{"id": "train_028567", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["coding", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "62136a0f4596ac4d56c434507d2d64a206a0deea"}
{"id": "train_028568", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "95436c3a24f8b2e82249525082bb88297cd67ee4"}
{"id": "train_028569", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "ecc80c7dcf06b1cb4d472bf1126df6732296bbba"}
{"id": "train_028570", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "127ef945b7a1e3544cc2b8d94e8f610ac8aaecd0"}
{"id": "train_028571", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2ed5908b71d4757b2ad56b12cd75418d08221f94"}
{"id": "train_028572", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "aab8eda2415f049e65f2f08339e45cda998c35a0"}
{"id": "train_028573", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "4ce6d79f1233c2e123adef635731e17ad28b8300"}
{"id": "train_028574", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "19630d7124f7403435bd6b172e8bdb24366383eb"}
{"id": "train_028575", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "30935cfc7ce847c791927a578f1966a4637689c1"}
{"id": "train_028576", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["performance", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "3712918e80ed5f05e1aa3517d285cd269e38cf5f"}
{"id": "train_028577", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "38e23e14fadbafe024400dfad7712523de3c6566"}
{"id": "train_028578", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "835882b3069529e5cb8476bc236a0291e901ec17"}
{"id": "train_028579", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "7dbdef615cb2e994d20af6d5591d3f953da0c56d"}
{"id": "train_028580", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "9a11c8c62e74b0553e2bd2f2d06d1fa478730fbd"}
{"id": "train_028581", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6a511b6d07ee7835a38e13e8ae636a050eaa9322"}
{"id": "train_028582", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "c8d65a9384a6238e4de2d5582b7f6aa6cf182781"}
{"id": "train_028583", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "6ab66490760f6f9d3b88edd5b8b25d55b38f4d5d"}
{"id": "train_028584", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "eda6a7c75e2cd121bd4b91ece4b6a6b68e8c22bc"}
{"id": "train_028585", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f10afed923bcfe9665240e0a41f3d09d813f902e"}
{"id": "train_028586", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["evaluation", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "71da6c734d2e30f90444263a61c826890de6ee82"}
{"id": "train_028587", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "90759ac86be26b6802a420206b9bc6880f60716a"}
{"id": "train_028588", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ae830ea1e6dbdf806cb7c164581a2bd0cff646b8"}
{"id": "train_028589", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "03dff8809e3af74f7b588652fb708470d07b5485"}
{"id": "train_028590", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "851833a4894c07e329fad289ed3b3568dc200e47"}
{"id": "train_028591", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "fc79191c16af4b65e49d1db5a8cc23df8f12af64"}
{"id": "train_028592", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "1b94a7850cce5419e1143edf8d609cdeff1a5d1d"}
{"id": "train_028593", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "0174d9c6fa429b8e0c670ec959e27f79ff95b8aa"}
{"id": "train_028594", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "676c9a8a21528142174f7325a299bcb6f600f005"}
{"id": "train_028595", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "76fb2d81f57fb22a4a671821fa814ecd411504fc"}
{"id": "train_028596", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "4c120672c059e028a119ed357fdb737650e6088a"}
{"id": "train_028597", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "fdc73edf8210b9aa37dda4f24caee0379cc5bc64"}
{"id": "train_028598", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "c9a0f481cef5921900e7a85159d863f5666ff5a1"}
{"id": "train_028599", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "cbff56cf82b7fde9c92ba331d0000fe8126f9e82"}
{"id": "train_028600", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "8885c729d85cb09de9e5d967635847e53d45175e"}
{"id": "train_028601", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "7ae162c0975474f0b220f572da9393bc14ff5b30"}
{"id": "train_028602", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "839844a7d7c735d72b90c406b2881d15cbdd89a4"}
{"id": "train_028603", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f907d9fa2399b051137a7123f4002dfb9ab24407"}
{"id": "train_028604", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c8d94c96b8b55711f8d0fc248c66edcebeb9cdf5"}
{"id": "train_028605", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "d97db548be99a559938a715f32d03f4a0312e204"}
{"id": "train_028606", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a79db43995cfe1d10e381fdce56640b34634abc9"}
{"id": "train_028607", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "9624f00b96a51342b0e792f5a133b47297bd6ecb"}
{"id": "train_028608", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "47258325d5c47cf93e724d74f77af5ac3811cc1c"}
{"id": "train_028609", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "9024aa05bc636f4de8e954faa964b2c4ad9499f5"}
{"id": "train_028610", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "35d00ee9b3f57c3af35d3edc3e76605180b9b265"}
{"id": "train_028611", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["data", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c0c8c8e8e4902f69ac1104111d77db7e3803e67f"}
{"id": "train_028612", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b0816926d3a12abb764f63e4134e41ea32f170d3"}
{"id": "train_028613", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ed3c000005998df0d397a64c9463619e6e2e345e"}
{"id": "train_028614", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "40e8e638ac4d3d46f33b6affcad558288192d327"}
{"id": "train_028615", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "aa1fa62b4db1a0ac818de8ace5ad33660bcdd111"}
{"id": "train_028616", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "e7cae698e2bedf062a59dcd2b0654158adf2c19a"}
{"id": "train_028617", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "31037750d913cf1f2eb372f7c3dc6a5a1c695562"}
{"id": "train_028618", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "49475817bde0b3bd7e408f99dc21debaf1ed6d50"}
{"id": "train_028619", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "19effca63d3dd31129b8720f54134e5b8cb55715"}
{"id": "train_028620", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "bbf1e9d1623e274ad1405a22faa11678b515371f"}
{"id": "train_028621", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "b00eb6e1ccf732628eccaf63d7b225dab7a12bc4"}
{"id": "train_028622", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "e0407a3abb26065ac1649e4b67efe2702c853e4f"}
{"id": "train_028623", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "d4b94e6725471a0cdb5b14c6e235e3885872eba5"}
{"id": "train_028624", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "f6e4a72098a6114941ab4e786947db94453edd69"}
{"id": "train_028625", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "29e87f34e6d1af7b021c039db9539951da60e68b"}
{"id": "train_028626", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "cd4a8e2cb502832ff6eb7908e9abdad946cf3fb4"}
{"id": "train_028627", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "5d5ff02b112401e35111ffeb095c41376fecdb71"}
{"id": "train_028628", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "30e2c0eeb29d2806ff60be6325a3adaf8ddc260e"}
{"id": "train_028629", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2978319f4ec6109adc3960783f5ec5767176f53d"}
{"id": "train_028630", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f97ea717caa30943a7bded196b55b56f9336c82c"}
{"id": "train_028631", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "3ce2398c9c92b7d4aa1c431fc9a564df2d328819"}
{"id": "train_028632", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "be08ed6bd0c1673c35cf476d7170cf6f5437229d"}
{"id": "train_028633", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "3f17779fc50cf8a460906852f72cf52d539e6019"}
{"id": "train_028634", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "9f51547a54c300e78f3576364e88cd5f297b22e5"}
{"id": "train_028635", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "c8cd176968f4db9a6aca6b7bc73336dff8ddc9b6"}
{"id": "train_028636", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "1be1c8cf54e58c051e7e74f64b8b3bfdf38ca106"}
{"id": "train_028637", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "608de79249dcb385ccfe203ee4bd2197be646b4d"}
{"id": "train_028638", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "4f366ead75b6b3a87bf636856ca0c4a494fab8ec"}
{"id": "train_028639", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "21de3803372f3c274c205088451db76a1bda5f5e"}
{"id": "train_028640", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2165f5d9e093d4f72a661d283b713a3591525ef1"}
{"id": "train_028641", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a05a3280fb14ed280de5e5cca67d6f95dea4d65c"}
{"id": "train_028642", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "b599cf079314b0f278592a7e9a726cc31934ca64"}
{"id": "train_028643", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "dea6f93a71c3ab79357fab396e4629b94ab2df38"}
{"id": "train_028644", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "f30d16c4c6b23bec248b600c4057e5afa74383b4"}
{"id": "train_028645", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "cf994d567be2c159a084fcaf5a7d76d1534c1fc9"}
{"id": "train_028646", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "ab5cf94eb459e80adcf7c4ad4a6916a57ea6d010"}
{"id": "train_028647", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "602d4eec52ae6ece7396137c415ddc51dbe89de2"}
{"id": "train_028648", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "2ba8e41b4fb33dd25837fd82a8bda7eaf6684a24"}
{"id": "train_028649", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9c0b9477adc8665fd14b8786893622d14d66b641"}
{"id": "train_028650", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "98fa5353bf2b2de9c3006907a0a1799e2a0b45c7"}
{"id": "train_028651", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "66fae0a587584e14e6dad59a2a0abcbe4096c2cc"}
{"id": "train_028652", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "097f1b175cc7aac816120becbde62f70f8909fa4"}
{"id": "train_028653", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "61586da759dbc4d813c2309e231d008eb52088b9"}
{"id": "train_028654", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "09bb1166de5a5e98d1c62100ffdffba31a4ceb18"}
{"id": "train_028655", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "89bba04ec4415a551e1266e29944783cfdfc884b"}
{"id": "train_028656", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "1e6405bca8aadcf0d23bfab7f6a5567604004e8a"}
{"id": "train_028657", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "55d6393bc9b3fdb453f87cf9effe84526a989932"}
{"id": "train_028658", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "658f4bc4ba68f06da57da3783a753a6713ed9417"}
{"id": "train_028659", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "2b420a87d6019bb8355c85b5f6f9d10428f731d0"}
{"id": "train_028660", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "618eea9862917e056eb8bdbc902ed322a2454225"}
{"id": "train_028661", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "b958ee50a212f70e3eb8b415ef8ec7099d2781c7"}
{"id": "train_028662", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "e93b5d5d4cb6656d91789da7ac7019c4aba5eeab"}
{"id": "train_028663", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "25d847b4e644739d7fc9269ba8ffd7e256369e87"}
{"id": "train_028664", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "55fc7b10fa0aa8a732d6be59d4d5f8412ca35734"}
{"id": "train_028665", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "92c3997aa33603c9af22e994668eac65fd6f6930"}
{"id": "train_028666", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f0de0ed3b715691b31ee672ffa44012295a14bda"}
{"id": "train_028667", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "90a6f4b85fb3f300547bbc75cdfa8c6a18c83f4a"}
{"id": "train_028668", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "1257341cbe26cdf5a9bb74498477edadac1778e6"}
{"id": "train_028669", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "f5dbb675c062aae94045c852a386ffd55b78c3cc"}
{"id": "train_028670", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "ce2c8bc662081d5320b60a0b53799c4910f09cf7"}
{"id": "train_028671", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "bd7a34d47c94cd17cc59fbad338c0cb514860607"}
{"id": "train_028672", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "53085934620a280679b11e8c51797ad64b0a2cbb"}
{"id": "train_028673", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "85d730c43b7106a11342bd1db02a945249d5a671"}
{"id": "train_028674", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "7bcedf4228cd0a5b721b60d8da8721976c831d16"}
{"id": "train_028675", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "fc655e590701e44d30e92eaa8047eaa46bdd92ed"}
{"id": "train_028676", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "adc954d11411e97725fdf91972ad670dd5996b54"}
{"id": "train_028677", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "42e83934894b07d2d61d2011c810b872ae43b620"}
{"id": "train_028678", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d36be2b6829e9d3fc2e7bfe8e899be99b667fdd3"}
{"id": "train_028679", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "3fbbdac39644e5186ef1131ef21cad08332d196e"}
{"id": "train_028680", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a24a0079d7c7b48da351c67ee247beda5ad56a44"}
{"id": "train_028681", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "64bed467e608ac3700b8af699ac704edb93a6fa7"}
{"id": "train_028682", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ac59ff758f563480600e2d00879c89367ed5e5d6"}
{"id": "train_028683", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "5690e89c6e183d927f82c82b93f3504625feb7a4"}
{"id": "train_028684", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "670482ad4257ee4b3867e2900363359d08cdd025"}
{"id": "train_028685", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "861b73950b0fc596e25e0e6d34fef07e4a7894a1"}
{"id": "train_028686", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "fdece1fcd2d38ce45372d27ae64b3b7101cfe39e"}
{"id": "train_028687", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "cfc4db1b593f48c405d0df59d5cf838c2356d702"}
{"id": "train_028688", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "eddda400855a65a2cada065c1b4604a9a7fd52a5"}
{"id": "train_028689", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "92677ead46be9f14ba52e56f68e990a6d59662af"}
{"id": "train_028690", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "f904954bcc31b7148fc4236928b531111a565ad8"}
{"id": "train_028691", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "a6a9ad55ef380e8959ffc1bb53d65953f37fa736"}
{"id": "train_028692", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "645013386f862c2ae542f049cdcf94d7d0655a26"}
{"id": "train_028693", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "faf3a368ec906d01e74752a748f0ff32dd3a961b"}
{"id": "train_028694", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["security", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "daae77428df578ad5cfc88980c6fb37917e652dc"}
{"id": "train_028695", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "0b57c420da0628deaa67478f5c91864f6f8a3844"}
{"id": "train_028696", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "790bd047af69816a4081b718fcf71ea55ae24621"}
{"id": "train_028697", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "7e46d3934818a3a1e8cf939986f1c5bf7704a0e4"}
{"id": "train_028698", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2f0d7f7faa142db675e254b0e00a8622337e39f3"}
{"id": "train_028699", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "c951e5c2ecb0e21f751fd9bb21331ab52eaf61c9"}
{"id": "train_028700", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2731b1cd75a9b80d3f4cac06ce90f99d7616b01e"}
{"id": "train_028701", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["governance", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b660db621d1cc63867939575e15c1d6c41732dc8"}
{"id": "train_028702", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "80487feef9b272b7f8aac9330a068c2c2435addb"}
{"id": "train_028703", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "3b8b7d4ddf1e51c69d383b39ee2e1b0955aabe07"}
{"id": "train_028704", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "0d863454aaa613ddd123d5fc683d3104b6e51615"}
{"id": "train_028705", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "262c27b6ae0a18bd3e5400bbac494a6b9fb16567"}
{"id": "train_028706", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["data", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "bc5b1cd542f8638bb38c1ad255e8699d8611d51d"}
{"id": "train_028707", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "9143af4c489cc91732db0178ee90d0fbde5a9b2c"}
{"id": "train_028708", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "8ce48b8ec65ae1c25a2f7246564eef11be8a2872"}
{"id": "train_028709", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "d8ed42d20b6d5a49f54f452118190d3acc198f5c"}
{"id": "train_028710", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["data", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "9e3c84bb480a1f2e62a81e03a56fe48497b3f63b"}
{"id": "train_028711", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "53099685e4945557267f794bf072ce230e3f43d8"}
{"id": "train_028712", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "b2de401305e138670eb19bbc55af752b7d02c892"}
{"id": "train_028713", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "780f2a651740ee57015068fbd46c96f28b5db811"}
{"id": "train_028714", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "22dd52b7f156adb3abc3a742809e6a5c187c3e37"}
{"id": "train_028715", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "aa9bed336083a71cd5a8d79fedbcf5cc8ecd22be"}
{"id": "train_028716", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "04e20fbc1ea1a0a66e3dd3fa17e51172430d7e47"}
{"id": "train_028717", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "69ae35b4339d95d6b975410497e1096fa012d13e"}
{"id": "train_028718", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["data", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "cc2a006289a115c40dd8acbc9a77a19b0c8ac63f"}
{"id": "train_028719", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2e842c9adfd0ff4fb35e4d60d45b2d11294f8499"}
{"id": "train_028720", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "fda08678becf4b139fe6f2f9c24f7f8ed3b4eed1"}
{"id": "train_028721", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "2219f274e784be672a18e9705f9fe138f7969530"}
{"id": "train_028722", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "02793cb6cce99399edee0f399a320117909fc699"}
{"id": "train_028723", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "bad5afdf3f0741233680c5af03d76bf90614326b"}
{"id": "train_028724", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "e6adc09667a74df8ae9766f0a2350d4401e53ac9"}
{"id": "train_028725", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "20ce001c2fcfaa4e55bc2f37c63f548bdb1da34e"}
{"id": "train_028726", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "cb045b3f32eb71efd938b3db3d001ff6067b2b9c"}
{"id": "train_028727", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "d03c1b48befd846a7d44148eb0b3dc55a85041cb"}
{"id": "train_028728", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "de43703b0357be6f4f566ca3c9c4b9b1d37be96c"}
{"id": "train_028729", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "7b72e195903a705eace20b82ef4748a712da0fd0"}
{"id": "train_028730", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "e9c61b5e62d470154670488108ed77e83e8d4e71"}
{"id": "train_028731", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "72958bc8e8a3bc38ae3fd4e27233d402d6e306a2"}
{"id": "train_028732", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "fc60df4ba04f526d8e2dadd22d5fc8161c6645fb"}
{"id": "train_028733", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "f5ba595ba17a588d39017f76f7619e01ca760f75"}
{"id": "train_028734", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "98df063d6e6d2239774ee1a742f26c8ce3468c7c"}
{"id": "train_028735", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "3a0c86c2a75fb51f6d1220fdb2929b4dc586aa0e"}
{"id": "train_028736", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "e43cdc8c99e8adffdc42dfa59c71bf45a7a92182"}
{"id": "train_028737", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "1ad4b8d264c4427a4322afce6f5e2367560d177d"}
{"id": "train_028738", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a7c05b303ffc44f003767010f25119c5de56be29"}
{"id": "train_028739", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a7135f87034ec0b24418e09ca8683dfdcb89a370"}
{"id": "train_028740", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a4b85ffd4a7437392f80d3dad234c28a046ec328"}
{"id": "train_028741", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "275cd483294cef0ddeff0bb5d8cb38a6995b0391"}
{"id": "train_028742", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "47cbb81512dafa3baded5d011f045873740b984f"}
{"id": "train_028743", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "e5cbfe41999bc3740849c018870dfb6136298a99"}
{"id": "train_028744", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c4023bda6e0b9090701791164e600733e8f434be"}
{"id": "train_028745", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "45d6b5bc1fb49829fb601943e6910e54b67b18a5"}
{"id": "train_028746", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "678acb45823105ce0bb8e85416aef939cc71d488"}
{"id": "train_028747", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["performance", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e902ba1b0f2eb66d8c1704baedecc84956812a91"}
{"id": "train_028748", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["coding", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "1225e31973272eb312af4f33bfdbea88d8eb01b1"}
{"id": "train_028749", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a7042edfd293ba84a7e31c6957834085a1f2a72b"}
{"id": "train_028750", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "f6693b65d017c6eefb73b12724238ecaf8762234"}
{"id": "train_028751", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ab5cf94eb459e80adcf7c4ad4a6916a57ea6d010"}
{"id": "train_028752", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "21a1b021fb0b63d864376d7746fd8d6c5cfe3d8d"}
{"id": "train_028753", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "8231ed40c212f25b1f3079c057477b52576c1fd7"}
{"id": "train_028754", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "3042a6e82ef2325163d55f2bba475f3f11d8a3c3"}
{"id": "train_028755", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ca003ebce55b8150c68d2cac1ece8e4831dd16cc"}
{"id": "train_028756", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "30564df5c950ee325354d1ea5ad2672cc82caed4"}
{"id": "train_028757", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "317172225dd7ab7f8af47aabfe8e951f792b280a"}
{"id": "train_028758", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["evaluation", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "e156fac250e4e91d405161d4d943c3303b7dd1f8"}
{"id": "train_028759", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "15e6e179d6be13cbf1dd2c8dba7f0bd0fbc3807f"}
{"id": "train_028760", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c47ef460361f44678a3dccac212b348ff2f09daa"}
{"id": "train_028761", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "e695e2971f7f617a4277a652742f47beb7f9faf2"}
{"id": "train_028762", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "8c09f1d6980098168f57c45b6b8418a633b5f204"}
{"id": "train_028763", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f078cfcec33b3624d105cbd880d8f2f686198973"}
{"id": "train_028764", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "4a1eb54c602015619f07d4397fa639978863546d"}
{"id": "train_028765", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "630ad727f77dad81b325298228570843025d23e9"}
{"id": "train_028766", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "df8991d850f9093c9b00323c892d02c050d33673"}
{"id": "train_028767", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "f513addb24db76c1888aa36a095d57b80b38f192"}
{"id": "train_028768", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "fa996613ca411e20eaeb2ddd4a8d141791eafe80"}
{"id": "train_028769", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "9e06b8376319825fd8269e2839e644c3160fd9c7"}
{"id": "train_028770", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "69c624b1e45658cb04eb8801d613386d72be7683"}
{"id": "train_028771", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "9d71bb0a6c26cfb83af2f1c5f4dd9d9ab1506adb"}
{"id": "train_028772", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "fcd2aa462d81e2ede7bb8d2d47159fa0c0b5758d"}
{"id": "train_028773", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "e21b7ce56aa6a51729a1c61ddaa8fc2e7cc51bb8"}
{"id": "train_028774", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0f21f63fa6b5c774decc2f3aeda671d493b86588"}
{"id": "train_028775", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ab54060901754270eb4930d2f0d9ffc67130772c"}
{"id": "train_028776", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "9c176affa456d13ddc14028b11e60f48f94cfba1"}
{"id": "train_028777", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "7f049c7183f2750c2e96900976b2abc4b51299e0"}
{"id": "train_028778", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "7d39dc9ee7905865d0bdf9b136b569ae91683507"}
{"id": "train_028779", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "69611e544692ce9c8471e58540f240b7716d4298"}
{"id": "train_028780", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "324527ceab358ec5ca7f59d71ee5be87b3c81739"}
{"id": "train_028781", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "d19b11a64bb699073f77dafb6535e79ee37af274"}
{"id": "train_028782", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "175a7e3f0337c3aadafca5f26c0bc93e814b9226"}
{"id": "train_028783", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "4c4b66914abfdd732bcf425a0af1021cba40520c"}
{"id": "train_028784", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "26926dbed6352d1f4fa2afb69c1d359956476504"}
{"id": "train_028785", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "519ade1f62b8662137f836c5b6e2cfc961b5c580"}
{"id": "train_028786", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e34a8705e211a56eb81d8e6b34c573144122c59c"}
{"id": "train_028787", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "8d45ee23d46bcbca05a155d295992ddb3722b1db"}
{"id": "train_028788", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "85c4c73491d08f834adf309515375e203e83b3fa"}
{"id": "train_028789", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "7123da6b4374619aba1b2c3c593f69dcee5f34a2"}
{"id": "train_028790", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9b1ee3e4e2bc7d14450b5a16735dfe9846d1d696"}
{"id": "train_028791", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "bac123d8a8d6612a5a6baf86a2492d3ef9a1b9b3"}
{"id": "train_028792", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "54d6460ee50b98250b8d91f4d7c5abf4d4104b70"}
{"id": "train_028793", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "4d15804f8908e9a837dd31d13ff35c7a0a60a481"}
{"id": "train_028794", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "64263ad462285c36b1a3fd3e2638992cc78c06d9"}
{"id": "train_028795", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "d6b8d64963bf79601fb27047bd524a33fed79052"}
{"id": "train_028796", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "321341e576c2dc3df0981de8b4e92bc775545343"}
{"id": "train_028797", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "6ac49a89a411ba72bf9185bb6c055de9ef228354"}
{"id": "train_028798", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2932b8cc7201af118989604965a51cc78cee46e8"}
{"id": "train_028799", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "503930869a0646061c735e796151fcad1847e2a1"}
{"id": "train_028800", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "7cfe96af3fe0958b400b986730bb5f05294986bb"}
{"id": "train_028801", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "f19c0b3d3c121a4c482abab72f7b48167b7e552f"}
{"id": "train_028802", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "64bed467e608ac3700b8af699ac704edb93a6fa7"}
{"id": "train_028803", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "12f8b00516f2d2617bcb0d479b2be0780e1b4b62"}
{"id": "train_028804", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "348fe4e0e893334d269e3e4c9adf647b24814766"}
{"id": "train_028805", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "5e526437c3773e7d02cb829d299940f82948bd64"}
{"id": "train_028806", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9c959812936dd0b622aa47821a2fb86559453071"}
{"id": "train_028807", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "e526acf33f41bee10a1e87306ea81e5b6760edef"}
{"id": "train_028808", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ed3c0e0e5219424ae82d92756d317fe035e3fcc3"}
{"id": "train_028809", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a0dbe8b5f15b312314e873237f78c87edf095648"}
{"id": "train_028810", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "787b2b00e935d908bcca36a4fa3179bc9980f592"}
{"id": "train_028811", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "d77bee4bf988cb2694ddd445fb71b192aa2936e9"}
{"id": "train_028812", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ba615a90ca688bbd0cca2b78575607a012eddd92"}
{"id": "train_028813", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "da8b38d463fe6fd33a480b22a113143adde8d7ba"}
{"id": "train_028814", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "2d8a27234a8488b4d984da7e061bdb8fd07bf8e5"}
{"id": "train_028815", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "08ba4fb101460fd2c3ac39e3f921e9570833a0d7"}
{"id": "train_028816", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "3c068f4dd88cd7e394cec2228f4e3cb2d7c5caec"}
{"id": "train_028817", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "1cc7e239dce5cd41ec0efdca700c009bf861c058"}
{"id": "train_028818", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "016ab9ffafecbab94afa1f2f12049015e179b480"}
{"id": "train_028819", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "1d6a0fd4651dfddd7ee23ed50ccbed6ae4986a7e"}
{"id": "train_028820", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "12cdca7aba02cbac7d47614414c375386212521d"}
{"id": "train_028821", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "3e3114a3b65512cf401bc7b1623f83034470211a"}
{"id": "train_028822", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "f6ce13cb5d49099160e6d6572e5bca4e73106565"}
{"id": "train_028823", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "ddb941ea1064f08eca73d9ee64769fb8a6de2e4e"}
{"id": "train_028824", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "294f696e97425e6943cd85a74e80c494f9c09f1d"}
{"id": "train_028825", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "d60cf066c8c12e84170cc9c829cad5981f093688"}
{"id": "train_028826", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "1530147e9c0a256927df2b47c0267fec7b53ccac"}
{"id": "train_028827", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "ebcc4227996926853a7f878b4554a5962ff32d76"}
{"id": "train_028828", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ce174051ecad360a5e8ee94e12c7186d4460fbc7"}
{"id": "train_028829", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "2c37560364ff125b94d3706bc128d4e31a2f75f2"}
{"id": "train_028830", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "83c6dfec8698bc2f820925355ef1ceb3ae702140"}
{"id": "train_028831", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "32ba8d274825938c0083bc7f44e6cc806b694539"}
{"id": "train_028832", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a8b790ef38229872a5c916987f1e116d0159165c"}
{"id": "train_028833", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "9e05c87417f5b55872e9f423b1c8b6fa76401a0e"}
{"id": "train_028834", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "aa4cb70e07eab886d15d53f4c8c7384b914346f3"}
{"id": "train_028835", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "7954aa1198e17a0a7fb67e2c1761569798a0deaa"}
{"id": "train_028836", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "993c1f43e6c21520868f9bdc9571b5c169542080"}
{"id": "train_028837", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "6cddfc7835daac641742f6adcea41786b4b11952"}
{"id": "train_028838", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "d69dd21840c50244ac55ab9f1aac4dd3092328aa"}
{"id": "train_028839", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "ce42a10765ed478c019099418a48be8d5f9dc2e9"}
{"id": "train_028840", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "786b80286c309cb0c15b00d510a249ec7465c1c2"}
{"id": "train_028841", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "7c1241e1561f216b4eff79ffbbaca4490cc78e32"}
{"id": "train_028842", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "40e8e638ac4d3d46f33b6affcad558288192d327"}
{"id": "train_028843", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "f01191f0d9d1c9d6abff7be81007785d86a20717"}
{"id": "train_028844", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f7a9c579b586c56f1fbe640c0d8835a5164f31b5"}
{"id": "train_028845", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "76d08805dfc19d06439e6debd321fb842ac7101c"}
{"id": "train_028846", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6c8a0675c04d1c9b14bdc3892b9738e88acc1491"}
{"id": "train_028847", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "bbf7a46af949deefe3a9ae954e7e64562763cdbd"}
{"id": "train_028848", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c3a51b37443d8dbb3cf0219cf40de48b80339f74"}
{"id": "train_028849", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c2f022fa82f18b728c1065f52c566adedd01b74d"}
{"id": "train_028850", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "58e68b4f04894799acb395f27c4455d11178b9ba"}
{"id": "train_028851", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "afdac676b1d779dccc8603ba9450ce7f245308a9"}
{"id": "train_028852", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "7ee534ca9c35ff8e731a4f1276e68d728d8d82bf"}
{"id": "train_028853", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c0f9fb880b04f71a41bc1601690ed479cd25fa7e"}
{"id": "train_028854", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "57de9c8205788c8369a5e0c40634421385a81fb9"}
{"id": "train_028855", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ab25c2300ae7e1cf68df21c1c28e718ac00754c8"}
{"id": "train_028856", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "3e8f1e714f5275b8fdaab2892ecd303bc41e1944"}
{"id": "train_028857", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c16fe1ded1c65e55aa0ba0e26d09b6f143ebcc3d"}
{"id": "train_028858", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "5c34e0aa405d194054c29b15650a20698cb87e98"}
{"id": "train_028859", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "8023548e81efcf90314c006c3bfbf23cd7ebff63"}
{"id": "train_028860", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "e7c450a086be6dffc62acf131daa5430034c7e44"}
{"id": "train_028861", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "87790017ec8db8bd43a32a6c0cab80936a454bb4"}
{"id": "train_028862", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "163d766a4bcbf52b676036610edc5b5f446e5e6f"}
{"id": "train_028863", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "cec73c47ec5b464774003f4b875a7471946e19dd"}
{"id": "train_028864", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "f8296a93eb3aeda83a9504f0cf1619e37f9820a8"}
{"id": "train_028865", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "30dadf7ed58c5abd77b5dda27d9d02cdc1f43b4d"}
{"id": "train_028866", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "b76ae652fcee4aafa1ad1b7c95650c61ce81c6a9"}
{"id": "train_028867", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["agent", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2c1db5c82fd07554fdee071d98079c83fc844451"}
{"id": "train_028868", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "62a2a2ffcac93c9a45f399d2aafb86e2f545af2e"}
{"id": "train_028869", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "fecb23716521993fef4367e4b91893cabd86f5ea"}
{"id": "train_028870", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ed4cf7da4780ba1390b461af2a7a00ef72646d19"}
{"id": "train_028871", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "f40b3ef4e1ef698e19cba60afde1aa261dd411d1"}
{"id": "train_028872", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "fd3ac7f3e2435937b6cd64ef1a18fa86f43cd48c"}
{"id": "train_028873", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f949b341c1c07e6f1fcd325e47de1b724b67725b"}
{"id": "train_028874", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "2584524c73b69f93eff21e6f2a48ed29849a79de"}
{"id": "train_028875", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "0ae50de68556dcbdfde99e7af0ffdbc0f0da4d96"}
{"id": "train_028876", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "2c5b177d1ec0bf724bcc76b60f89881d288b8d11"}
{"id": "train_028877", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "afec30f18a9c74767f1860c531172b19f2786134"}
{"id": "train_028878", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2baa590ccd02d557a27d4708308bb21bca0dd277"}
{"id": "train_028879", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ecf0587a967e8da99527764793018dd6678849a5"}
{"id": "train_028880", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "27a20a62c0d525016bbd8623edd096aea850d4de"}
{"id": "train_028881", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "1a5d786a4aea51f905714e2101746c8549cf6f35"}
{"id": "train_028882", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d7f0880068cb18edac299cb0a311bc9b04eab582"}
{"id": "train_028883", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "2d4aa5481fcf2a10b98d9721c7766547e0e2f586"}
{"id": "train_028884", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "2e7616c1a224e78b788fa392c9611603eff3eae2"}
{"id": "train_028885", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "95021f07405593824c25208b3bfda150faea727b"}
{"id": "train_028886", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c9f468542b7fd9e667219f2c967bdca388747617"}
{"id": "train_028887", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "245cb2ef156b3af4ca6af524d3306114c73bede8"}
{"id": "train_028888", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["security", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "99efa2484c2f32da3976f52f97d64f5a9535b7b0"}
{"id": "train_028889", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "e7fa00aa3cea99a5fdc3d6c37c26fedcab5567c8"}
{"id": "train_028890", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "23af050a48196d3c839cbfbc1ecf0be709190d6d"}
{"id": "train_028891", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "c99247f3448ae88acd6e1f601f08a10cf81d309f"}
{"id": "train_028892", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "cd1db49fa51086f52e368e71278d7717d9eff71e"}
{"id": "train_028893", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "b6fd8f752b0cd652d448b94fb907377f5b50394f"}
{"id": "train_028894", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "5a6e29b2f831c2d198f72fb89404b7aa2492270f"}
{"id": "train_028895", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["governance", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "d6a8b7d8886de3be6cf1746ce2216e390773b54e"}
{"id": "train_028896", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "798a9c6555fe790ef0d93ec30dacf941054e68ac"}
{"id": "train_028897", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "7f9026e749403a702cedf73b6dc6cc4d01503cf6"}
{"id": "train_028898", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b31dff83c8611cc6123754e31e1ebcb4ecb8417d"}
{"id": "train_028899", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "8ce48b8ec65ae1c25a2f7246564eef11be8a2872"}
{"id": "train_028900", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "aad0d9309338ac3f7b1d3751a4a5e02af64321bf"}
{"id": "train_028901", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "2a068cd3d248429eb53fa56b72ddc203353a59fc"}
{"id": "train_028902", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "87ce239165d9a1a315f243ce89ac9509652a2d2a"}
{"id": "train_028903", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["research", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "fdb3905f93a5ca7b4b002ec0aebb334b02994757"}
{"id": "train_028904", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a626003ad6efee3efc2b53034c134b50390c8fae"}
{"id": "train_028905", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "16c4879ddd4deb99720f0f6ba5e1e1df4ae14d1a"}
{"id": "train_028906", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "0007bdcd575c43806e52d8cedb568a7ae32d6ef4"}
{"id": "train_028907", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "ddb3f3b44057b9011c747705afa24418f40671af"}
{"id": "train_028908", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "5190e4993a6e72e8198d916e2e134024748f9a0c"}
{"id": "train_028909", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "680eb2e6b1042209d16fde07a52c8cb46c45801f"}
{"id": "train_028910", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "ffa6f9ba3819d99b7a707b743d20967af7c0e773"}
{"id": "train_028911", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "01ae623a36892393d211f0c1696ec3be65edff9b"}
{"id": "train_028912", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "55eb79ab1a0b41d87a618dcb1ed56db7efc0943d"}
{"id": "train_028913", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "9a3fa15cad0888d8b3a180ba31bfb2a3dc22d435"}
{"id": "train_028914", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "5392bb25dc5cc63b10b9233cb500c3a0c157430f"}
{"id": "train_028915", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "88bfbaa8f00de2ce699d8572e95a1b22d1b6ef7a"}
{"id": "train_028916", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "caad43e68a2389eb2f3d96afc2def7ee9a7ea083"}
{"id": "train_028917", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2aab7e9508b72d4704d1dd2a193f6adc5bba1fb0"}
{"id": "train_028918", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "f904954bcc31b7148fc4236928b531111a565ad8"}
{"id": "train_028919", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["performance", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "71df555ebccabf54211d782180ecbd2b2f78724b"}
{"id": "train_028920", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "797d156a37231ee25f0459b11ee4748ab57fe255"}
{"id": "train_028921", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b07e58856d37da2b11ccfacf071941896a4b8fa1"}
{"id": "train_028922", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "18b9b20b5b02b8f1f8e1a1dbc8ae89982077657e"}
{"id": "train_028923", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ad1e472ebb627422c0863192c51dcd0cc8184851"}
{"id": "train_028924", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "70d78986822c7500b779b5d520adf99efda79cdb"}
{"id": "train_028925", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "ddb3f3b44057b9011c747705afa24418f40671af"}
{"id": "train_028926", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "fd5a9adef1acf599d4ab9f58a85250b2d3716132"}
{"id": "train_028927", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "169560cd9b0b0ce69f6193d4cc2b5a3098baed0e"}
{"id": "train_028928", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "4baade02d402298f31ce0f6ee7881344a2338639"}
{"id": "train_028929", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "22ed4cca7fc47b2fc00bb4e5f918bb69bb1bebb6"}
{"id": "train_028930", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "a7d1933a796cb1cc36f3228181ee9fa291d4d157"}
{"id": "train_028931", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "db72f324dcd9fc0f57bb645b8366946cf0845d38"}
{"id": "train_028932", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "9435b31d62a85d61db88c7ffbd273e24ccd68f41"}
{"id": "train_028933", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["data", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "fe41d604e3d5c2c4814d9b725852f831a26a2b43"}
{"id": "train_028934", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "93b341d3ea3406b577868f7f474299562e9dbfa6"}
{"id": "train_028935", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "995bab96ed6356da50ba713d22c6ea16ee1c6b4b"}
{"id": "train_028936", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "ffc57ae6121b98e9597999367d472a2be1da93ae"}
{"id": "train_028937", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "28e10dc59ced3f6500fc68438731174ef4751885"}
{"id": "train_028938", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "11a695c56b44af449e593e0c9f00f234a3fe69b6"}
{"id": "train_028939", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "2661f329f89fb8ba1c8c4b7fd447d48fa138ab6b"}
{"id": "train_028940", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "009c86852e0325875b9ca8cd17c849e540160c7f"}
{"id": "train_028941", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "db361ef9a69f7ffbc0b43319dd1ba297bfa7471c"}
{"id": "train_028942", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["data", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "8232ad87e1703aa98a0d0a715bf926a0c2704841"}
{"id": "train_028943", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "943e954340a048334c7c5ad96d37c201f8093ee0"}
{"id": "train_028944", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d1ed87dc1fe5687f378932a827b4aca26386c152"}
{"id": "train_028945", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "0559294bd48926b2abace1f81f9b8d9072ad362f"}
{"id": "train_028946", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "54bf111be7918eb841af2b7b047da3f3ef6f4e41"}
{"id": "train_028947", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "50ef229b0bbdad1f0f760eec1981b1866aec73cc"}
{"id": "train_028948", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "43f7e8ef0ea1cebfcbb1538e77dff5b14ae36109"}
{"id": "train_028949", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "2a23d3003ba37c30d60ee3936bce435573cf7e70"}
{"id": "train_028950", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "0256e5e9bf71b7040b1c1511c78757c908f94d4a"}
{"id": "train_028951", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "8bd99303ee0d854a661a9894ab9b05ebbec2989a"}
{"id": "train_028952", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "58e68b4f04894799acb395f27c4455d11178b9ba"}
{"id": "train_028953", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "3c29c3a0641169ee0d7d3b65e106a24f1f974aa6"}
{"id": "train_028954", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a11361e816d9bb09f1016c9ec06cf62e438e5716"}
{"id": "train_028955", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "9376a8a6389455b5c1274dcd5a7e4e60ade7c89d"}
{"id": "train_028956", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "028bbfd8a485a0d3b905582ad7724966a0cd91e3"}
{"id": "train_028957", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a21666d42098344f1b1363e385525dcade18842d"}
{"id": "train_028958", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "2571590d066a2fd51196db0225981a9befe64fee"}
{"id": "train_028959", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "127dcd783cb41ae169cce9b6ae73b2ada4cfea7c"}
{"id": "train_028960", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "95633fe12a6e9d42ff466659fb2b2ce1826cded8"}
{"id": "train_028961", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "2791b2070ea720804d05047c56df9e8e6a8645d3"}
{"id": "train_028962", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "4aa3be3891417b456a0927ceea1034d7870ed272"}
{"id": "train_028963", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "76dd58236d95e393a78d037f3fabf4898a8ba2df"}
{"id": "train_028964", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "843e9200bf5fa8839559e42c725a285a9034d818"}
{"id": "train_028965", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "a00cf69f9d50e651d407bc29759cc98183c71c39"}
{"id": "train_028966", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "5b9f48bba08fee478fa5266a201b12d86e329f5b"}
{"id": "train_028967", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "138b9672552ed58305d4f4fb8f6ed66bfa92c032"}
{"id": "train_028968", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "571ba31fd7d7f898c8f225055f4d164cdc935756"}
{"id": "train_028969", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "79a68ac9c237972cac9a15ef2093ac3ad5cb6c23"}
{"id": "train_028970", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "f1ae743d374f8df4f2efb517701d8adec67248ed"}
{"id": "train_028971", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "7b2795405c67ddcdf091c58a2937f09b647d1581"}
{"id": "train_028972", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "72f5c7ca6bef009c26dea177a36fda5eb2ad2dbb"}
{"id": "train_028973", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["evaluation", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "ec75846631d3d3e6713204bd036584c28f0a715a"}
{"id": "train_028974", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "28e37e4cc6673b1a0cef632289f06237ee7a0090"}
{"id": "train_028975", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "da2aa776c5e3ff08a98e07be80d34030c8073d4f"}
{"id": "train_028976", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e0b6e341739b7020fff0365c5aa19661e808aa56"}
{"id": "train_028977", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "3247d3499a743878191221f1b07eac265618f498"}
{"id": "train_028978", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "29efc11ddb1920461ebf441729623591cb574b87"}
{"id": "train_028979", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a2225cb3ead123252930f3014d48d6e1e3bc303c"}
{"id": "train_028980", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "307bfa2a14121eaa8ca96e252793ac4010db4b99"}
{"id": "train_028981", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "0ac85336095dbd141f9afddf3f16054784799e35"}
{"id": "train_028982", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "2bc3592679ad57b1ed3a63ad063961ff761696b3"}
{"id": "train_028983", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["data", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "ea131f9d047ec0bdb76628018a09a741d1e5b0b9"}
{"id": "train_028984", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "20b82f5e4b6807ee6e076875a423d38976845ece"}
{"id": "train_028985", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "b7d1f578ff59694c698e20273140f2bf6e9db796"}
{"id": "train_028986", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "971e0ca25eb57b7a310cb219165e6b8f7661155a"}
{"id": "train_028987", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "77e2e402c1d3dfaa5a824daa4cbd15ebadad75ce"}
{"id": "train_028988", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "7e0d8e0e0ea265403df7cf41a7ad85bb4fc07adf"}
{"id": "train_028989", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "92edf949c7f22cbc0fc3dd64d6bc7f5198b0c12b"}
{"id": "train_028990", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "656dde842c288074b6c72b46aae2dc8bd819c23a"}
{"id": "train_028991", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "a566fb8a39d1f1624818fd322ba419dfdb70689b"}
{"id": "train_028992", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "91bdca75b617653dc21ecb741c4971008c993ecd"}
{"id": "train_028993", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "6d021c2e286926e19e2af47d9d2750c46d7261ce"}
{"id": "train_028994", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "f96a35b3ae7238878ccf453d3b6d509e2c19497a"}
{"id": "train_028995", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "2fbf696cc33439cb49c219586fdd760bac37d9f7"}
{"id": "train_028996", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "a78a3a30f993f7160e9f7961e2382a74bd5ce9f4"}
{"id": "train_028997", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "eac3da1aad123f68ef96666a5c5055b450218f39"}
{"id": "train_028998", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "5708631620e2e0c4f7b72d04528de7f725aeb545"}
{"id": "train_028999", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f5b8495fd2023aeb7cfc382d4e9ba8048ee3b153"}
{"id": "train_029000", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "eb6e7685725710adc5be99e21abb70d33d041272"}
{"id": "train_029001", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "cbcec364ed384ced79028c2239e7e2a957d020e3"}
{"id": "train_029002", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "5f43a2d864b3109a5bd7fee88596227d201478e9"}
{"id": "train_029003", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "905c267a06042e66dd1f41a2d255186439c56ae4"}
{"id": "train_029004", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f34d755dc328d05fe63a889b86f38dadc8fd56b6"}
{"id": "train_029005", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ef7a008f6191a9dd7f951bdf38b6c32bf9461620"}
{"id": "train_029006", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "eee1145211863ef251fdef9cdbf898bd3906b2e6"}
{"id": "train_029007", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "e3563f395dfe2ffa13ff30d6ab3d155e20ea2599"}
{"id": "train_029008", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "5fb254dc96de9017db59092a2fbff43001baca95"}
{"id": "train_029009", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "651204f672767f27902f5d866b43942dd474477f"}
{"id": "train_029010", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "b0b1b2490061d987e1db17b3b3537c8bcc40e278"}
{"id": "train_029011", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "ce9f4ed6fb088b9757e5d9a6a5f3889bf7ae0ac9"}
{"id": "train_029012", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "630ec416b219c9b336567ddd4646bf88936a65ea"}
{"id": "train_029013", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "63642c32ade39a253dd0212badbd9017bccaa0a3"}
{"id": "train_029014", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "6b696a97b899f8012d1becdc07ce8a139640a471"}
{"id": "train_029015", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "4d0f36b39a533bf320a3775143da12f05ddb8794"}
{"id": "train_029016", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["data", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "bedd6e02b56707cfc25500b04dd631c76f4d80f6"}
{"id": "train_029017", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "db4a1c71d8007a82495bf5be94a1725ae6686463"}
{"id": "train_029018", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "cdc53e638f2fd88df3af4f653bcbaf33e7059cbd"}
{"id": "train_029019", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "805b9711f10fca2cd9b8c5e38c41227bcca3cdbd"}
{"id": "train_029020", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["coding", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "15cc01d1241dd94764f8de9cf1aace9404567500"}
{"id": "train_029021", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "53f20d336374bad1d86e6db899a9c0eaf7ccce5b"}
{"id": "train_029022", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "bc56891e8cb15b074e9413e70c8c6a99c5cac848"}
{"id": "train_029023", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "4386d72f48fe50b72d10e5af82ec663f54bb21cc"}
{"id": "train_029024", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d98ee39b8c31e7048359a5f260429f5cce7bd749"}
{"id": "train_029025", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "1179ffb82e2fca3cce6c47eebd448f33b2a6b6bc"}
{"id": "train_029026", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "cf5a2cc375594a1bf71436e793095c5c0524f9b8"}
{"id": "train_029027", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "7f90fccd86239d4b88d660edbebcd3c20d900b83"}
{"id": "train_029028", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "915afac4832676a45d9ba2394b15a18254252183"}
{"id": "train_029029", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "61efd778fff7ada7ec406a565efbe1ac6d5e4a9a"}
{"id": "train_029030", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "49ac8d318fb4e5f672c5f198e43630da3c2ecf5b"}
{"id": "train_029031", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "66db99b328a5175bb49239b2e7fa4f883eebb085"}
{"id": "train_029032", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "5de118773cc4f7ce4e090d210f760f16801e8d8d"}
{"id": "train_029033", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "b86c8f172a4cfe604717bc2b7a30993c22117620"}
{"id": "train_029034", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "14c39db7a3cd0b7990ff10f4e5c7c3ebf5dcb4f2"}
{"id": "train_029035", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "4aeea3fbe48f37bee20aeeb801fa6a6690f10357"}
{"id": "train_029036", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "b0c8e9583f07c6ac98b765981b6c75b01d60e25b"}
{"id": "train_029037", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "928843c3eba177ef30cd8efa32842d65d746a34a"}
{"id": "train_029038", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "29b25f32e425cec18cb38ad032b85af06c9476b9"}
{"id": "train_029039", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "99fdd164420e5eb2c4cef9b44ada8181ccbc6a3f"}
{"id": "train_029040", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2c87be6b97fa9a8d420e49c919e898562fb29c5d"}
{"id": "train_029041", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "a5c67d18494c142d0233c767690532a2e9c3d162"}
{"id": "train_029042", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "950fb29aaab14335e258d50637e4fc96fabeb85f"}
{"id": "train_029043", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "923febdca577b74f2e0a0b2a13e98da1b2c85787"}
{"id": "train_029044", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "a2f69b71141da58b0744f85f0873bf419c38adfe"}
{"id": "train_029045", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "93be4d3436d479c4cfa43ff40a0611051ecc2b36"}
{"id": "train_029046", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["security", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "037f2633ab413c80c98768df8f5287e90da8bc47"}
{"id": "train_029047", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "1caa276a90ce13e12d67b3aaf04e4d2873a0d713"}
{"id": "train_029048", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "861e89f8e9734a66c2fec1088f7e98f017ec25b5"}
{"id": "train_029049", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "841b1e1de15ab0d96222a22ac2cd73a47725008e"}
{"id": "train_029050", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "5079eaecf65cf14ea534b08ed07eda5738ea30c0"}
{"id": "train_029051", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "513c6f062141cdea5b96b59821bc029f63862886"}
{"id": "train_029052", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "b7bce14fb4c89b7222598f0b7771a167ce595606"}
{"id": "train_029053", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "2821f9423350fc2b1ab92e8c07acf17d9d8ded5b"}
{"id": "train_029054", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "3d79b19abcce8441c1ce427b83e7fece379a0cd9"}
{"id": "train_029055", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2a9f244eba2747497d82d74bc72998102b4d7e73"}
{"id": "train_029056", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["governance", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "f4440cb83e0d86ad851ceabaf07ff6e02bf35054"}
{"id": "train_029057", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "5aaf5d6f63e739466de648b21b6b9b078fdabcf3"}
{"id": "train_029058", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "516da9d4dfa3767f9e650921a97a31d465e3a94a"}
{"id": "train_029059", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "5738cd33ebb465fb321e2483e922a338656f05b3"}
{"id": "train_029060", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "136750d260c38d88348afff5f9be67a0e2b4abcc"}
{"id": "train_029061", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "c20dea5f13661d0e8690f5a668867b9a09813ac8"}
{"id": "train_029062", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "0500e3ca0bca16ebb039b0240eaba3a424720228"}
{"id": "train_029063", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "d64efabe726093192460ae46f0411fec82f73cb1"}
{"id": "train_029064", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c4ed4c713b395abf8478f98799e9c1ae123e36af"}
{"id": "train_029065", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "9dd0402d79d66b2aa02c27d62c2aa591117a9390"}
{"id": "train_029066", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "6a22999336b64388358e41faa20b0e78d2646d7a"}
{"id": "train_029067", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "1ac95a99261302015ece116b933e4ea70d53303e"}
{"id": "train_029068", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "d0141eed2eaf673e80969a2c6e233419af851107"}
{"id": "train_029069", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "28daf676d9397205bd414883b1cffafdc2f38921"}
{"id": "train_029070", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "1b7d66d88e4966db25cff24dee632c7156e2e7f8"}
{"id": "train_029071", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "0830035c8ddaa7a3ee5baafaebbd49bcfdee2166"}
{"id": "train_029072", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "e3883fac3bb79bf2748baa5e53c720435bc631f4"}
{"id": "train_029073", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "58ce64d2866a83e8ed72ec6a7ddae54a896848f1"}
{"id": "train_029074", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ec2274d7e3c2d35901e990c64152ecf6369b8208"}
{"id": "train_029075", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "4ee1e843e2b182046a58b6c842874b2e0fdec8cd"}
{"id": "train_029076", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "048bdb5093ec68a0a72eda0ad63a92b3aaf9077e"}
{"id": "train_029077", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "421cc5219e96f86f51e349fade9b7b6d4d5756bf"}
{"id": "train_029078", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "12fe6ac8b194c61f381f6fbc4153edd3584dafd0"}
{"id": "train_029079", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "9f336e8f2c21db553bbada54030e625a2f36cb81"}
{"id": "train_029080", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c69d41ba412ec421889e3d2f6b68a4ab84441c00"}
{"id": "train_029081", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "4fa060aa95068d2dc1f35e74ed965070ed8a8709"}
{"id": "train_029082", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "8e5820bd57d0c79c88b21e454961a9afb1d8de24"}
{"id": "train_029083", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "5aa9ab865ba54b52c68de2b63975146e24700a6d"}
{"id": "train_029084", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b159297e4a0285be01bd9f7ca20aca620399662a"}
{"id": "train_029085", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "2074d1b96d9d6fc7dc27902720b68ddf1fceae0b"}
{"id": "train_029086", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "54d2f174e43748eae94bd2244cacaac0172e58f7"}
{"id": "train_029087", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ecfcd7866d145045edbd89881382536f42db297a"}
{"id": "train_029088", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "372b5008b0937b6a7becd6629c111cf877985e2f"}
{"id": "train_029089", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "acd25e5082e0e00e09d44e99c7f889c9b1419519"}
{"id": "train_029090", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "add5a4949434063dc3485bc01975e3508005f98a"}
{"id": "train_029091", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["data", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "6be3475790b646c71d3032d613266c56289d9aca"}
{"id": "train_029092", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "441286549af551f1a1ed1099af5178b059b2c047"}
{"id": "train_029093", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "77aaf47b35fff9adf3321ed8625bb8b7c42428ab"}
{"id": "train_029094", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "3eaa74e10f3c20ecaf3578968757e953dcabd520"}
{"id": "train_029095", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f3d28b77666c89dec6086c8992b241be715fc285"}
{"id": "train_029096", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "48d0f8ba61d01a4b82982b7e34de3f9b139f8743"}
{"id": "train_029097", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "8a8234f5c9da6379cc77920e72fcabf8ff68b25d"}
{"id": "train_029098", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "765529af724dad4a601efeeefc05ae364f9caa96"}
{"id": "train_029099", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "5e627d63c08b0da79ed677d7d7097f8b51517cc3"}
{"id": "train_029100", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "cc6901d324f723e4fe22d0b19c451aad0d8dda2b"}
{"id": "train_029101", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "41a064279a7bdf82234a82a9c8962bc3431f5032"}
{"id": "train_029102", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "ad8cc31899e0fc37de400aef675ed4420fcf681f"}
{"id": "train_029103", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["data", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "24ec2627a34db4077b53c6533927958f3d0b47c9"}
{"id": "train_029104", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "40b4e2ffdba30a8fdb0b89b207271236aacf1cd6"}
{"id": "train_029105", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "4f374f9149a97b66e8a405019c950efeaa6d940d"}
{"id": "train_029106", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "8652f669289290cc4fd2534c1347ad96a8afc88a"}
{"id": "train_029107", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2e6b23202b348213482f9de4866033fd0513d103"}
{"id": "train_029108", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "977593847fedbc272a6ae3e6b8ca3116199047af"}
{"id": "train_029109", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "0f7e72a07ff727af8b145aa3d05f9adea569e5d6"}
{"id": "train_029110", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "b2d1f307a0709e33d583c2ccc3d6ed2830d74786"}
{"id": "train_029111", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "7bb83327af74288dab89b84a1896ef6e152da571"}
{"id": "train_029112", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "7626ae550cce9691b0934178a2bea179171df907"}
{"id": "train_029113", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "144a7b9be20bdad3301344db78afd82b5a028234"}
{"id": "train_029114", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "63532500d677f87a7c9c134ca941094655d7f621"}
{"id": "train_029115", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "62a2a2ffcac93c9a45f399d2aafb86e2f545af2e"}
{"id": "train_029116", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f0babb971f52704f9392cb4455becf808eddda81"}
{"id": "train_029117", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "9c93c8e20bdbeb64fdf758e0981400268c7ca37d"}
{"id": "train_029118", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6d2969710539fd62637d9336a1608427eb96503e"}
{"id": "train_029119", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "ad8cc31899e0fc37de400aef675ed4420fcf681f"}
{"id": "train_029120", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "3e413bc874760cc9358a29aa347f89847e7a7a0e"}
{"id": "train_029121", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "0179c5a36332e560120f8ac9f21dfbcd30a9bf43"}
{"id": "train_029122", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "d2ab8ff50a6f0bebf51a53e681933cbc9a3235c2"}
{"id": "train_029123", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "b54d92ee11d656efd5c84727e7c88550c2295bfa"}
{"id": "train_029124", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "bd1bdc736389d468077cda22988808b0894dabed"}
{"id": "train_029125", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "29db0db86b7e26d6aea894f70cc5346a7534f5b9"}
{"id": "train_029126", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "d4762daff3867dccba490857d2babbe54cacbaa2"}
{"id": "train_029127", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "f20e96f7de1349243db4f005e7e2c9711292b023"}
{"id": "train_029128", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "60969c7cc54d0c505ed1ecce8e5c0b33d94f98b0"}
{"id": "train_029129", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c55b98db57ce80f0c222653e11ae389ff81d8c14"}
{"id": "train_029130", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "58787be06985ee26ed65f13e1f9303e03f5a5c34"}
{"id": "train_029131", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "17e108ba739dbda2526688b6afca840f3c69d6d2"}
{"id": "train_029132", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "57f0697dd57be146127536516c915466ff52a611"}
{"id": "train_029133", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a8a1db1dabfe472fc9e59a223768d258e7487b09"}
{"id": "train_029134", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "53cd86576024c677925bcf50b72a8e14b01325f1"}
{"id": "train_029135", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "3d620b6bb95a54a08071bc97a398f7ce40a084ec"}
{"id": "train_029136", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "aaca6ef8735993247caf475c8df5fcb1fe434d9d"}
{"id": "train_029137", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "54bbd19f1c95480ec25b66fec41819b4d8da32ae"}
{"id": "train_029138", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "8e9d2272e20420a32b7ba943e981d61d401a3d6d"}
{"id": "train_029139", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "1a82a1e5286b174a8b3afb5df7f9e865e542e4c8"}
{"id": "train_029140", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "72971499ffcdf30cde9eea808f7c81083f27c1e4"}
{"id": "train_029141", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "397fb668c8e1df0ab36a92fab80544325d1f7b84"}
{"id": "train_029142", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["coding", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "588f4546bea4cadd2de462f3763e25a7904c7294"}
{"id": "train_029143", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "af5ed1cc5b98123f89087db04c7c92861a15f843"}
{"id": "train_029144", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b77eb28a3be1ac1de02b584883cddfff45dbced1"}
{"id": "train_029145", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "5fd425c2b886603abe4abbdd8e113ce06525e6b1"}
{"id": "train_029146", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "0151eeda47407636a52a81c6b0701fddad76c5a2"}
{"id": "train_029147", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "0dddb3d260e763bf7de4d14594e5a9143a1a01d2"}
{"id": "train_029148", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2c1db5c82fd07554fdee071d98079c83fc844451"}
{"id": "train_029149", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "7587468adeaf9dec7e2c3c0c721fde1b18a4d017"}
{"id": "train_029150", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "40ef3009b5770a6708123bc2382dd93a3d168553"}
{"id": "train_029151", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "32b5704c93a7bb9297d13d6804c3d5729fc52e0d"}
{"id": "train_029152", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "070148d76120909eb938a017520ea39eacf9ae48"}
{"id": "train_029153", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "8c2734c1f93bd4be8648f54d0bda9ea4f90d0068"}
{"id": "train_029154", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a2332ddbdd02ed104a1786fc82f414f7dc393f39"}
{"id": "train_029155", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "aead422c597dffc627a277ecc2da77f6440b9e90"}
{"id": "train_029156", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "3074050de7f1400c4162c0d75dd7dbf4e59b2e37"}
{"id": "train_029157", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "d9a9e018efdce2b7bbd7d8f0b6c731cad80e9878"}
{"id": "train_029158", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["research", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "8466bcb1862021d0359bd6b01b2bb12ff0bbc886"}
{"id": "train_029159", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2ea451d71b69e77d1576d0b0a3404572614568cb"}
{"id": "train_029160", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["data", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "5bf3c0614442b58737dbe37aa39b3fafd9187e15"}
{"id": "train_029161", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["security", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "5006fe2d4255162b7db3fa7befe3f3f4732a21b4"}
{"id": "train_029162", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "92887e57729a397dc5f7f0acc594a20a9f917223"}
{"id": "train_029163", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "e45c062bef59d980394f064b5e06e891ad5e1f61"}
{"id": "train_029164", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f48c6a1eb630b9ea38fd3c96dda6cb3734eed898"}
{"id": "train_029165", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "ab4172cfa0f3b7e4030d0822d8aba74e46e0bd76"}
{"id": "train_029166", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "c3696a9038e93a7bffe23cc3a92a13074f387445"}
{"id": "train_029167", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "4985f2777deea97c6fc59cad6e5371186ea29104"}
{"id": "train_029168", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["performance", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "32c15d839362b4c15a93f7fc8658e1fbf6541f7e"}
{"id": "train_029169", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "55beb3fe79a0bfabbd24d3a0ff9c1e5bba6b2da0"}
{"id": "train_029170", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "023265147a04b0d237dbf6ec13a17a70a073e279"}
{"id": "train_029171", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "dac48c20072c4dd511dbd4aa086c779c6cda3450"}
{"id": "train_029172", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2f7ec88138d2b709b9d0c9ace328cb8a05cd04ef"}
{"id": "train_029173", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "cff6c69d8da0f579559f01cc675038b958fb58f5"}
{"id": "train_029174", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "be8d75c2bd0712735e7381f077a0957c71c11780"}
{"id": "train_029175", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "716e4bf9b3369aa512beaf99984b20c81d391c67"}
{"id": "train_029176", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "6ddc33e78b4bf43d07f4c1fb5c9a55efe9089245"}
{"id": "train_029177", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "cf53a6e5195b395b5909cea90e203a0ed153e520"}
{"id": "train_029178", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "40fa8657846abce04bd5b2bc9a09442c328b2792"}
{"id": "train_029179", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "4d739aa562ce44b97ff84fa35893c990f79397dd"}
{"id": "train_029180", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "35d8aac1cdb2981c6728a99cec0ad9f5556848df"}
{"id": "train_029181", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "1a48e5b42c0d3ef72e2a4664497497aadc8d821c"}
{"id": "train_029182", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "1794833462fc090696c87076a3f8267fa447178f"}
{"id": "train_029183", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "0e855dbcb4546d3176e59359547f636a89601772"}
{"id": "train_029184", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "4013b64fb2ce072aab5587f7edccd1b451ac3645"}
{"id": "train_029185", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "523bf72a8fbe0c5e0d11a3d31a4915c82e93cf21"}
{"id": "train_029186", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["data", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "d3b475afc89710a62ed6beef3a796994904015a4"}
{"id": "train_029187", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "3d50d0388a748d1555bacc72b22cd1dc21988858"}
{"id": "train_029188", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "1c17b349c863db8d5518d0328310ecb9fe9654d6"}
{"id": "train_029189", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "5935f67da3175024a783998a6eabcc518941b701"}
{"id": "train_029190", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "bb2bbd290de73dc1114e5cb119e3856e6a515c69"}
{"id": "train_029191", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "224e9feccd9d53f9ee52389dab80241db8e322e2"}
{"id": "train_029192", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "6bf28e4314b461a5144bbb1fd4dd9b35584b2a0c"}
{"id": "train_029193", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "c5187488bfd948540b20d3d615cfda058b73fe27"}
{"id": "train_029194", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "b6ad0d505a279579f0f66274c0fcabe2be3ac285"}
{"id": "train_029195", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "9f2b7edd9f3088877f964e48433d0f54ff4eedf9"}
{"id": "train_029196", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "98ef6882189f5fb92ecc4eb67a31bdb00ff327ef"}
{"id": "train_029197", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "a0deeb729c4589546c2b08c989917d5fcd979391"}
{"id": "train_029198", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "86b35cc38ea4e53ac41dd90db6eec27a408f0ef8"}
{"id": "train_029199", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "3a516dc5064803966a2d034f3c87f62578e98596"}
{"id": "train_029200", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c5ad52f96d974652e7f33a63e3bb6b167464d3b4"}
{"id": "train_029201", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "5fedb19357ad7013eff136d13a6c612eee8dda9a"}
{"id": "train_029202", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ceb518842f8a3a44e475c535ba0b97530d6894ad"}
{"id": "train_029203", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "a697437882dd85e853c3d3b983affd1e23667b4c"}
{"id": "train_029204", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["coding", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "8cc569becc99125ba12b4c830b4b68a74884fc46"}
{"id": "train_029205", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "7c4fec3d89ee584fbf021d6cc9ece1c92bdfda65"}
{"id": "train_029206", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["research", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "594309bf3f38a9e2d840d32da57b5a3163b4a808"}
{"id": "train_029207", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "4c9983803552b6880411af09501d5a22bd11e351"}
{"id": "train_029208", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["performance", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a1873874d069fc5b7e1a5ef933ccab88b35438f6"}
{"id": "train_029209", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "1f5598cfc922dc95e5107f3a460236a0c638844b"}
{"id": "train_029210", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["data", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ed420b904f1fcb984ae700102f92afcf467ebe3d"}
{"id": "train_029211", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "e933abd0e77f31c086e18011a9f04389bd41bff6"}
{"id": "train_029212", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "d3011cceb730f1aaec52ccda4ef19d14423b5a07"}
{"id": "train_029213", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "133b8c60eff95201c977079c5dc6e4463324a8f3"}
{"id": "train_029214", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "8c857284be9298057184c0ed22aaa37ac8ded217"}
{"id": "train_029215", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "981be08bc94713ecff6d7b6978ef95e50aa9069c"}
{"id": "train_029216", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "6d3872d7afaf8fbcf82b98d8ba7a77ee63024e4c"}
{"id": "train_029217", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "7c34ffc1307987364d467d3ceb7af9bc9000e6bc"}
{"id": "train_029218", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "e7440794621bd5acb5172631700b2129f89546c6"}
{"id": "train_029219", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["coding", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "7c1161c7f59118faa01a5e5957fbe2bc88939640"}
{"id": "train_029220", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "3eee0f5ce6492e935f9fe65a97ec54047d27445d"}
{"id": "train_029221", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "74bfb5c1b69ecac24cb5c916f03bdfca01dcfc4a"}
{"id": "train_029222", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "e065cbfad85267a6886687990b4288c6e91171fc"}
{"id": "train_029223", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["security", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "604e392eea495aa3feeb3aed2436fb588109380f"}
{"id": "train_029224", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "1fb9b3e6c33aff7e2811bd074fd7e1ce5f415625"}
{"id": "train_029225", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "797a54c93bb0c4a3b7e666fcb31beebe13a9673b"}
{"id": "train_029226", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "adfd68f2cc2db804b182ba1ddd9c0e033d2d53bc"}
{"id": "train_029227", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "333396aae072ba7439b1709ef2bb0d39267ae927"}
{"id": "train_029228", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "346f2887269d782207b15ebd39fd4f6e35cd35f2"}
{"id": "train_029229", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "b91def1f6fcfdfeba5b2154167319e0e832259a9"}
{"id": "train_029230", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "f518cfff66c62b87ca1370aaa1944b08009fa007"}
{"id": "train_029231", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "cb2b17a49c794e17c16b4c06e5d71706ef61bb14"}
{"id": "train_029232", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "4daec53665aada658780393e63c1f85e1a0f82e9"}
{"id": "train_029233", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f7274a9e2a0d6f52a9bf2fece4057a5606625473"}
{"id": "train_029234", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "02a29901bf2b3d1e5af04891bab6685c4607c920"}
{"id": "train_029235", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "74b7766143b188da1b2053f36f37f514681c3daf"}
{"id": "train_029236", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "3d42819932f7cf9723c2ebfe9f6c7b0d4453a209"}
{"id": "train_029237", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "cec172dfa3dc12f0136923e5bac01b15030830af"}
{"id": "train_029238", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "982312b2f1458721f9b911671d4c8d57638a807a"}
{"id": "train_029239", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "6f2fad9ceaf8ee67afe4611f0e9cc6385245d83d"}
{"id": "train_029240", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8465b916c5d38feda5a6d0d66cd393e416e18afd"}
{"id": "train_029241", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["coding", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "b5f99917db5e3e09fe2607cbde6a45b3f2956773"}
{"id": "train_029242", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "6f942639479c07786c5063cca45759dce8881505"}
{"id": "train_029243", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b4e90431574091b59148b6e1221aa3d50ceb6afd"}
{"id": "train_029244", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e355ff449fe017b3a0e28112d13bc0f7e46432c2"}
{"id": "train_029245", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "88bb11f808244efa9d96e6d8dc49a1d7257d12ec"}
{"id": "train_029246", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f319a63abff6c96fc981a992634fce510c71d135"}
{"id": "train_029247", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "364005fa62dc2035cc8d62a783aa96280a1ef3d1"}
{"id": "train_029248", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "6dbe1701fcf0c6d628b50a91c84629d9602f7cd0"}
{"id": "train_029249", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f85dd995f047011c01ad1f5b24153141389626fa"}
{"id": "train_029250", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["data", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "004bb1e27cafed9eb0abba609beff2f2fe64d68f"}
{"id": "train_029251", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "fd5a9adef1acf599d4ab9f58a85250b2d3716132"}
{"id": "train_029252", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "cf435b83ce4e1f32c8359471087b7001c64fc08c"}
{"id": "train_029253", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "15cae4c1cd5ca4015dabe7cb72189239b10c0197"}
{"id": "train_029254", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c383dd2b1da123bbcf96f8c3fab6db951293e62c"}
{"id": "train_029255", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["performance", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "404ea1aadb6c5791f4d69cdb13e66b2a87007106"}
{"id": "train_029256", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "8b26ff2da70de2af107a91f66cfa25e84228d049"}
{"id": "train_029257", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "34fcc04a002e2ccd1e248b116ce06a2da7c34338"}
{"id": "train_029258", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "df2a9d4a415d3e022001b9ec1241b11e55d0f015"}
{"id": "train_029259", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "226b87d2c56d02dbef25e9349f49a880df152f91"}
{"id": "train_029260", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "d3cfd5f5af80ec9c3e886e6535469d160322972f"}
{"id": "train_029261", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "952a674bf79773494fb6d5c43208aa305aaf8554"}
{"id": "train_029262", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "0b3a9067201654a46a6eae189fa0be7622295110"}
{"id": "train_029263", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "800bbab91ddedd858c6b5f0fc5cc6652503f5d33"}
{"id": "train_029264", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "8bc41e84b3de6b137c1e41454bfd43cf91b1e1af"}
{"id": "train_029265", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c4395d614a2f153a27b2f8d6a5b961faa35fde7c"}
{"id": "train_029266", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "4b2bdd038ad40d7ce921b49dbb9e7385a757cfe1"}
{"id": "train_029267", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "725f161412cf4a61d030fdbcab438f916c84d227"}
{"id": "train_029268", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "d97250aee671c68ee2e265d97e7abe4cc5b5cb85"}
{"id": "train_029269", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["data", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "61394271a259bc80384d0355e0c97c0f8ffe7230"}
{"id": "train_029270", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "665a5133214faea1f2d4c5fb98b4ed3a31dea175"}
{"id": "train_029271", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6378971f88e520cf65fa6dca3ee12f9bca8c533a"}
{"id": "train_029272", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "6f0c3bc1ca64289979cd2de53254946033815e37"}
{"id": "train_029273", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "da39230c1c0fc186346a9dd5738ee6fd20b986f8"}
{"id": "train_029274", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "0f3e6d4fb57c6edbeaf1131cd16f63cb5b73c7ef"}
{"id": "train_029275", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "c6576eec34559a52a3c25cb339616eaabf98cf3a"}
{"id": "train_029276", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "bb92290ca82a6758551c470365f74ff17771d818"}
{"id": "train_029277", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["coding", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "284673fd58ed41287202691d1a68113ddbe28131"}
{"id": "train_029278", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "9435b31d62a85d61db88c7ffbd273e24ccd68f41"}
{"id": "train_029279", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "5accd4a5604d96c88824f113e98818a55e62e8bd"}
{"id": "train_029280", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "a1b64b794a401b6635f67f91e4461109abbd1475"}
{"id": "train_029281", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "d6e801aefaf9fcc9be38eddd05fad2d54a0266ab"}
{"id": "train_029282", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "5bd0fd09421f766a8472b3488dfab6fca822abb3"}
{"id": "train_029283", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "cb57b467a72dfeb44c482c8f714d9d226f3d38e5"}
{"id": "train_029284", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "96b0319295a060f4b34adfc8e45f14dec1e19a7b"}
{"id": "train_029285", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "d1d9e361a8f7700f5d27dae7d3f49eaaa4d10145"}
{"id": "train_029286", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "afdfda85759e1f9b9c0de34be96407e33c11b688"}
{"id": "train_029287", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["coding", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "d2538b0b141d28d70bce2226589742393d1f5bbd"}
{"id": "train_029288", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f83bfaae89c51211b6cd7889edf734bc86e92715"}
{"id": "train_029289", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "540b290d5d9fc8ce317d17905f2e4db11d90eb38"}
{"id": "train_029290", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2b5fce926d1d2978eeba691174fd42d3a40d4ad0"}
{"id": "train_029291", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "b5bc7a0b7c77a0cbaf5e0e64dd300e6b98735dc2"}
{"id": "train_029292", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "cb3c6b6d523466fec8c6d11a83d30b321f0820b1"}
{"id": "train_029293", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "7e722a5492e54bab000f0951ec8a527fd8fe14a5"}
{"id": "train_029294", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "6f6a418467f8013c11939c8cb1dff37d3b2d5cf2"}
{"id": "train_029295", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "1d4061b0afb028f3b9b12f45113691dbfca92765"}
{"id": "train_029296", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "0792983c1c4bafcf36bd9f2e0c4f3fd21c27c677"}
{"id": "train_029297", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "8ebe6f5cc1cd894b20b034d2fa2c53beef4b6613"}
{"id": "train_029298", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "bdfdd2bcb8b3ce897c0a5a3aec2f42ae0480aa30"}
{"id": "train_029299", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "034664b2bba7d64ea5881379794af9c27abb39a7"}
{"id": "train_029300", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "930e98bd5c3bb19ed9d6123281f067acee570fe9"}
{"id": "train_029301", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["coding", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "0775d6ad3403231ff4552fb38e83127a82609b88"}
{"id": "train_029302", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "4b539f53a395b9a47b79f8fcad4de2804a4dca06"}
{"id": "train_029303", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c3c6a4884c687d12d40690c0a9ae6bad8fc0d514"}
{"id": "train_029304", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2b5fce926d1d2978eeba691174fd42d3a40d4ad0"}
{"id": "train_029305", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ceb518842f8a3a44e475c535ba0b97530d6894ad"}
{"id": "train_029306", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "b201e6cc81065003e77e0e2f825f35e01f05a2d6"}
{"id": "train_029307", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "c67f15d35555d0daa6c76f02e40f0eea3e5158d2"}
{"id": "train_029308", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f36b8eb208fbf9b740cc1e1c812d237dafe2a778"}
{"id": "train_029309", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "20f4f08036fb94fe1f71e70709db468dc380bb85"}
{"id": "train_029310", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "dcf22b3b82ceb1d86a8ebe8f32b45d5f0e31bf9b"}
{"id": "train_029311", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "a179acb981927e01bbef94c0d0800b3d06f72780"}
{"id": "train_029312", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "44d6de83f38a24f227198303cade0839da82e447"}
{"id": "train_029313", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "05676b04bd270892109662a4555b65448b444eb9"}
{"id": "train_029314", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "68b154edde897c9b4d5efc7803846ae06f31e618"}
{"id": "train_029315", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "8528fffeb08c21398f901f6fe7f6fe4bca5d531a"}
{"id": "train_029316", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "dce93b042cfa2518401973623e98b1080781fc8b"}
{"id": "train_029317", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["data", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "36adf1f8b1899bfc4e49d6d9f26aed8f58a06e16"}
{"id": "train_029318", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "767bf58ebd0b1c3ce108c0892d5ed79229334c17"}
{"id": "train_029319", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "8b73e6de9fcf103aa30ed731da75dbc0256e3296"}
{"id": "train_029320", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "946f330286f3cb5d71b38e48c6960399f2995899"}
{"id": "train_029321", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "97efb7fcda1a55c0000db9c8e1dae78dbd02d449"}
{"id": "train_029322", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "a4dba4051426a23c7d1703ff924b8885bb454976"}
{"id": "train_029323", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "865563eff409c09392fda94e5cbe62c31334a310"}
{"id": "train_029324", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "f2ea2c7c68082df74a6b35a4e742466d383807af"}
{"id": "train_029325", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "fc25ac099ca089f646109e2a094a067dd3b89e84"}
{"id": "train_029326", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "e67773b682d8e309c4b00edf7e3ca7be59db59d1"}
{"id": "train_029327", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ca498854a5f5c45d971b28c1a2c963a86c2660cc"}
{"id": "train_029328", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f6db4b081fc99270b556920b9514c880ea52140e"}
{"id": "train_029329", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "2341a1dde2e0bf239c56f2965bf904a34ae2a143"}
{"id": "train_029330", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "945e8dc7e7b6e4b2bc772de6c2b3d2c21939f136"}
{"id": "train_029331", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "6d144fe0e68ac7183a17862d95ce12814d6e28b1"}
{"id": "train_029332", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "bbf1e9d1623e274ad1405a22faa11678b515371f"}
{"id": "train_029333", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "d17be0472afe0c6f46655fda8bcb978f88c4e9fd"}
{"id": "train_029334", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "0bc8f65b8f6d3a42e7c12a21ec619503c2666330"}
{"id": "train_029335", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "80bf3a9a0d4e734c18ac31f57d1a52f552637077"}
{"id": "train_029336", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "46bb19a840d62fbc35f3c39b692778b3d39aa73a"}
{"id": "train_029337", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "3779a3e3fbc27b2775fad3b2e8168aef5b6b0e4a"}
{"id": "train_029338", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "7461b9ba290695ed3248e03151b7bff56af8cdbc"}
{"id": "train_029339", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "90ccbcff99924004993218fbabb5c57ac9cbbeff"}
{"id": "train_029340", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "661f74f3642eba5f273e079ec0bd467372adff76"}
{"id": "train_029341", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "006b9bd737c6ef6fdcfefe6b7c0e86e4ff90491d"}
{"id": "train_029342", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["coding", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "169a3890b50e582e4f56524a7e5cc535aae57a0d"}
{"id": "train_029343", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "88aa090ab42312ab63d918ce03f538c2508b4518"}
{"id": "train_029344", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["data", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "276ed2d17cbca25ac2d126c95f5beb87adb73ba0"}
{"id": "train_029345", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "6a22999336b64388358e41faa20b0e78d2646d7a"}
{"id": "train_029346", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "0d327cfa51a5b2c61da65cf162e6e2ed40f5a635"}
{"id": "train_029347", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "92ef1d0293b35cb727af88b41f0268b0de858cd5"}
{"id": "train_029348", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "049690ccfe069c94a1135dd2330558dbfaff75d5"}
{"id": "train_029349", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "78ee7058c1e51ec000a1388aefe46c00eb0aae9a"}
{"id": "train_029350", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ea797ea0e5deb124467dcab099129af29fd1a83b"}
{"id": "train_029351", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "529c2da6924aac04a914021bc1ab36677bcc4660"}
{"id": "train_029352", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "b54f110924dd1b0abf3266ae5b3996a5d7cb82f5"}
{"id": "train_029353", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "10061fe6fc31177307aeca0f0bef2020192a595d"}
{"id": "train_029354", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "aa90e882ecfd6a8c87a3d36c77e0f8538a380b91"}
{"id": "train_029355", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "78cf6addfdc5256989cce4018819ee7321437a39"}
{"id": "train_029356", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "5e178342408bd8f2dc9fe423d62f1304bce47245"}
{"id": "train_029357", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "546af586271a2e78f22c51a08efaf7db5a77dbfd"}
{"id": "train_029358", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "3339f155ec240b898086c448721d0e31d74bbdbc"}
{"id": "train_029359", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "a6cc73a86f983ff9ec14cad436db506b1c790bc0"}
{"id": "train_029360", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0776f2e8a58ec06d09508ec2b82de253413ea959"}
{"id": "train_029361", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "3a896971c99c97fc2338517b77c1521ce5aa813f"}
{"id": "train_029362", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "afe31f83904180ba0ba9fb8c0d2de91669fb15bf"}
{"id": "train_029363", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "a13fd4d6a72990f9a94fe81a1b65a711b9b0d830"}
{"id": "train_029364", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["data", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ce0e604b4a98c98e9c98874f84bb5e2cee0c2fc3"}
{"id": "train_029365", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c12cf3d3e069fe3d4a7295a6faba1d9df712f61f"}
{"id": "train_029366", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "47e53c9aa41d4892073f3708a94bc933290a40c3"}
{"id": "train_029367", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "03976419de663a324e6e887f5327f2fe85039426"}
{"id": "train_029368", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8dea89501e07ee3a81fcf324dd03ceb70b526dcd"}
{"id": "train_029369", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "78f6430397b75ece3a78d079e73e995d08986122"}
{"id": "train_029370", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "057e66004362b2e784ce637df79ce62245dd4069"}
{"id": "train_029371", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "25bf353a755acbdf53fc85de492aa8104862af69"}
{"id": "train_029372", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "098195dc70529122c61218b685c5c037b5b46571"}
{"id": "train_029373", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "43fed9166b8ffabfceb049547c7b0687da114fae"}
{"id": "train_029374", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "79046f5e81eaca49b9b4ec0f837d891849bda8df"}
{"id": "train_029375", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "e5b2d2ff1d71905ecb82fbad4e70346f57f03519"}
{"id": "train_029376", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "32a7f4347b9d2c0cd6320f35b9b17158fd2c6e8f"}
{"id": "train_029377", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c8d0783c2ec7efada6eeab6751b892c0ac73624c"}
{"id": "train_029378", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "b3ff06cb8c154e257b3dac731fcab3d4f90a863a"}
{"id": "train_029379", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "3fcfb570a349e19b0419c3c32faeb7d8d0c2f7a3"}
{"id": "train_029380", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "7130b30995d49006d37692d9c5c982f0896c0a47"}
{"id": "train_029381", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["coding", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ce3478a2b7e7b7e3c4fcea37c3626e8b12285e02"}
{"id": "train_029382", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "034664b2bba7d64ea5881379794af9c27abb39a7"}
{"id": "train_029383", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "1bdc37d4b47d89a78f2ade1e858524c3dc6846a0"}
{"id": "train_029384", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "48d60c823df17306c1ccebe12c38b6c1bc03e1ad"}
{"id": "train_029385", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e719eb1f3c3a0e62c042789b718a1c4386789a95"}
{"id": "train_029386", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "abb0ec6e65f5985879b6208d86d14c1f7834b2b2"}
{"id": "train_029387", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "85fece1b48f9d4f8334d106d1939474d47e4bc5b"}
{"id": "train_029388", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "3bb348a83261ec93f1d433c1d4b75ec4fe5d9b76"}
{"id": "train_029389", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ec6cf59ae9f612a2763d1296b0418f57d5cca4fa"}
{"id": "train_029390", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "b7708d2695534a5f05a8fed59fc67d5c23153a64"}
{"id": "train_029391", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "fa85f8e0edabf95354addf90e38545ed29af96d2"}
{"id": "train_029392", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "324ef87a9165c1cfd2fc6a57cd4feda15f3923f2"}
{"id": "train_029393", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "05bf3f3c0b5519bd49704c1368a85a8b799613f8"}
{"id": "train_029394", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "589e33598eabc8f0bfaaf3747f1df158b6c56e0d"}
{"id": "train_029395", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f082ff81a8456cf279ede37563b6f846e0acb549"}
{"id": "train_029396", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "445ba238370a97c70114a91bd8a518ba68a59668"}
{"id": "train_029397", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["governance", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c23d55930d6963f172752017bd52edca7c7c0bad"}
{"id": "train_029398", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "e58c86090df842412820f3ac22be4969d38d3342"}
{"id": "train_029399", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "93691cbf3cf750f24629f60fffec29fbff5d50e0"}
{"id": "train_029400", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "62758fc927b1f40968c0b7a6ad41dca98d415c1a"}
{"id": "train_029401", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "fcc25575bdbe6f2cab48d82547764ae0245260ef"}
{"id": "train_029402", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "c41ac12bdae55b93d47c8267fc5675db4e9adf78"}
{"id": "train_029403", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "88a7ba23a9a48028755baa1716f4b8a432b53ca5"}
{"id": "train_029404", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "1540d558ffcd7a1b6df058864ec75a7c354f7d14"}
{"id": "train_029405", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["coding", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "f7eafae2c5f202bac64ea29e5de96b5fa52f7b9b"}
{"id": "train_029406", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "e93aa2a27fb7a0a25f6a3a82dc870d55f19cd706"}
{"id": "train_029407", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "9d9b078766c587489a4c4c1d35b5bf307a7d6732"}
{"id": "train_029408", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "99efa2484c2f32da3976f52f97d64f5a9535b7b0"}
{"id": "train_029409", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "7bc6d2a6361168697e80f49f0ba148edbc8e65fb"}
{"id": "train_029410", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["coding", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "eabe69670a2c1c55c04cc9837f93a85befa013e7"}
{"id": "train_029411", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "ac9fc62e99a5136b59690f7a28c165eb9ceb7942"}
{"id": "train_029412", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "f9e2f3d184f2ef1b0c7f38dfbac3fc8d5fe2c232"}
{"id": "train_029413", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0114584cf8fc207c2513484d7b8db48f4a35fe59"}
{"id": "train_029414", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "dbe281be2a9cbe561830d01fbbf6d0b5b28c899b"}
{"id": "train_029415", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "fd499a704f15acaae9798eb819997a4746fb0261"}
{"id": "train_029416", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "2159802635f317f29013dc23e6096ffd2796b733"}
{"id": "train_029417", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "4fa060aa95068d2dc1f35e74ed965070ed8a8709"}
{"id": "train_029418", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "7cceba65bf28245220a9297d1119701bcda1f57d"}
{"id": "train_029419", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["agent", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a5a1327cbae1387934643ae308805164f1a5459e"}
{"id": "train_029420", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "325c048fb92ad616d4f3ec46351e1003669d9057"}
{"id": "train_029421", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "bc511480d625049aa6b3978cdf6fb725b6cd1a17"}
{"id": "train_029422", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "00002db1ac8f904315b39eba6376ba8cb73bba2e"}
{"id": "train_029423", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "a0d794e2373ab7f43458ba4c509c7314984beff2"}
{"id": "train_029424", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "33b787a933f1346ad72ed21fa40d20464d8d4358"}
{"id": "train_029425", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "708e5058ca8f3432a6380dbb28a4037ce704eb19"}
{"id": "train_029426", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["agent", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "390e1aee0c43b84f693d0410d313f56806807e08"}
{"id": "train_029427", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["security", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "75b622de54d3c91cf3cd5389cbf0855b3c470605"}
{"id": "train_029428", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "52c80b8077777f681a13b009349581e3d8c1ba46"}
{"id": "train_029429", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "b24a6cd1c9da599fa37c4f965aebab56e074298e"}
{"id": "train_029430", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "1891d06b4e17cb6e2d7a3c77efb0ceb693461e43"}
{"id": "train_029431", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "8157cfeeeb35acc615c43e618242d9ed08ffa6ad"}
{"id": "train_029432", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "e135ebede4d6c1cdc4773a23812702ea5bcbc8ab"}
{"id": "train_029433", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["governance", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "8c60bb6a907b93c8b7be7aa7373db9f9283fce27"}
{"id": "train_029434", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c9b241958be6396827ca0feff7969cd5cb2712fd"}
{"id": "train_029435", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nCode:\n```python\ndef agent_loop(plan, edit, test, issue, budget=4):\n    cost = 0\n    for _ in range(budget):\n        p = plan(issue)\n        patch = edit(issue, p)\n        ok, _ = test(patch)\n        cost += 1\n        if ok:\n            return patch, cost\n    raise RuntimeError(\"Budget exhausted\")\n\n```", "metadata": {"language": "Python", "moe_experts": ["evaluation", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "1f9dc8759ba45c2aaee65c9bc6d7ebb6d1e535f2"}
{"id": "train_029436", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "c9b241958be6396827ca0feff7969cd5cb2712fd"}
{"id": "train_029437", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "0a48658d163330d34e0e3e15af566188dba9fa3c"}
{"id": "train_029438", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "e27dda8422c74bd772a791a9bc55581e52e038c5"}
{"id": "train_029439", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "db2bfc36a31f333a50dbed2d990f76cb0c14af6d"}
{"id": "train_029440", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "cbc429a544baed70a62a8659fd7beb838f88785c"}
{"id": "train_029441", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a6588a053bfe7a14b3f359980987809c93cdf5b7"}
{"id": "train_029442", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f66d204619127df84652532250c0e1d90f60d784"}
{"id": "train_029443", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "059b858667e46769576abd14eb25551ff0f9bca3"}
{"id": "train_029444", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "03f5eb1c26b37718c636389418671be0cdc59d5c"}
{"id": "train_029445", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "54b81a2c06d878f70f79dc6238f9854fc0e5a735"}
{"id": "train_029446", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "1c60ab7a201df4679a6ed7244d3e65e6f388fec6"}
{"id": "train_029447", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ecf68be4913f3de8b49f6add502a2e3dc4802d64"}
{"id": "train_029448", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "ef1277b93b77849cb02897a44a7e60c09e8719bf"}
{"id": "train_029449", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "61f954bca15299274332459e7b66f12b67679e7a"}
{"id": "train_029450", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "8c784d393c1af6619ec32b2c02561c82a1a245c5"}
{"id": "train_029451", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "fc9b13f213c63cf6559ecba9e44affd4f46bba14"}
{"id": "train_029452", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "dd8c212cfcf71f359d7e30806997c9c5ea198f3c"}
{"id": "train_029453", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "dcd31052e98dc45a8189d76bdef53ff322a2cf45"}
{"id": "train_029454", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "15454d94cd9dc537f4953d80cc80d05dd8530510"}
{"id": "train_029455", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "a64f2796365a21052c264ff225f031202cc5dab1"}
{"id": "train_029456", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "d7ca8155eb72d41d7178783da791f0c96588ad18"}
{"id": "train_029457", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "1bba0756a3d4bd69e88a70d01f11eaa61ae8f274"}
{"id": "train_029458", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f8c335719622279152a56d563cf0a3efbdc272a1"}
{"id": "train_029459", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "335ea61cb14c7f2b7523dd0033acdc8ce7cc2201"}
{"id": "train_029460", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "9dfd330b5595c10aeb933e826a1a0c753a29f593"}
{"id": "train_029461", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "6c5c9e32d6b927a0d5e391d80fa9067f3de4b763"}
{"id": "train_029462", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "bdd72856597bdab9f988d7172e209fbfbc6da6ab"}
{"id": "train_029463", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["performance", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "4063ed85825cdf52671c9dd8a90bf7c54a9ab052"}
{"id": "train_029464", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "fcfe53791b9be96720839478215a94cc3c5c3615"}
{"id": "train_029465", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "8f6342c3e921dad31cc03eff7f5e94e3211e4c90"}
{"id": "train_029466", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "fb2e7132bb0e59090c5558108d29ca479a290313"}
{"id": "train_029467", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "b67c6d9a046c81ac956a829a2e72347599a229be"}
{"id": "train_029468", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "1e1883369a89c849d6e9ecc18bb25537405472c4"}
{"id": "train_029469", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "764a9be7768dff1ac9199a91dfa3013fd0ed4b42"}
{"id": "train_029470", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "bb2377646bba88835a67225bb50df44d7d3dac7d"}
{"id": "train_029471", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "afebfb889043549eb2d20f1eff15b29913184e2e"}
{"id": "train_029472", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "44b604d04c085eeea2e6adf09d68c82b458035d8"}
{"id": "train_029473", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "1ef8d01030d12f43cfdec4e5fcfa6e6f81230445"}
{"id": "train_029474", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "3bad62da9c3792a40648b0e1f7158f392e1f388e"}
{"id": "train_029475", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "1939a5e7d8b79a6e7ef694713b94b3096468d3d3"}
{"id": "train_029476", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "531469944fca36a8c88e3b5e754b6cff8967b341"}
{"id": "train_029477", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["coding", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "273f458ace3960f2ec5254de4fd8b0902c12be19"}
{"id": "train_029478", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "6b58325991dd3f7163cbf5895582d14dbb545322"}
{"id": "train_029479", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "5ccb4e376dca7ebe6a9c7909349bba0fce838d48"}
{"id": "train_029480", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "63cd31b936571754b15af485b618c6a5c879654d"}
{"id": "train_029481", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "7ce100daf917d1c2288a4196e44754518ac6569a"}
{"id": "train_029482", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9697bdb3c594a1d41ef1d43906825764bd03507b"}
{"id": "train_029483", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "75391582fabb2f87794f2f305ff8487d5b0fc2a5"}
{"id": "train_029484", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8a6bdef1422a98cac26e4a84bb5e3f9a232e202b"}
{"id": "train_029485", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "8d83ad86ef8ce764ea9af59f3a36e94792e21e91"}
{"id": "train_029486", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "5207be653eea1d024d896207c4dc729d7c065dfe"}
{"id": "train_029487", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "badff359b8d3e8c7787086ca90cc7230d7cfb9ca"}
{"id": "train_029488", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b7cb5c79b97118180c486f2cd7db5dda095c7d46"}
{"id": "train_029489", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "6e4e72f7dc36e9d63150e2508bfdfe9821de292e"}
{"id": "train_029490", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "871d5ecb73722b715f279be2905cfd9f448ef4a3"}
{"id": "train_029491", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "7875d759c8eb131cf8f9efec42252ceb0c70cf7e"}
{"id": "train_029492", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "eaf0e21b4d89899d3e98cd7c9e388188ba0b928c"}
{"id": "train_029493", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a6e437c9d709ed20f8efa1a3e91209a0f7e1a88e"}
{"id": "train_029494", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "a10191c11a1b08b4512466da8116379aa8a39fb6"}
{"id": "train_029495", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "e048f6b0343bfbb24093da390e100781e38892ff"}
{"id": "train_029496", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c0fba6fe857b3cb1173e5547ec9bb1c2cb5acdae"}
{"id": "train_029497", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "25baceac7043f47728e733e4e8a6d1e17876d7e1"}
{"id": "train_029498", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "45a08087d6aab872f3820350d9da0a696f97e2d6"}
{"id": "train_029499", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "25e304abf5d25c9204059804843a295c8737969b"}
{"id": "train_029500", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "f772af691937f0455bde968372aca1078cdc65cc"}
{"id": "train_029501", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "d53a39ecff1be636686a7f3562f2f49a76d435a4"}
{"id": "train_029502", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "961d0220e8ad017070ff1e72d7632dab3925ce68"}
{"id": "train_029503", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "0730147dff0173be577b846d7568a06040045f92"}
{"id": "train_029504", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "8a1aea630ef7add9ca297e4fb00dd6320cd702da"}
{"id": "train_029505", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "f9ff0813cca58c052b1236640ab52d7feebf0a97"}
{"id": "train_029506", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "74bc04c800824daa1e19ed81b2014a9593a37107"}
{"id": "train_029507", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "a249edfc64a4a9dd4aebb51d060eb72b7d274da8"}
{"id": "train_029508", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "db5e9ba8f950e6ede050ca5e3ecc76842e690572"}
{"id": "train_029509", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "29a01613721047d4b7a33ab10a13439144f355f9"}
{"id": "train_029510", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "3247d3499a743878191221f1b07eac265618f498"}
{"id": "train_029511", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "bf6eb5b4367f51f3bb020d222b75a17c1d43e028"}
{"id": "train_029512", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "bd280ae21ea53567a42c61c5d1aac1cad1518a57"}
{"id": "train_029513", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "0d79316061e63ea30224bf85a2b7c51b67f18b95"}
{"id": "train_029514", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "3e989a07da5eb5d9528317ddebf24c0543981685"}
{"id": "train_029515", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c4639f0eab90e303cd908f95b975a8f1744f0593"}
{"id": "train_029516", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "804f8fcd9a45ec30c54134bfcd6bb0fdf19ba97f"}
{"id": "train_029517", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "beabb263c9bf6ada4a73be63f786d82557f96ca2"}
{"id": "train_029518", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ffa2ec4f10bf3de1bba763aed946558fe8de462a"}
{"id": "train_029519", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "9b7b726e1b6bd89d84a5a30dfdfe120cbf759824"}
{"id": "train_029520", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a8ceac23746c9cbcd6b8133a880b64561ae8479f"}
{"id": "train_029521", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "da187dc434fad34d99ea81122701cbaeabb2e175"}
{"id": "train_029522", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "5800933bf77c37824068a2902675778c7353071e"}
{"id": "train_029523", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "fee5f95f6cbac1134747a93509e0e3c326707e4e"}
{"id": "train_029524", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2ae91c85f768f0a080e68b9d3a40ab35b0da7fff"}
{"id": "train_029525", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "fc50f527cd983137614dc677692d954e1550d54d"}
{"id": "train_029526", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "64651cf03ca4d7de39cebd225571fca3833108c6"}
{"id": "train_029527", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "aca413a18faa722b2b8beff812b1758572b4c2b7"}
{"id": "train_029528", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "f3249e27035feb5c9088d49e4b3ea072c52ad23a"}
{"id": "train_029529", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "672cf6a65211822914f0bbe1d03ead5272aedf78"}
{"id": "train_029530", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "d008029c3500353b1bcb09fdee1390deafc1c1b0"}
{"id": "train_029531", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "20802ba09e303a1edcc8a701cdbc874220caf724"}
{"id": "train_029532", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "3b3be00b18382f50a58f74a7fba86ff676c406e6"}
{"id": "train_029533", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "b89df570eb320fe17bc1aae04f89057168ff8654"}
{"id": "train_029534", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "8573e06815e890d653b899a6dd506b6e655831df"}
{"id": "train_029535", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "8bf8f7842f7937df406a50a2312041f57140cfdd"}
{"id": "train_029536", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "7c94b71c332177f88230ddf0657a6332283a5462"}
{"id": "train_029537", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a2e18dbc8f2638444f052eb6ff296329c0090576"}
{"id": "train_029538", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "1aa832dd82a79935d9f6c034bbd6ba235e690d44"}
{"id": "train_029539", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "4c9897e16869e8936d548258cce49be9df65ed5d"}
{"id": "train_029540", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "d2a29c677b29f7faca338a82ae59d961d7431ad5"}
{"id": "train_029541", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ea12e1ae624e7d45fa21271555d9abf812813211"}
{"id": "train_029542", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "coding"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a5bb6fce36b558e31ea139f3bdb408d37411cc48"}
{"id": "train_029543", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "ab1a9b6385e56ce1eb89d9d55cb87da3c9595ed7"}
{"id": "train_029544", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "7573627b54d9aa4a94153e7e7b51820ecd8f8e22"}
{"id": "train_029545", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "73be284a991302808d9461c94ebf885f0fb9255f"}
{"id": "train_029546", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "94dff5d847d74778a861961ecbf3ca3288664518"}
{"id": "train_029547", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a2e372484588aa88be9d8527b806b6100a847848"}
{"id": "train_029548", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a1432ecfc03e336e917352563ddb659c0b8f8264"}
{"id": "train_029549", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "923533d291f01fce771d2485c8df6d24cf48a057"}
{"id": "train_029550", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "5a93f66acbc8e4e78d2e8d545fdd2b80eac0350a"}
{"id": "train_029551", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ff57ad22365121a01fe54fefadd17953f43d4618"}
{"id": "train_029552", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "64f2871ef0aafc9710074b28bca62195808dd6d8"}
{"id": "train_029553", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "0d79316061e63ea30224bf85a2b7c51b67f18b95"}
{"id": "train_029554", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "656f3d7c68a68fc0953dc16ac5be2c1bd00a1234"}
{"id": "train_029555", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "72d7ba744c8041dd600d33ff139f91b1621f865c"}
{"id": "train_029556", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "6ec0a6422410d5407d08f9b35580722cfd95379e"}
{"id": "train_029557", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "3ab7ffd25d01ee6d810e4e9d01c5800efba90e77"}
{"id": "train_029558", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ae45301aaef3124fb26f0949e9b90c033df32fca"}
{"id": "train_029559", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "340a751b4d7acf2ba1702c8d02aa6b824f3d1a6f"}
{"id": "train_029560", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "f334495ccc8a732ba320c4bb8cfa9bd0dde56663"}
{"id": "train_029561", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "9ca76f6cc12a74a06c79689c7b5b3122ed252c62"}
{"id": "train_029562", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "329c9a65ab736480b14835d5691683e29451be0e"}
{"id": "train_029563", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "b9283878999ec42dcd8c864e4b847ec36942cff6"}
{"id": "train_029564", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "6d59f385c44ecde317ef1fe43edee2b1dd4e212f"}
{"id": "train_029565", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "97ac702dd2621cad7d3080c83406e91ee0ee6500"}
{"id": "train_029566", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "a776c6c813064fdb075bcc9f209e13b871cd795b"}
{"id": "train_029567", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "dd952bf7fbcdc98b1bcae62ffd653e4610e28126"}
{"id": "train_029568", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "bad8286510b7240cc198b05e8cdcc96e5d52fd30"}
{"id": "train_029569", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "5376870639793ef26e3da13b67e5cae209c4cd4b"}
{"id": "train_029570", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "7728f081ea56afdbd28afd62e3fabb8d163aaf2e"}
{"id": "train_029571", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "d8ae3168be15c6d8be50c93c21710c8720bca542"}
{"id": "train_029572", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "7ea14f6ee0567321f547bf0813c3ccad7349c8d2"}
{"id": "train_029573", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "3efbfe5dc53934481e586779ec1366ad998a7fc0"}
{"id": "train_029574", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "b034e64799d298b1883b186a757a4eee158ffe83"}
{"id": "train_029575", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "615a7c57ace559326e3bb81b5d14800b3e30d801"}
{"id": "train_029576", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "2e875de61f7abf7c3d8d7e73a13d14fa7bdc51f5"}
{"id": "train_029577", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["research", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "0899d484a49a1d52fcd139ff2050981c0d475c62"}
{"id": "train_029578", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "c974ad536e7908ccdd6b80b80f7ca5a3f7cfaecf"}
{"id": "train_029579", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "b86c5fe7d14bf6da58fff644c6d7586b51aeb22d"}
{"id": "train_029580", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "5ba4de1fa720fb9e8f6aac290f175c1b340ef9a3"}
{"id": "train_029581", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9e11bbfd7f6173ef4893b23a43009fcbd3fe7ed2"}
{"id": "train_029582", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "03daaf7e52f400ebff8481878c7a5312c956bd75"}
{"id": "train_029583", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "af4b863b108751cb1f69f70c2c20bb7311f12436"}
{"id": "train_029584", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ed25d6ac740d0062792685ea1d3fc70c90869652"}
{"id": "train_029585", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "3ca486007b978d0955ecda3197e72cc376dea17c"}
{"id": "train_029586", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ce265b91a5b2d9a1c2f6bd275d0787311d16f0f2"}
{"id": "train_029587", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "c17bfc7ed4ee4ce820889e04eff1b9afdc27c048"}
{"id": "train_029588", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "0d7601834335080442244e7a2dcb307f4a6591de"}
{"id": "train_029589", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "d3e2b46255c27a92e0443ec63df864a84001b97d"}
{"id": "train_029590", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "3e0073c07f144878a2384b6443f1e67a3c8f1cc4"}
{"id": "train_029591", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "6f8d679774ebd4f04ca2a9e1a03dca56b3fbffeb"}
{"id": "train_029592", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "4c001b963af9ecde9d36065917db39e0ef7e5375"}
{"id": "train_029593", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "ad0ae4f42e8b4f0f50fb36ff2c60f7db2c8da09f"}
{"id": "train_029594", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "3c068f4dd88cd7e394cec2228f4e3cb2d7c5caec"}
{"id": "train_029595", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "4d84188fd7ff3ea0a798583b71917ae9927a2f48"}
{"id": "train_029596", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "f6ce13cb5d49099160e6d6572e5bca4e73106565"}
{"id": "train_029597", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "33da9e6e6ad33044fca6dac232afd65041aa1b60"}
{"id": "train_029598", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "81ab849d1e608a509c358a854a8c7aa78c8bdd18"}
{"id": "train_029599", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "031135a1eea3c1b6a0224e5ec37e9b205c72cb78"}
{"id": "train_029600", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "e546d68164a14647c899ddaef0ac9028598488ae"}
{"id": "train_029601", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "d7429b28a4fe1b3af141fc4f72a30951522cbffa"}
{"id": "train_029602", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "3b86152527e153e1bd92cad0c2e516043f8aa02b"}
{"id": "train_029603", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f36c83a896e9f15d1911818b0bdb3518d4847610"}
{"id": "train_029604", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "16e2614d53ac4707a4ce46a92ebefc665e239f1b"}
{"id": "train_029605", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["coding", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "d23123b06a678a8c781ae23eeae55d8a122790a3"}
{"id": "train_029606", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "5ff23b45c3e2fe7d256444916ce76564a350df2a"}
{"id": "train_029607", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "cf9b5cfc378f414ed0d1954a4e0c7900da1af670"}
{"id": "train_029608", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "8231ed40c212f25b1f3079c057477b52576c1fd7"}
{"id": "train_029609", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "4cb2d692e0ce6b5b277d42ecb091a5facf172bc0"}
{"id": "train_029610", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "2220e9037b8f9861a9e3e93cd569c8beaf5d5ba5"}
{"id": "train_029611", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "00212f0aede1929407bbe314a48855fd09c0a447"}
{"id": "train_029612", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["coding", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "924f4373aced7e0cbc68d1a2c026e0c5e36f39da"}
{"id": "train_029613", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "1430fb6b7ce7e2cb070f66ba05329cd6febdca7d"}
{"id": "train_029614", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "0bad4323a0af4e0ddef61538cd48e7031ca8f76c"}
{"id": "train_029615", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "91125ae2022236376df9a92a10c927b79b6379be"}
{"id": "train_029616", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "bcee7e04702005b3bb391ff85e037370f894950a"}
{"id": "train_029617", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "243471facbe2e3d9712064994a48bc035b430ddd"}
{"id": "train_029618", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["security", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "d53a39ecff1be636686a7f3562f2f49a76d435a4"}
{"id": "train_029619", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "60756845b11b36cb8a74c27319d1dcb864d638b2"}
{"id": "train_029620", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c0e6a289bfb58b189c8ea9c33d2759d8e6a10362"}
{"id": "train_029621", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "7c695b37a770efc74ed866dd503e1a079898674c"}
{"id": "train_029622", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "6b077c53eb58ca22de3382578910911d195a3263"}
{"id": "train_029623", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ba0a9aafb7e6aa88d972e11284e58bdfa7e19128"}
{"id": "train_029624", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["agent", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "105d5c4621db1484e4e597b20509325cf36b0ef4"}
{"id": "train_029625", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "3849969f940c0fa51567f217d291c752549df16c"}
{"id": "train_029626", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "832cc9da58606dc8637e276df40dde1f078593de"}
{"id": "train_029627", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "ccb9cd391fb3a4737412fefb49803069ce081894"}
{"id": "train_029628", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "96d10e67ebda9a25ea51dad6613e16a080c29135"}
{"id": "train_029629", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "643dd91f09f8b50c9b9915f411cf3eb2928fde29"}
{"id": "train_029630", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "63d6a5183fe9568b70c2b9349d8324e9c3072504"}
{"id": "train_029631", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["agent", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "37789d3918a51788ecaa6c53182a5e424b99bcac"}
{"id": "train_029632", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "189e1489df6b93c0b84a5366e439ea734d4e67eb"}
{"id": "train_029633", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "0f23cf7c158f6684a45a839852650346b5777577"}
{"id": "train_029634", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "4cc6861e77f1cd0f40105a839fc01156f8378878"}
{"id": "train_029635", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "38afbf1f77e820baff489d13a1af4b59850a6e73"}
{"id": "train_029636", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "954f3976e0babb2d989f9c6d95b3bdec442bdccf"}
{"id": "train_029637", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "e70e2d7a970dc0a98f7e845fa5d70ce469cf9a39"}
{"id": "train_029638", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "75150bba4e8e2461e4085d1e8ae55f0e71b7d23e"}
{"id": "train_029639", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["agent", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "0fc9d9dd49ddd5faeda70503502a52d027da56ea"}
{"id": "train_029640", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f007f2995b7ceac2f73016122d09f279ab015d51"}
{"id": "train_029641", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "d1d44ec39684f68471b1f427f17d6934cb75babd"}
{"id": "train_029642", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["data", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "7324d52dde97ab46fa9f1465467f3c981a78789b"}
{"id": "train_029643", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "18f4e3c742570b5a1c295cc4a15230119f86f818"}
{"id": "train_029644", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "b5e7e85eb84aaa73ba4f5c604e0cc221cf4275d0"}
{"id": "train_029645", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "27dfd2ab45d5f0ad233f564a675853eb33f080fc"}
{"id": "train_029646", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "0ebee5c84505e5854d3476e9607aee69c2be7c09"}
{"id": "train_029647", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "e23252377cd12e1bca2f8c9b31aa902d5ddbe4d7"}
{"id": "train_029648", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ff7e221ad09aa4d39d045eda453598af05b55037"}
{"id": "train_029649", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["performance", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "b73c3f7304c86ed74b338611de7250d09235c274"}
{"id": "train_029650", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "871fd0d1f2343a0a6005548d49fcc0d4e16b5662"}
{"id": "train_029651", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["security", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2b86de57dc8983f279e9cea677bb888d1746ba55"}
{"id": "train_029652", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "b072b044ca78c17a848e71fab88e20ca59a6770a"}
{"id": "train_029653", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["performance", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "4f195faf2b2d6384b4ef4f9e166e18f367a7e0af"}
{"id": "train_029654", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "76ada0ff613cc1b8e3222a81b118789ec4035a91"}
{"id": "train_029655", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a7f21763f3f8fdae68f5cc1d1ace3c0fe0acfca6"}
{"id": "train_029656", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "839b2b211adc2c1c9088a047882ac7b9739e13ab"}
{"id": "train_029657", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "006aa5125fdf3283f2f2f8aefd92daf69eb84994"}
{"id": "train_029658", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "b30aa3ac9369968dcb615827cb0f0258cadbbe50"}
{"id": "train_029659", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "802c77246ffcbd44c2a6fcba42e945888f589158"}
{"id": "train_029660", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c56a2ca0977cafbffb053a84011815680ee61913"}
{"id": "train_029661", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["security", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "e643a405f526fffeebdd8f3673da0c6031a3deec"}
{"id": "train_029662", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "185b987cc7723820a626db33c901833dd2cabb13"}
{"id": "train_029663", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "3475bf55a6badc4067cd4c2b5f92e2a7f6ac012e"}
{"id": "train_029664", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "648a85ab7b8a6a189e3e684fe1eaf4957dfde675"}
{"id": "train_029665", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8c5df1ed2612b2a755173d3572799e01c7f18753"}
{"id": "train_029666", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "3b68b41d5c15b0a13b248c9d8eabbd52383ea6a0"}
{"id": "train_029667", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["performance", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "15886ea4b8fbce078762bde40bd2145ef1100a53"}
{"id": "train_029668", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "e98a156026bac626e1828409fa8149180d5db277"}
{"id": "train_029669", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "85c490cf7c3f0e5c326bec8f56b618fe526a6966"}
{"id": "train_029670", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "9f9d9e0248b63441ef5af60f7292ec3842fbb0ed"}
{"id": "train_029671", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "bbf1d210d0406f6d3372adc40aef6a06d501cfca"}
{"id": "train_029672", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "beb5acf4b3237394aa0e7491f423a7cb7fc89da9"}
{"id": "train_029673", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "d9e74981935432d5f0541cc4433242079b2b5130"}
{"id": "train_029674", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "6d85899bf14a27e9cd577fa71d0f6c566033b368"}
{"id": "train_029675", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "ded5ee4454fcedf360ba5a6167a7112fd11d6265"}
{"id": "train_029676", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f0e4d2e32be86a4005d8aa3aba3b183145034df2"}
{"id": "train_029677", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "12aadfa00a0c5cd91d1c0dd498d0175ae97e559b"}
{"id": "train_029678", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "20570622e50bc71c601e5727cecdf844fcd8d7f0"}
{"id": "train_029679", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "45bb948b381a655dabef03521f4e689d1f04ab12"}
{"id": "train_029680", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Java", "moe_experts": ["performance", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "57e1a2f18f725d2cba3d59492884cfdc6ec7920f"}
{"id": "train_029681", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "80828d3b771ee18f8edd5c153eb47607651ac844"}
{"id": "train_029682", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "67488cc01cd0e19a0cd50aef600f5cde14ce9b03"}
{"id": "train_029683", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["coding", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "dcfdece8466f01bd6a94d90c25c889a98f5cff1b"}
{"id": "train_029684", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "80629acbaf9ae6d87e6007805d88bf91b777daf5"}
{"id": "train_029685", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "3b20f09d7c8cf84fa6f0ff3129e3f3b08be3d5a5"}
{"id": "train_029686", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "eb0e8a12def9e8bfa7e13278efffa901e26df61e"}
{"id": "train_029687", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "13a6b5c6861160c4b5860b35a0d0c8d49f8cbebc"}
{"id": "train_029688", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "ef0ff2783932194e58863b59d2222eb4b3b13deb"}
{"id": "train_029689", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "591238f4af1a84b64c8615d2176945ecee268b15"}
{"id": "train_029690", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "6cce94668f3f644857ea18bac2cc9b1796159685"}
{"id": "train_029691", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["data", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "ecd516f6472b93c68f0cceee4a86dbde3cfd719b"}
{"id": "train_029692", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "dba0ed4cd9a71c6a1b123d8deb8840bfbe779f85"}
{"id": "train_029693", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "98564065d328819074191e11af5106eb4b6a2136"}
{"id": "train_029694", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "d32b633f95f1a5af35ea564c5fdee07c39cb997b"}
{"id": "train_029695", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "eaf7cc440439b9786cc081de38a3eb38e8904e84"}
{"id": "train_029696", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["agent", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "1a9fc1ef956a9d0a14aa2f979cac50ed9697ce40"}
{"id": "train_029697", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "acc12ec69f7f6e28b6056bcf1f94f7bf2d649a70"}
{"id": "train_029698", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "301ba6b640c8fb424c7538cf14537d7b37fda008"}
{"id": "train_029699", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "d42127c1975c2a1d7144900213980d186ed8ed1e"}
{"id": "train_029700", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "6a1122c2b97c6372005bd24b97fa2a80b8f56d27"}
{"id": "train_029701", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "d648889524a4258b0722e1532638e25412c8f4e7"}
{"id": "train_029702", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "a44db2c6d3dbff7ac7d16a6e6751d2d5ba4bd322"}
{"id": "train_029703", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "ba554b1b1c30b5a087224f412aefee2dbd115bd8"}
{"id": "train_029704", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["research", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "9435b31d62a85d61db88c7ffbd273e24ccd68f41"}
{"id": "train_029705", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "df5c26216b6fe800d0d9066b00cd09145c594cea"}
{"id": "train_029706", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "00347b40b709eca111ad6e4b65df0c0c8a1b2b81"}
{"id": "train_029707", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "623f506288e0e8dc95700479b1f826996d89e964"}
{"id": "train_029708", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "3549d0830e8174ab3ce8738324f8882a08c467f1"}
{"id": "train_029709", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "745c3671c70a5a8f281386490c84ef4e928bab1c"}
{"id": "train_029710", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "a6f05e0178ffd311d280207ed0ce916097612714"}
{"id": "train_029711", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "3ab39ef108ed501e4b57447f7683cda6891de395"}
{"id": "train_029712", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "dd85ab6ee7137cb3342793095d1a89e2a9adc72f"}
{"id": "train_029713", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f022528a3aa726dc8a0ccd2e797df0afad6bfad1"}
{"id": "train_029714", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "956f9ccdd45312c30f2bacbee54498667a9bb939"}
{"id": "train_029715", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a5cfa2038956f6fd1cd00167e53001048dbf54a7"}
{"id": "train_029716", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["security", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "cbbe3fabd3d0681bed6ceb3dca0703da33b1fafb"}
{"id": "train_029717", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "281ce2e057fa0feb172f3e90adcade717a295466"}
{"id": "train_029718", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "bfa580a39ae2cd09dc9fbf2e44faaa548633f775"}
{"id": "train_029719", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "d98a9250748d52d968707f6805b763697ba49f63"}
{"id": "train_029720", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["security", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "78ba56a43af2f3af61424cf9cc8424cff79c329d"}
{"id": "train_029721", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "eb5ed1e0bf1514c0b1507db8c086698c851cd518"}
{"id": "train_029722", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "a7f21763f3f8fdae68f5cc1d1ace3c0fe0acfca6"}
{"id": "train_029723", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "601fad7380f323cfc2beffac7cc2a5d66358e00e"}
{"id": "train_029724", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "35cd85b9e586668bdf9727f98d5aa0b7be1a963c"}
{"id": "train_029725", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["agent", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c59c33dba3ac0700ba0142bf8f3c483a259b48e0"}
{"id": "train_029726", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "TypeScript", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "cfc7a750c4c753acea3231304368b9ab2662e380"}
{"id": "train_029727", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["security", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "fd044e02593d76f397c98095accc0f7de391eaf9"}
{"id": "train_029728", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "a47c6959465216b510354b7bd7d419857b76de61"}
{"id": "train_029729", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "1e43a9eff46635a2597ac611010e182e591d52cf"}
{"id": "train_029730", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "8a7af8253fa1c1e2b7fc38affa7037f70498d4dc"}
{"id": "train_029731", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["agent", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "847e4030a0d3b7a8cef1bf2034d71ef04671bd28"}
{"id": "train_029732", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "c9b8325254fd79e3b0f08a0557792dc093c43cc2"}
{"id": "train_029733", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "adfb667c13d0f63571cb83fb0d562a68d53fe561"}
{"id": "train_029734", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c58597d7da75faebd1aa9f4a640c2e911667ea4c"}
{"id": "train_029735", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "e9fbbf3dbf44aa1af2ef4d160759ab15bd42292a"}
{"id": "train_029736", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "f8289300c623cd2281fb8c758407aaff2ba4f920"}
{"id": "train_029737", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "bd0592448b31ddef29e8d764c91c687d782c5099"}
{"id": "train_029738", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "92145bd696130031f4d7029bf9be0caf9c95c3bf"}
{"id": "train_029739", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "7312203d1f0ecd2d3013c0755214d761b4e80f14"}
{"id": "train_029740", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "a1b994659bd2665c55aa8a8cec205432b15757a5"}
{"id": "train_029741", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["coding", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e31bab7f4186c2be83673bf5075d49d00adf3e41"}
{"id": "train_029742", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "af78777682407608382d401b1f65467158b64153"}
{"id": "train_029743", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "5ff23b45c3e2fe7d256444916ce76564a350df2a"}
{"id": "train_029744", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "ed4d9365fd48764aa83c63c4aec6f0f8718f9d33"}
{"id": "train_029745", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "626cb477fac7e47a44e52258cee5b9d808e78e62"}
{"id": "train_029746", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "0821b4565a8050b5a6efe5a856ee08196c43d7f9"}
{"id": "train_029747", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "e8aac0c5c9571d102a6710e23ab30a5402633872"}
{"id": "train_029748", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "19e915d30bef0aff5e28bde3bae05135db8b8d11"}
{"id": "train_029749", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["research", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "2074d1b96d9d6fc7dc27902720b68ddf1fceae0b"}
{"id": "train_029750", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "273d109727efb9dc97847f50edc2f44fefca80df"}
{"id": "train_029751", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "f770d2268b6d87343ec0710a52506f80aa1305bd"}
{"id": "train_029752", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["governance", "security"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "504a1aff4c3b0c03fb482cd60c0835dc80799299"}
{"id": "train_029753", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "0dfafbac7fb3f52dcd57f77e65658b1ebca16d73"}
{"id": "train_029754", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "b11f7a1b869f5dc92ee01e908abe6619a9b9dd8d"}
{"id": "train_029755", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "f65c44f0d7a8cd8fb5348a2ac6e6e0d3fe01208c"}
{"id": "train_029756", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "7063e9ffe5a8edd9535a11e45290413fe9fdfc41"}
{"id": "train_029757", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["agent", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "22a279dc8fa4dd73ac0c8fe1db982861d1df6b65"}
{"id": "train_029758", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["coding", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "4750a66eb4ac4b0c71b9870a2fee97f5ed4a87f9"}
{"id": "train_029759", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "58fb4cc3eb5458c72419de7e322c89c91d3da0e7"}
{"id": "train_029760", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "2a0f09b4090a07a9f513a51d1230bb418b96a9e9"}
{"id": "train_029761", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "e6d1a9b0cb6032da75812ef1140e25d06efe990a"}
{"id": "train_029762", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["agent", "data"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "cf994d567be2c159a084fcaf5a7d76d1534c1fc9"}
{"id": "train_029763", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "22568f77bd04a421ec0433bcbccd89499ac215a2"}
{"id": "train_029764", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "dcb35585ec497b593712574ac468a61fac8e19d0"}
{"id": "train_029765", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "a78f456a8c0a8140838528e7d02f300e9d7e6d04"}
{"id": "train_029766", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["research", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "807bc5165677cee765119da2ce449678f1e727e4"}
{"id": "train_029767", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["evaluation", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "279788254cf318b8fd5749eb754dcdba01625326"}
{"id": "train_029768", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["security", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "9a4f2cc7ec1795fa7a502bb94c88bde20d864d92"}
{"id": "train_029769", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ac6a1e826e01edbba9bc1d81d2caba71abcc2f54"}
{"id": "train_029770", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "3c0b07085272557e1446baba0c4c52874722457a"}
{"id": "train_029771", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["data", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "8ff6f06a2562883cc7554ef77a0ea7d21c799e3c"}
{"id": "train_029772", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "381b48a3960591eb9e555734c4f4b40aefa646e1"}
{"id": "train_029773", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "8bf2457e21dd9e209fa33c8b472c215362b53819"}
{"id": "train_029774", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "db361ef9a69f7ffbc0b43319dd1ba297bfa7471c"}
{"id": "train_029775", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "65fb16b1f27693c20d4d60a09aa046556a12f26c"}
{"id": "train_029776", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "85c2af5a32a94e39ca064332e87d44db6c8d11f7"}
{"id": "train_029777", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "4f1817cc1be69b7662fdf1de4da125570568dd80"}
{"id": "train_029778", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "08b6e2b6cba0a05c503ee4359b4c1a3837193e44"}
{"id": "train_029779", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "703a109d0a6e00fcb3a5f4e3731fa00d5a03ffce"}
{"id": "train_029780", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "ff1581c9b1af595319f9aa5eb03a0224879bdbaf"}
{"id": "train_029781", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "db9f191d375e238b941ab2843856c071b7291fe2"}
{"id": "train_029782", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "6003489d3453b4fe520a3a49a34388791013f071"}
{"id": "train_029783", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "11149c4b1237fe8581234fd99a5d68b97088ffde"}
{"id": "train_029784", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "45e13dcdcb1d2183ede6f6102d22ff1910844214"}
{"id": "train_029785", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "aff4052573abfa43b5cc1a0e020216cadb4698a2"}
{"id": "train_029786", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "871fd0d1f2343a0a6005548d49fcc0d4e16b5662"}
{"id": "train_029787", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "bdf0414ec167a3afff51a645c6c8b5806a37cd6f"}
{"id": "train_029788", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "7b214346998baf0b7373de99197731dc0ad371f0"}
{"id": "train_029789", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["research", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "af2a45933085dae5744b3dd1fa8a99556ba9259c"}
{"id": "train_029790", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "16837312ef99235246d1f3d657c6d35648c7c788"}
{"id": "train_029791", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "evaluation"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "3816ce2136d336dfd7905932b8fcca3997bd4b4a"}
{"id": "train_029792", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "01b376f2f65db9a053e9f1686f433dc7820fff5d"}
{"id": "train_029793", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "eb5e3346b6b26c6d2b69bc9510e89033c6973d59"}
{"id": "train_029794", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "60263ebaba7e398ee5751a757fa6c3d311f79d37"}
{"id": "train_029795", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Go", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "919c939b953ba128b662b608ede6babb5014bfe5"}
{"id": "train_029796", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "ce566587b46d9dcdff60738c1ae45eb3cb4e0200"}
{"id": "train_029797", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "3092af9113d4f025e3e4ec86fb9241f883a738a4"}
{"id": "train_029798", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "208d9709e1c3e80afd8ee6612c9ea4fd0e5dede6"}
{"id": "train_029799", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "a872f7b2c2091f1c226602bd0383f7364ee9616a"}
{"id": "train_029800", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "da8bcc1883d19a730360ec3df08e3c71b2cff5c0"}
{"id": "train_029801", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "7ee534ca9c35ff8e731a4f1276e68d728d8d82bf"}
{"id": "train_029802", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "5c39b5a93a6e304e479c38c96aa91805af226b82"}
{"id": "train_029803", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "C#", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "42a935a96c03f4d649d6fb7d0623a9f1d3f5b580"}
{"id": "train_029804", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Bash\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "7dc3beddddba031838126826f7387a3f6f20df8c"}
{"id": "train_029805", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "fc4f0fbf55ca578218946e64bcc924428dd6c028"}
{"id": "train_029806", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "ce789b865f393dae3a63c1bb6af1ea938ea3314d"}
{"id": "train_029807", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "8041ca5c9195f62702f2e920671758fcd5f980b2"}
{"id": "train_029808", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["research", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "17bd03ebfb420417a7488fefa789d213cd66a86c"}
{"id": "train_029809", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["security", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f032e42b800f15b1d45d0331bdbf6fd8054172a1"}
{"id": "train_029810", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["coding", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "d7c0035e5bd48c25f91768b50a10296060821fd0"}
{"id": "train_029811", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "performance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "1bfa6cd02277e637eba882f34aeb2a75822c19df"}
{"id": "train_029812", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "4a1eb54c602015619f07d4397fa639978863546d"}
{"id": "train_029813", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "ac796a832868d00b8f93a39cf7cc4a44dd6e8cb8"}
{"id": "train_029814", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "agent"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "641cff9996fb4ab584b8b60c7aec518e746ed5be"}
{"id": "train_029815", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a11cdfb44cafe8d50bc471f464e5ff96c13da2f0"}
{"id": "train_029816", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Bash", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "1752c7cac39e8aec861a16709141fcec4a52ddb8"}
{"id": "train_029817", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "9b4514bb2102056aa9631adb732eca1a1d80d458"}
{"id": "train_029818", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "825a2aecb2ece9de399d80ed8c7f8a3839bd66f7"}
{"id": "train_029819", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "30d7aff1ada2d5986daadd8a7e9466540904e00d"}
{"id": "train_029820", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2de24f3018c2218734bca261c99a7f5bbc3af0e3"}
{"id": "train_029821", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "011d09822b75785bcb14d1bf799398769439e6e6"}
{"id": "train_029822", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "d638f6c58d0f377d83f0183ee0d0fe81e06cb7ba"}
{"id": "train_029823", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "4fa060aa95068d2dc1f35e74ed965070ed8a8709"}
{"id": "train_029824", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c9b8325254fd79e3b0f08a0557792dc093c43cc2"}
{"id": "train_029825", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "9a7582e9cc16165c37f5150d27877d25402371d2"}
{"id": "train_029826", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "34d4b803c54903642c936ab80fe7a6b47e2b7889"}
{"id": "train_029827", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "4f2c0d13f82019ccad0c536ae5ee1cbec967cd3a"}
{"id": "train_029828", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "0c0eb42b401fa49e9de1565879837ada6e9753b2"}
{"id": "train_029829", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "106219da6843a4610641fe7dfce7d62e341e9b13"}
{"id": "train_029830", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Go", "moe_experts": ["security", "governance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "0ef683122b4d0b0f5aa3d683b3a625ac6f9e7c59"}
{"id": "train_029831", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "security"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "af0e8cf3d68bf0e4f0760e6937b27a3fa3fae4cf"}
{"id": "train_029832", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "1aab72284b4fee0e9b821117d790bd365ec0c683"}
{"id": "train_029833", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "SQL", "moe_experts": ["agent", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "4625e5d973ff07bebc1247fac32162dc3415071a"}
{"id": "train_029834", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "dfcb18e9c126f833d8ec11cdb9ccd31c979072cb"}
{"id": "train_029835", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "ddb5a4aba9740c8831b435d7313cf7904ee6ac9d"}
{"id": "train_029836", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "security"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "bdb8767fe0bd8a9a205a0ba88398b8911a404d1d"}
{"id": "train_029837", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "a91a14e4e9230cf215f240f9f48cfb4eb70574d5"}
{"id": "train_029838", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "9fa8d32500e1a8d5d12cfe21a5ece5b17d2624fa"}
{"id": "train_029839", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "03fcac1f3511a09b227c1380f54dc0876afbc0a3"}
{"id": "train_029840", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "4c9897e16869e8936d548258cce49be9df65ed5d"}
{"id": "train_029841", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "c96af1e60678347d63268bc58ef7e91a377ee111"}
{"id": "train_029842", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "64a09c8b2551d69303e4b58f6d24f52dbb7dd580"}
{"id": "train_029843", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["security", "research"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "14da01080daf8025c57a661a26da2de12d9167e5"}
{"id": "train_029844", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "8970a518a7b3c4bcbf9da7e494b2a519418c8fe5"}
{"id": "train_029845", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "dddaa4e42b74be8dd1e0c587e676b7523581182a"}
{"id": "train_029846", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "c74564c3a7dfa9ac660b5d16f3f58402cdcdf876"}
{"id": "train_029847", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "e3b1bd1eb0389b667db8fded56f25f695fe9cb87"}
{"id": "train_029848", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "bf4117d0f42a525689c1ab5208720ce255d6a4ee"}
{"id": "train_029849", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "970e20aff35e242759c287750093e25ba186c8e9"}
{"id": "train_029850", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "7cdbf4cca908e012a1682daa8fcf43f2c4ea4a80"}
{"id": "train_029851", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "bf54c5a88bda7485bf25359740107ce11d308478"}
{"id": "train_029852", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["data", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "47631f58e45a13af5951283a170d1372277e0e3e"}
{"id": "train_029853", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "fd6c55ff3664b0e9c71581cb31ea49bca90bb311"}
{"id": "train_029854", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "9b95119ef4d0c71b740e742f66aeb3638faa2766"}
{"id": "train_029855", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f6e2de2c69bc23afe47f0f66eeb60b6e38fef418"}
{"id": "train_029856", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "3779a3e3fbc27b2775fad3b2e8168aef5b6b0e4a"}
{"id": "train_029857", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "f1385edae21e937906dc71629e6056d9bcfd4e71"}
{"id": "train_029858", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["evaluation", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "aa792f413e9b9270e98d15c4aa937dd8f0f9686a"}
{"id": "train_029859", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "research"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "82b38ac1caa3e78d2b4b22fbf09298e94550973c"}
{"id": "train_029860", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "3472e89b501bcec4689d7a3a2d21d9cda562ce45"}
{"id": "train_029861", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["performance", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "1008e16d5c2a89aa6e8468cdec2f3cb888356c79"}
{"id": "train_029862", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Java", "moe_experts": ["data", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "4ed3ddd0013387166535b151a7b8df552b064d50"}
{"id": "train_029863", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "c32a5345c3b9ce5130b39e1c7e9cdc5b2b7abc48"}
{"id": "train_029864", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "data"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "dc71dea23eac815c23263652f91e5971fb26d155"}
{"id": "train_029865", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Go", "moe_experts": ["coding", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "419b9fb4e982c31b78b211f51c8b6e6397aa80c3"}
{"id": "train_029866", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "985a0a5a72d02a64f65e4c95740cb36ee68a4f67"}
{"id": "train_029867", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "62cf44faa50d5daa79fce65485039fde428d45dd"}
{"id": "train_029868", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "research"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "e697a778d8612f70d8b6bbdf6a56704dc03db0ad"}
{"id": "train_029869", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "research"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "67d1264110dc921dc18d0440ea96609c0a799acd"}
{"id": "train_029870", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "ca69fda604c618adc3b92d0b905f1d3b9ea98755"}
{"id": "train_029871", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "agent"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "694d959dabc9905fc76a5da3fdb8c7535d020911"}
{"id": "train_029872", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["performance", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "1a082629ad138bd2f4e83ed4bbb7d5878d46e938"}
{"id": "train_029873", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["coding", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9143af4c489cc91732db0178ee90d0fbde5a9b2c"}
{"id": "train_029874", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 8,\n  \"confidence\": 0.8,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "SQL", "moe_experts": ["performance", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "4eaa9d61153fd9c5af0b0d7953a2887bc747c888"}
{"id": "train_029875", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "aab8eda2415f049e65f2f08339e45cda998c35a0"}
{"id": "train_029876", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "71f361557195f1b665354c252a1824866c1527ce"}
{"id": "train_029877", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["performance", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "70e7eba5a1cba97aa63ceaadccf2fd9cfc316135"}
{"id": "train_029878", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Java", "moe_experts": ["coding", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "c6b391fe48c6a86a6405f16b1f79a1acdfcfb11f"}
{"id": "train_029879", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "SQL", "moe_experts": ["performance", "data"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "6d85899bf14a27e9cd577fa71d0f6c566033b368"}
{"id": "train_029880", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "5a2c836c28ceb066996f670f8dd76ea42459937a"}
{"id": "train_029881", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "performance"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "7e09dd612c6b14d9fd7a753b31130f410fb97605"}
{"id": "train_029882", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "ba47cf720cb790907620a804e2edde7c9d41b66c"}
{"id": "train_029883", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "a61b8f7165a1eb382fd460725a0e8efabb51966c"}
{"id": "train_029884", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["research", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2531a0dcba53d875d010ea5a11831bbf41915648"}
{"id": "train_029885", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["performance", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "d8847f03c0800f45b6b7d5eaa315540a131905a8"}
{"id": "train_029886", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "cdaed515267c02a263df7f0ecf4000c588f2fa4c"}
{"id": "train_029887", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["coding", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "b6e595c66d6b1cb61320b6f82962e86717a552fd"}
{"id": "train_029888", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "2c366f71f837efc9c05a8555c78eaea59672d69a"}
{"id": "train_029889", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "fd3ac7f3e2435937b6cd64ef1a18fa86f43cd48c"}
{"id": "train_029890", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["research", "agent"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "828a8448b4041e2f4fc7532213f90d5f9e18c585"}
{"id": "train_029891", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["governance", "performance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "1fb9b3e6c33aff7e2811bd074fd7e1ce5f415625"}
{"id": "train_029892", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Java", "moe_experts": ["evaluation", "coding"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "74c27ff88d7a394317caf6a010bc501a43081dfd"}
{"id": "train_029893", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "b01b0521b9eca147fafde1abc1c915c948a22290"}
{"id": "train_029894", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["coding", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "131930229615de2ee0070ed8528b3f5c14af9eac"}
{"id": "train_029895", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["security", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "505762040219f1659b725a342e8ce863879da074"}
{"id": "train_029896", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "f731a0bb49f2ac109dbc52e9343738bbdfa074ce"}
{"id": "train_029897", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "58f8359ff8046648706cfe0d6b05cd9d43daccec"}
{"id": "train_029898", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["agent", "coding"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "b603bb998eab2b3612c4b3d77e80753f3447c3de"}
{"id": "train_029899", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Java", "moe_experts": ["security", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "03c3e363a603106891367ff190658de95fd1ab81"}
{"id": "train_029900", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "56174617c7292e0dff3c8ac45386c1b275f3bf3a"}
{"id": "train_029901", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Self-grading and automated model critique\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "coding"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "46e99d60a6d20bd254819bb3146f43cfe6e02962"}
{"id": "train_029902", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["coding", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "4d3fb7d81b55ed68a2352a6fdc6dc0841eca3243"}
{"id": "train_029903", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: C#\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "2d2f23c35f3d9ca23a257f973d1a19638a2421e6"}
{"id": "train_029904", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Python", "moe_experts": ["data", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a8e968982663fcbf14699b954df55103d52c3f69"}
{"id": "train_029905", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "C#", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "771029c0ea23c0c15c0b15f94cf14d048ec73b40"}
{"id": "train_029906", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "governance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "b3f2270c166060f36333c70c34488665fb312e67"}
{"id": "train_029907", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9beafa1a25c8ff678b98409cca3b642f30cc12de"}
{"id": "train_029908", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["evaluation", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "916890bcb5e40f6ea5d402940ffb9f7fffe13b68"}
{"id": "train_029909", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "3ce2398c9c92b7d4aa1c431fc9a564df2d328819"}
{"id": "train_029910", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "802c77246ffcbd44c2a6fcba42e945888f589158"}
{"id": "train_029911", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "4dcba614b274ab1cb1341ba3dbfa8dfc5a13e4fd"}
{"id": "train_029912", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["governance", "security"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "5232abfb9176a2c13741df0c3efc301fafbb84fc"}
{"id": "train_029913", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["evaluation", "governance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "5dba510add539e415862793460b0b58e96c2039a"}
{"id": "train_029914", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "d9edc7d2cb3ac42ba38020647b56bac401868c5b"}
{"id": "train_029915", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "7ff2fbd7c9ddc42c68908a23d53d803be8ea3979"}
{"id": "train_029916", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "9a5a6aec6f94d868b08aafe62df2f0217eeac9db"}
{"id": "train_029917", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "tool_trace", "difficulty": "expert", "instruction": "Task: tool_trace\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "b575d5a6127a4e920e67ab611e2b2a447f53a656"}
{"id": "train_029918", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["governance", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "883ea5512d802561041338a674fa82d1ba56da22"}
{"id": "train_029919", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "0ec0f6b82af6153cc096e979b1add1f95aeaa191"}
{"id": "train_029920", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "performance"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "a9da9b27be60fd46dd5311861f0aa120a8253661"}
{"id": "train_029921", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "security"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "eaf09b41539299b2db2d8353f22670b6fd5af487"}
{"id": "train_029922", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: TypeScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "security"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "9bd62c08347bda01afa80c8505f6e841378f9979"}
{"id": "train_029923", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: C#\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["coding", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "02927fd0819147b80470640a7d6ad7ffc406dcac"}
{"id": "train_029924", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["research", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "47093086290d2312d19ce02d9ea7cff864d8fe75"}
{"id": "train_029925", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Python\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Python", "moe_experts": ["performance", "data"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "73401425426a5fdbd5828246a897a58b72daf6c5"}
{"id": "train_029926", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "performance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "0f6f260a37e67301165692a565735807eec1af8e"}
{"id": "train_029927", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "093f621efd7d3b75d573db5d6857a9cdd4bd2f0c"}
{"id": "train_029928", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Rust", "moe_experts": ["research", "evaluation"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "c77b9a8c892433d38cc767cf55add023522f3289"}
{"id": "train_029929", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "data"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "ab1e6fb13888dff826ec160e379daf0a1c2dd1d9"}
{"id": "train_029930", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["security", "evaluation"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "5a84e258797742bb2401ffccb7dcbb50862998b4"}
{"id": "train_029931", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Java\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "551c96d9229e95f25d491e6505fea62274443bad"}
{"id": "train_029932", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 10,\n  \"confidence\": 1.0,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "a64318e310c4d11f003914c694f5e75c821efd14"}
{"id": "train_029933", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: Go\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "coding"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "e10814a7fd4c34319ff4cdf1335e9f5fa7ab89f6"}
{"id": "train_029934", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: expert\nLanguage: TypeScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "c3805a3fe8d639ea8b2c377fa0842e8bb0b8a4fa"}
{"id": "train_029935", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "JavaScript", "moe_experts": ["agent", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "beac60cc201acccbe5f41fd11692dcdbccf3cdcd"}
{"id": "train_029936", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: advanced\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "df97888e042c97f257bdf3a4c935d0d5cb326036"}
{"id": "train_029937", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Rust", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "a5a70f6b43509cadff468cf8b6407258fa9f533c"}
{"id": "train_029938", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["performance", "agent"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "4a261f3840d93992880f971d100825907fd36097"}
{"id": "train_029939", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["security", "performance"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "96b65c9fb066f1eb7085c40765d0135d6690edcb"}
{"id": "train_029940", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["security", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "e8190521c7d8a7ff6c4b2303d82d6fcc8715d37f"}
{"id": "train_029941", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Rust\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Rust", "moe_experts": ["security", "research"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "defb025f431cc28790b8af28073a51dd6dcc59f4"}
{"id": "train_029942", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: JavaScript\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "961d0220e8ad017070ff1e72d7632dab3925ce68"}
{"id": "train_029943", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Rust\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["agent", "evaluation"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "344563f5f359fdcae729d2bf32c9b6365a59b8db"}
{"id": "train_029944", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: Go\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["governance", "security"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "977593847fedbc272a6ae3e6b8ca3116199047af"}
{"id": "train_029945", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["agent", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "f4be643f6ed2f66456ba0a593fef794fd3e0cda3"}
{"id": "train_029946", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Rust", "moe_experts": ["agent", "governance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "6f0c3bc1ca64289979cd2de53254946033815e37"}
{"id": "train_029947", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Multi-agent coordination and role separation\nDifficulty: expert\nLanguage: SQL\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "fb5c27eda16b2cdbc5428931c4b8865d38adef49"}
{"id": "train_029948", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "b9ffb2636bb4ca2bd880262ddc39c4be1357bea3"}
{"id": "train_029949", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["governance", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "860d90e4e27b1eafc914266c02f1718836a0eba7"}
{"id": "train_029950", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "review", "difficulty": "advanced", "instruction": "Task: review\nTopic: Governance, provenance, and audit logging\nDifficulty: advanced\nLanguage: Go\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["evaluation", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "839a60b9393fee661c6d8242b5a3108554faa0a5"}
{"id": "train_029951", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "data"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "ae5cd7b9bb398c1a59138fe7b2c5fec737b66196"}
{"id": "train_029952", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: JavaScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 9,\n  \"confidence\": 0.9,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "JavaScript", "moe_experts": ["evaluation", "research"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "23393195e24300a27a5cfd049853b06e3bf8c7bc"}
{"id": "train_029953", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: C#\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["performance", "data"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "b78bf5f1155c0def430a5c5cb8b3295a213e2027"}
{"id": "train_029954", "created": "2026-01-01", "topic": "Governance, provenance, and audit logging", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Governance, provenance, and audit logging\nDifficulty: expert\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["research", "evaluation"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "38d52395a460afea6be2d2064b28791c68259aba"}
{"id": "train_029955", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["security", "data"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e7db194e4dc84f4d2f6b85b229327ca2dbd00eb4"}
{"id": "train_029956", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["agent", "evaluation"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "dfc5c8d583a0b0cf1ebcbb40b1660b3acaf1bc3b"}
{"id": "train_029957", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "research"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "b04f31e7a8bf2950c656cf90feecb70c5068226a"}
{"id": "train_029958", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "coding"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "2313d27830bfcdecd3f855be90bff03fa62955d8"}
{"id": "train_029959", "created": "2026-01-01", "topic": "Secure code generation, red-teaming, and policy gates", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Secure code generation, red-teaming, and policy gates\nDifficulty: advanced\nLanguage: JavaScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "agent"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "0344448b8276cb63c91d6f06c16e78bc6442b2ee"}
{"id": "train_029960", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: advanced\nLanguage: Python\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["agent", "security"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "53dacb5f2152e854f5ecfe66596cc8ab3ddda375"}
{"id": "train_029961", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: advanced\nLanguage: Bash\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["evaluation", "coding"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "9a914fb6862dedadc31ba4b4470126068c2fab97"}
{"id": "train_029962", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: advanced\nLanguage: TypeScript\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "TypeScript", "moe_experts": ["performance", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "53f0910c2f1c2359240f3fe1e3d1786c8c5a1656"}
{"id": "train_029963", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "tool_trace", "difficulty": "advanced", "instruction": "Task: tool_trace\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n[\n  {\n    \"tool\": \"repo_checkout\",\n    \"args\": {\n      \"commit\": \"abc123\"\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"unit\"\n    }\n  },\n  {\n    \"tool\": \"apply_patch\",\n    \"args\": {\n      \"files\": 1\n    }\n  },\n  {\n    \"tool\": \"run_tests\",\n    \"args\": {\n      \"suite\": \"full\"\n    }\n  }\n]", "metadata": {"language": "Go", "moe_experts": ["evaluation", "governance"], "cost_budget": 4, "latency_ms_target": 200, "audit_required": true}, "hash": "d620483ef593c9b2502b358a1f3433e0d12027d9"}
{"id": "train_029964", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["coding", "security"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "62cf44faa50d5daa79fce65485039fde428d45dd"}
{"id": "train_029965", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "review", "difficulty": "expert", "instruction": "Task: review\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["coding", "research"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "8685b31a4a9fa114200de1a89276ecff9be844a7"}
{"id": "train_029966", "created": "2026-01-01", "topic": "Extended context, memory, and repo graphs", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Extended context, memory, and repo graphs\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "a73359137c890c84dff29765906a3eb96f4d96e6"}
{"id": "train_029967", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: advanced\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "C#", "moe_experts": ["data", "agent"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "dce93b042cfa2518401973623e98b1080781fc8b"}
{"id": "train_029968", "created": "2026-01-01", "topic": "Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Agentic coding systems (plan\u2192edit\u2192test\u2192reflect)\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["performance", "governance"], "cost_budget": 5, "latency_ms_target": 100, "audit_required": true}, "hash": "eefd915c00b15604ab030238ff11a87bd12c6731"}
{"id": "train_029969", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Failure engineering and resilience testing\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Bash", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "d992fbc9c45d6e168e8b57b53d6c3d57f13bdfb3"}
{"id": "train_029970", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: SQL\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["evaluation", "agent"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "9146ed390417648c0f2e710ac55253e1c6466dc8"}
{"id": "train_029971", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "compare", "difficulty": "advanced", "instruction": "Task: compare\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Java\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["security", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "e86fff6cd171118ee3339841cd10a9b2f748cc7f"}
{"id": "train_029972", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "code", "difficulty": "advanced", "instruction": "Task: code\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["governance", "evaluation"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "5e52c86c498077ead358f7b78263876381630690"}
{"id": "train_029973", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Bash", "moe_experts": ["governance", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "2c1db5c82fd07554fdee071d98079c83fc844451"}
{"id": "train_029974", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Rust\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "performance"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "0073d9f517528c565ec725ac49f21a6e11dd687d"}
{"id": "train_029975", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "eval", "difficulty": "advanced", "instruction": "Task: eval\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "Python", "moe_experts": ["governance", "coding"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "594b2e35fbc49b6b223ed1807cd0c8f35513b64a"}
{"id": "train_029976", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "code", "difficulty": "expert", "instruction": "Task: code\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["security", "performance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "a4cbc0060a589adaec3c5201c3377990ba2dfbaf"}
{"id": "train_029977", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "data_pipeline", "difficulty": "expert", "instruction": "Task: data_pipeline\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Bash\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["research", "governance"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "9f7c0ef414418fd0156f3c0dab53d0acf7d893e9"}
{"id": "train_029978", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "design", "difficulty": "advanced", "instruction": "Task: design\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["governance", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "33c5b6dd3286dc7838e74bc2c04ac7bd6fbccf31"}
{"id": "train_029979", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "explain", "difficulty": "advanced", "instruction": "Task: explain\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: advanced\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Bash", "moe_experts": ["data", "governance"], "cost_budget": 5, "latency_ms_target": 50, "audit_required": true}, "hash": "bbc971daf3983d139ed50a1c168043ff5828d0c8"}
{"id": "train_029980", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Go\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "Go", "moe_experts": ["evaluation", "agent"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "8b7f6f40cf6dc1855ffb97d291996d48dfb06e71"}
{"id": "train_029981", "created": "2026-01-01", "topic": "Code-specialized model families and scaling laws", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Code-specialized model families and scaling laws\nDifficulty: advanced\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "C#", "moe_experts": ["data", "performance"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "cfc5edd04e885958fdfeacf1e8929e2fe5a8722b"}
{"id": "train_029982", "created": "2026-01-01", "topic": "Failure engineering and resilience testing", "task_type": "self_grade", "difficulty": "advanced", "instruction": "Task: self_grade\nTopic: Failure engineering and resilience testing\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 7,\n  \"confidence\": 0.7,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Python", "moe_experts": ["agent", "performance"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "0e078921e1e7c6bdcc5c3dd0ec49901580805a12"}
{"id": "train_029983", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Java\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["coding", "security"], "cost_budget": 2, "latency_ms_target": 50, "audit_required": true}, "hash": "2a8cbff1c7a936ee609f5c05687dfe482914f654"}
{"id": "train_029984", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "compare", "difficulty": "expert", "instruction": "Task: compare\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: C#\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "C#", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "dcd12e1af010989330b632ac295aae5aaf7a7e6e"}
{"id": "train_029985", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "patch_diff", "difficulty": "advanced", "instruction": "Task: patch_diff\nTopic: Model merging, distillation, and continual learning\nDifficulty: advanced\nLanguage: JavaScript\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "JavaScript", "moe_experts": ["coding", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "fdb9bc84cd73785e32952b514cd554740abc9cc2"}
{"id": "train_029986", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "advanced", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: advanced\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Python", "moe_experts": ["research", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "de12ea4bba40c465b98612d1b02086dbe8613a22"}
{"id": "train_029987", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "patch_diff", "difficulty": "expert", "instruction": "Task: patch_diff\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: SQL\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nPatch:\n```diff\n- result = a / b\n+ if b == 0:\n+     raise ValueError('division by zero')\n+ result = a / b\n\n```", "metadata": {"language": "SQL", "moe_experts": ["agent", "coding"], "cost_budget": 3, "latency_ms_target": 200, "audit_required": true}, "hash": "cfacb4a57288b93b97502ff2c329bdbcd676a420"}
{"id": "train_029988", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Python\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["security", "governance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "67bfeb5bf076ca579530245505e80a1d3469ef92"}
{"id": "train_029989", "created": "2026-01-01", "topic": "Self-grading and automated model critique", "task_type": "data_pipeline", "difficulty": "advanced", "instruction": "Task: data_pipeline\nTopic: Self-grading and automated model critique\nDifficulty: advanced\nLanguage: Go\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Go", "moe_experts": ["performance", "governance"], "cost_budget": 2, "latency_ms_target": 100, "audit_required": true}, "hash": "a5628cd15897b7428a3f07c46de3a837efc1723e"}
{"id": "train_029990", "created": "2026-01-01", "topic": "Tool calling protocols and sandboxed execution", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Tool calling protocols and sandboxed execution\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "TypeScript", "moe_experts": ["data", "agent"], "cost_budget": 3, "latency_ms_target": 50, "audit_required": true}, "hash": "a28f5bbdec447fcab623a2349ebe63a524f0551c"}
{"id": "train_029991", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Python\nContext: Large monorepo with flaky tests and strict CI\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Python", "moe_experts": ["research", "coding"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "c293bd74e67988a00552c4c69b958fae6c283545"}
{"id": "train_029992", "created": "2026-01-01", "topic": "SWE-bench real-repo evaluation and leakage control", "task_type": "self_grade", "difficulty": "expert", "instruction": "Task: self_grade\nTopic: SWE-bench real-repo evaluation and leakage control\nDifficulty: expert\nLanguage: Rust\nContext: Offline air-gapped deployment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\n{\n  \"score\": 6,\n  \"confidence\": 0.6,\n  \"notes\": \"Passed tests; minimal diff; no security flags.\"\n}", "metadata": {"language": "Rust", "moe_experts": ["agent", "data"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "c84e420c2afc3e010834f9c3885d28f1d46a6482"}
{"id": "train_029993", "created": "2026-01-01", "topic": "Latency, cost, throughput optimization", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Latency, cost, throughput optimization\nDifficulty: expert\nLanguage: TypeScript\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "TypeScript", "moe_experts": ["agent", "research"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "7e2872d5b839ddffc00054530b1f122d9bd7b646"}
{"id": "train_029994", "created": "2026-01-01", "topic": "Mixture-of-Experts (MoE) routing and sparsity", "task_type": "failure_analysis", "difficulty": "expert", "instruction": "Task: failure_analysis\nTopic: Mixture-of-Experts (MoE) routing and sparsity\nDifficulty: expert\nLanguage: Bash\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nFailure: regression detected\nCause: missing edge case\nFix: add guard + test\n", "metadata": {"language": "Bash", "moe_experts": ["security", "research"], "cost_budget": 5, "latency_ms_target": 200, "audit_required": true}, "hash": "698804ee6110287cfdff4bee04787388b80ad193"}
{"id": "train_029995", "created": "2026-01-01", "topic": "Dataset curation pipelines and quality attribution", "task_type": "agent_loop", "difficulty": "expert", "instruction": "Task: agent_loop\nTopic: Dataset curation pipelines and quality attribution\nDifficulty: expert\nLanguage: Java\nContext: Research benchmark validation environment\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Java", "moe_experts": ["evaluation", "performance"], "cost_budget": 3, "latency_ms_target": 100, "audit_required": true}, "hash": "0925d32df13c8139447fbd569dc6c7cbe110db12"}
{"id": "train_029996", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "design", "difficulty": "expert", "instruction": "Task: design\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: Rust\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "Rust", "moe_experts": ["research", "agent"], "cost_budget": 2, "latency_ms_target": 200, "audit_required": true}, "hash": "ad1382a808453b3be1e4370e586e345f5e6b4860"}
{"id": "train_029997", "created": "2026-01-01", "topic": "Reasoning-first coding models and tunable deliberation", "task_type": "explain", "difficulty": "expert", "instruction": "Task: explain\nTopic: Reasoning-first coding models and tunable deliberation\nDifficulty: expert\nLanguage: SQL\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["data", "coding"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "e5a956497da098b4b4d12d1181901c56e7ce31d8"}
{"id": "train_029998", "created": "2026-01-01", "topic": "Model merging, distillation, and continual learning", "task_type": "eval", "difficulty": "expert", "instruction": "Task: eval\nTopic: Model merging, distillation, and continual learning\nDifficulty: expert\nLanguage: C#\nContext: Regulated enterprise with audit requirements\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nMetrics:\n- pass@k\n- time-to-green\n- cost-per-fix\n- regression rate\n", "metadata": {"language": "C#", "moe_experts": ["evaluation", "security"], "cost_budget": 4, "latency_ms_target": 50, "audit_required": true}, "hash": "6be3475790b646c71d3032d613266c56289d9aca"}
{"id": "train_029999", "created": "2026-01-01", "topic": "Multi-agent coordination and role separation", "task_type": "agent_loop", "difficulty": "advanced", "instruction": "Task: agent_loop\nTopic: Multi-agent coordination and role separation\nDifficulty: advanced\nLanguage: SQL\nContext: Latency-critical production API\nProduce frontier-grade, production-valid output.", "input": "", "output": "Facts:\n- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.\n- Agentic and multi-agent systems outperform single-pass generation.\n\nExpert guidance with risks, metrics, and governance controls.", "metadata": {"language": "SQL", "moe_experts": ["security", "data"], "cost_budget": 4, "latency_ms_target": 100, "audit_required": true}, "hash": "beff04cdf4289880e8eeed251af4868a5d9b74f7"}