validation.jsonl

{"id": "JT3K-002245", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "easy", "title": "Handle empty payloads", "domain": "healthtech", "user_story": "A webhook handler crashes when the JSON body is empty. Make it resilient and return a proper error.", "repo_context": "Repo: analytics-core. Domain: healthtech. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Handle empty payloads. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-001677", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "python", "difficulty": "easy", "title": "Map the flow of a request through the service", "domain": "developer-tools", "user_story": "Given a Python Flask service with routes, services, and a DB layer, explain the request path for creating an order.", "repo_context": "Repo: dashboard-web. Domain: developer-tools. Python service with app/, services/, repositories/, tests/.", "instruction": "Explain how this codebase handles the task: Map the flow of a request through the service. Identify the files and modules most likely involved.", "reference_solution": "1. Start at the route/controller.\n2. Follow service layer calls.\n3. Check repository and schema modules.\n4. Verify tests that cover the flow.", "evaluation": {"checks": ["correct entry point", "correct data flow", "relevant files identified"]}, "tags": ["agentic", "coding", "repo_understanding", "python", "developer-tools", "persian-friendly"]}
{"id": "JT3K-001257", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "javascript", "difficulty": "medium", "title": "Extract reusable hook", "domain": "fintech", "user_story": "Refactor repeated fetch logic into a custom React hook.", "repo_context": "Repo: ml-pipeline. Domain: fintech. React components and shared logic in a frontend repo.", "instruction": "Refactor the following: Extract reusable hook. Reduce duplication and keep the public interface stable.", "reference_solution": "Extract a hook or utility, preserve props and return shape, and add tests around behavior.", "evaluation": {"checks": ["duplication removed", "interface stable", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "javascript", "fintech", "persian-friendly"]}
{"id": "JT3K-000790", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "typescript", "difficulty": "medium", "title": "Test a formatter", "domain": "edtech", "user_story": "Write tests for a date formatter that must respect locale and fallback rules.", "repo_context": "Repo: cms-platform. Domain: edtech. JS testing stack with Vitest/Jest and React Testing Library.", "instruction": "Write test cases for: Test a formatter. Start with the edge cases and expected behavior.", "reference_solution": "Cover input validation, boundary values, async states, and error rendering.", "evaluation": {"checks": ["edge coverage", "async behavior", "clear expectations"]}, "tags": ["agentic", "coding", "tdd", "typescript", "edtech", "persian-friendly"]}
{"id": "JT3K-001887", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "javascript", "difficulty": "easy", "title": "Plan verify-edit cycle", "domain": "developer-tools", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: ecommerce-backend. Domain: developer-tools. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "javascript", "developer-tools", "persian-friendly"]}
{"id": "JT3K-000517", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "javascript", "difficulty": "medium", "title": "Simplify reducer logic", "domain": "creator-tools", "user_story": "Reduce branching in a Redux reducer without changing behavior.", "repo_context": "Repo: dashboard-web. Domain: creator-tools. React components and shared logic in a frontend repo.", "instruction": "Refactor the following: Simplify reducer logic. Reduce duplication and keep the public interface stable.", "reference_solution": "Extract a hook or utility, preserve props and return shape, and add tests around behavior.", "evaluation": {"checks": ["duplication removed", "interface stable", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "javascript", "creator-tools", "persian-friendly"]}
{"id": "JT3K-000317", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "javascript", "difficulty": "hard", "title": "Plan verify-edit cycle", "domain": "edtech", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: ml-pipeline. Domain: edtech. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "javascript", "edtech", "persian-friendly"]}
{"id": "JT3K-000996", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "javascript", "difficulty": "medium", "title": "Locate data fetching boundaries", "domain": "saas", "user_story": "A Next.js repo separates server actions, client components, and shared utilities.", "repo_context": "Repo: chat-agent. Domain: saas. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Locate data fetching boundaries. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "javascript", "saas", "persian-friendly"]}
{"id": "JT3K-001170", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "python", "difficulty": "medium", "title": "Sequence tool calls", "domain": "healthtech", "user_story": "Construct a tool-use plan for fixing an import error across multiple files.", "repo_context": "Repo: payments-api. Domain: healthtech. Agentic coding workflow with file reads, tests, and patches.", "instruction": "Create a tool-use plan for: Sequence tool calls. Include the ideal sequence of read, edit, test, and verify actions.", "reference_solution": "Read the relevant file set, inspect failure output, patch the narrowest surface, rerun targeted tests, then broaden checks.", "evaluation": {"checks": ["tool order sensible", "minimal edit strategy", "verification included"]}, "tags": ["agentic", "coding", "tool_use", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-000560", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "typescript", "difficulty": "medium", "title": "Fix stale closure", "domain": "developer-tools", "user_story": "A hook callback reads old state after rapid updates; patch it.", "repo_context": "Repo: ecommerce-backend. Domain: developer-tools. React/Next.js application with hooks and server/client boundaries.", "instruction": "Fix this bug: Fix stale closure. Keep the change minimal and safe.", "reference_solution": "Guard missing values, avoid stale state, and keep server/client-only logic in the right layer.", "evaluation": {"checks": ["bug fixed", "safe boundary handling", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "typescript", "developer-tools", "persian-friendly"]}
{"id": "JT3K-002244", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "javascript", "difficulty": "medium", "title": "Design client methods", "domain": "fintech", "user_story": "Design a TypeScript client for search, pagination, and retries.", "repo_context": "Repo: dashboard-web. Domain: fintech. TypeScript client or Node API layer.", "instruction": "Design a TypeScript API for: Design client methods. Define method names, inputs, and return types.", "reference_solution": "Prefer explicit types, typed errors, pagination support, and composable options objects.", "evaluation": {"checks": ["typed contract", "clear ergonomics", "extensibility"]}, "tags": ["agentic", "coding", "api_design", "javascript", "fintech", "persian-friendly"]}
{"id": "JT3K-000497", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "bash", "difficulty": "medium", "title": "Simplify script logic", "domain": "fintech", "user_story": "Refactor a long shell script into functions with clearer error handling.", "repo_context": "Repo: payments-api. Domain: fintech. Bash automation with functions and environment variables.", "instruction": "Refactor the script for: Simplify script logic. Reduce repetition and improve error handling.", "reference_solution": "Extract functions, use strict mode carefully, and keep reusable command pieces in variables.", "evaluation": {"checks": ["repetition reduced", "error handling", "maintainability"]}, "tags": ["agentic", "coding", "refactor", "bash", "fintech", "persian-friendly"]}
{"id": "JT3K-001874", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "bash", "difficulty": "medium", "title": "Plan release verification", "domain": "saas", "user_story": "Run lint, unit tests, packaging, and smoke checks before publishing.", "repo_context": "Repo: cms-platform. Domain: saas. Release and verification scripts in bash.", "instruction": "Build a tool-use plan for: Plan release verification. Include commands to inspect, run, and validate.", "reference_solution": "List the exact commands in order: inspect files, run targeted checks, then package and smoke test.", "evaluation": {"checks": ["command order", "verification", "safe execution"]}, "tags": ["agentic", "coding", "tool_use", "bash", "saas", "persian-friendly"]}
{"id": "JT3K-001731", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "python", "difficulty": "easy", "title": "Trace a failing unit test", "domain": "infra", "user_story": "A test intermittently fails due to shared mutable state between cases.", "repo_context": "Repo: ecommerce-backend. Domain: infra. Python service with logs, stack traces, and tests.", "instruction": "Diagnose this failure: Trace a failing unit test. Provide the most likely cause and the verification steps.", "reference_solution": "Read the stack trace, identify the failing boundary, isolate shared state or missing key paths, then add a regression test.", "evaluation": {"checks": ["root cause", "debug steps", "verification plan"]}, "tags": ["agentic", "coding", "debug_trace", "python", "infra", "persian-friendly"]}
{"id": "JT3K-000455", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "hard", "title": "Fix timezone parsing", "domain": "fintech", "user_story": "A datetime parser returns naive objects and breaks scheduling. Patch it to handle UTC offsets safely.", "repo_context": "Repo: chat-agent. Domain: fintech. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Fix timezone parsing. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "fintech", "persian-friendly"]}
{"id": "JT3K-000248", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "typescript", "difficulty": "easy", "title": "Design client methods", "domain": "saas", "user_story": "Design a TypeScript client for search, pagination, and retries.", "repo_context": "Repo: analytics-core. Domain: saas. TypeScript client or Node API layer.", "instruction": "Design a TypeScript API for: Design client methods. Define method names, inputs, and return types.", "reference_solution": "Prefer explicit types, typed errors, pagination support, and composable options objects.", "evaluation": {"checks": ["typed contract", "clear ergonomics", "extensibility"]}, "tags": ["agentic", "coding", "api_design", "typescript", "saas", "persian-friendly"]}
{"id": "JT3K-000395", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "python", "difficulty": "medium", "title": "Reduce duplication", "domain": "ecommerce", "user_story": "The repo has repeated validation logic in three endpoints; centralize it.", "repo_context": "Repo: payments-api. Domain: ecommerce. Python codebase with repeated helpers and service objects.", "instruction": "Refactor the following situation: Reduce duplication. Preserve behavior and reduce complexity.", "reference_solution": "Extract helpers, keep public API stable, add/adjust tests, and avoid mixing business logic with I/O.", "evaluation": {"checks": ["behavior preserved", "complexity reduced", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "python", "ecommerce", "persian-friendly"]}
{"id": "JT3K-001344", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "javascript", "difficulty": "easy", "title": "Design client methods", "domain": "infra", "user_story": "Design a TypeScript client for search, pagination, and retries.", "repo_context": "Repo: ml-pipeline. Domain: infra. TypeScript client or Node API layer.", "instruction": "Design a TypeScript API for: Design client methods. Define method names, inputs, and return types.", "reference_solution": "Prefer explicit types, typed errors, pagination support, and composable options objects.", "evaluation": {"checks": ["typed contract", "clear ergonomics", "extensibility"]}, "tags": ["agentic", "coding", "api_design", "javascript", "infra", "persian-friendly"]}
{"id": "JT3K-001544", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "javascript", "difficulty": "medium", "title": "Plan verify-edit cycle", "domain": "developer-tools", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: auth-service. Domain: developer-tools. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "javascript", "developer-tools", "persian-friendly"]}
{"id": "JT3K-000838", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "javascript", "difficulty": "easy", "title": "Review an API client", "domain": "fintech", "user_story": "The client mixes transport errors and business errors in one catch block.", "repo_context": "Repo: dashboard-web. Domain: fintech. Frontend app with shared utilities and API wrappers.", "instruction": "Review the implementation for: Review an API client. Comment on correctness, UX, and maintainability.", "reference_solution": "Look for async race conditions, unnecessary re-renders, error handling gaps, and accessibility issues.", "evaluation": {"checks": ["quality issues found", "practical suggestions", "frontend concerns"]}, "tags": ["agentic", "coding", "code_review", "javascript", "fintech", "persian-friendly"]}
{"id": "JT3K-001534", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "typescript", "difficulty": "medium", "title": "Guard null props", "domain": "edtech", "user_story": "A component crashes when optional props are missing.", "repo_context": "Repo: chat-agent. Domain: edtech. React/Next.js application with hooks and server/client boundaries.", "instruction": "Fix this bug: Guard null props. Keep the change minimal and safe.", "reference_solution": "Guard missing values, avoid stale state, and keep server/client-only logic in the right layer.", "evaluation": {"checks": ["bug fixed", "safe boundary handling", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "typescript", "edtech", "persian-friendly"]}
{"id": "JT3K-002496", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "javascript", "difficulty": "medium", "title": "Trace state across components", "domain": "creator-tools", "user_story": "A React app manages auth state in context, hooks, and a small API layer.", "repo_context": "Repo: dashboard-web. Domain: creator-tools. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Trace state across components. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "javascript", "creator-tools", "persian-friendly"]}
{"id": "JT3K-002651", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "bash", "difficulty": "easy", "title": "Diagnose build failure", "domain": "creator-tools", "user_story": "A CI script fails because one command exits early and hides the root cause.", "repo_context": "Repo: scheduler. Domain: creator-tools. CI scripts, release automation, and shell helpers.", "instruction": "Diagnose this shell failure: Diagnose build failure. Explain the failure point and the safest fix.", "reference_solution": "Check exit codes, quote variables, and make failure handling explicit with set -euo pipefail when appropriate.", "evaluation": {"checks": ["failure point", "shell safety", "clear fix"]}, "tags": ["agentic", "coding", "debug_trace", "bash", "creator-tools", "persian-friendly"]}
{"id": "JT3K-001625", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "sql", "difficulty": "easy", "title": "Plan migration checks", "domain": "saas", "user_story": "Inspect tables, compare counts, run a diff query, and validate migration impact.", "repo_context": "Repo: auth-service. Domain: saas. Agent workflow over SQL files and validation queries.", "instruction": "Create a tool-use plan for: Plan migration checks. Read schema, inspect query, run checks, and validate output.", "reference_solution": "Inspect schema first, isolate the grain, run a comparison query, then verify counts after the patch.", "evaluation": {"checks": ["schema first", "comparison query", "verification"]}, "tags": ["agentic", "coding", "tool_use", "sql", "saas", "persian-friendly"]}
{"id": "JT3K-000876", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "sql", "difficulty": "easy", "title": "Plan migration checks", "domain": "fintech", "user_story": "Inspect tables, compare counts, run a diff query, and validate migration impact.", "repo_context": "Repo: scheduler. Domain: fintech. Agent workflow over SQL files and validation queries.", "instruction": "Create a tool-use plan for: Plan migration checks. Read schema, inspect query, run checks, and validate output.", "reference_solution": "Inspect schema first, isolate the grain, run a comparison query, then verify counts after the patch.", "evaluation": {"checks": ["schema first", "comparison query", "verification"]}, "tags": ["agentic", "coding", "tool_use", "sql", "fintech", "persian-friendly"]}
{"id": "JT3K-001334", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "typescript", "difficulty": "easy", "title": "Diagnose promise flow", "domain": "infra", "user_story": "A promise chain swallows the original error and returns a misleading success.", "repo_context": "Repo: mobile-sdk. Domain: infra. Frontend app with hydration and async data flow.", "instruction": "Diagnose the issue: Diagnose promise flow. Explain likely cause and how to verify the fix.", "reference_solution": "Check server/client output mismatches, stale closures, promise handling, and state synchronization.", "evaluation": {"checks": ["root cause", "verification steps", "frontend-specific diagnosis"]}, "tags": ["agentic", "coding", "debug_trace", "typescript", "infra", "persian-friendly"]}
{"id": "JT3K-000355", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "python", "difficulty": "easy", "title": "Review a cache helper", "domain": "healthtech", "user_story": "A cache wrapper forgets to namespace keys and can collide across tenants.", "repo_context": "Repo: analytics-core. Domain: healthtech. Python backend with API handlers and background workers.", "instruction": "Review the code for: Review a cache helper. Call out risks and give a practical improvement plan.", "reference_solution": "Flag correctness, performance, maintainability, and security concerns. Recommend the smallest safe refactor.", "evaluation": {"checks": ["risk identified", "actionable fix", "production awareness"]}, "tags": ["agentic", "coding", "code_review", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-001048", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "typescript", "difficulty": "easy", "title": "Plan verify-edit cycle", "domain": "infra", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: scheduler. Domain: infra. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "typescript", "infra", "persian-friendly"]}
{"id": "JT3K-002934", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "typescript", "difficulty": "easy", "title": "Trace state across components", "domain": "infra", "user_story": "A React app manages auth state in context, hooks, and a small API layer.", "repo_context": "Repo: payments-api. Domain: infra. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Trace state across components. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "typescript", "infra", "persian-friendly"]}
{"id": "JT3K-002603", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "bash", "difficulty": "medium", "title": "Simplify script logic", "domain": "healthtech", "user_story": "Refactor a long shell script into functions with clearer error handling.", "repo_context": "Repo: chat-agent. Domain: healthtech. Bash automation with functions and environment variables.", "instruction": "Refactor the script for: Simplify script logic. Reduce repetition and improve error handling.", "reference_solution": "Extract functions, use strict mode carefully, and keep reusable command pieces in variables.", "evaluation": {"checks": ["repetition reduced", "error handling", "maintainability"]}, "tags": ["agentic", "coding", "refactor", "bash", "healthtech", "persian-friendly"]}
{"id": "JT3K-001806", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "javascript", "difficulty": "easy", "title": "Plan verify-edit cycle", "domain": "ecommerce", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: analytics-core. Domain: ecommerce. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "javascript", "ecommerce", "persian-friendly"]}
{"id": "JT3K-000380", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "sql", "difficulty": "hard", "title": "Fix duplicate aggregation", "domain": "developer-tools", "user_story": "A query double-counts rows after joining two one-to-many tables.", "repo_context": "Repo: mobile-sdk. Domain: developer-tools. SQL analytics pipeline with reporting queries.", "instruction": "Fix the query issue: Fix duplicate aggregation. Explain the bug and provide the corrected pattern.", "reference_solution": "Deduplicate before aggregating, use correct grain, and be careful with time boundaries.", "evaluation": {"checks": ["correct grain", "no double count", "boundary handling"]}, "tags": ["agentic", "coding", "bug_fix", "sql", "developer-tools", "persian-friendly"]}
{"id": "JT3K-001648", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "typescript", "difficulty": "medium", "title": "Extract reusable hook", "domain": "saas", "user_story": "Refactor repeated fetch logic into a custom React hook.", "repo_context": "Repo: dashboard-web. Domain: saas. React components and shared logic in a frontend repo.", "instruction": "Refactor the following: Extract reusable hook. Reduce duplication and keep the public interface stable.", "reference_solution": "Extract a hook or utility, preserve props and return shape, and add tests around behavior.", "evaluation": {"checks": ["duplication removed", "interface stable", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "typescript", "saas", "persian-friendly"]}
{"id": "JT3K-001943", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "javascript", "difficulty": "hard", "title": "Extract reusable hook", "domain": "fintech", "user_story": "Refactor repeated fetch logic into a custom React hook.", "repo_context": "Repo: dashboard-web. Domain: fintech. React components and shared logic in a frontend repo.", "instruction": "Refactor the following: Extract reusable hook. Reduce duplication and keep the public interface stable.", "reference_solution": "Extract a hook or utility, preserve props and return shape, and add tests around behavior.", "evaluation": {"checks": ["duplication removed", "interface stable", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "javascript", "fintech", "persian-friendly"]}
{"id": "JT3K-002103", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "bash", "difficulty": "hard", "title": "Simplify script logic", "domain": "creator-tools", "user_story": "Refactor a long shell script into functions with clearer error handling.", "repo_context": "Repo: mobile-sdk. Domain: creator-tools. Bash automation with functions and environment variables.", "instruction": "Refactor the script for: Simplify script logic. Reduce repetition and improve error handling.", "reference_solution": "Extract functions, use strict mode carefully, and keep reusable command pieces in variables.", "evaluation": {"checks": ["repetition reduced", "error handling", "maintainability"]}, "tags": ["agentic", "coding", "refactor", "bash", "creator-tools", "persian-friendly"]}
{"id": "JT3K-001330", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "python", "difficulty": "easy", "title": "Explain a stack trace", "domain": "healthtech", "user_story": "A KeyError appears in production when a field is absent in a nested payload.", "repo_context": "Repo: analytics-core. Domain: healthtech. Python service with logs, stack traces, and tests.", "instruction": "Diagnose this failure: Explain a stack trace. Provide the most likely cause and the verification steps.", "reference_solution": "Read the stack trace, identify the failing boundary, isolate shared state or missing key paths, then add a regression test.", "evaluation": {"checks": ["root cause", "debug steps", "verification plan"]}, "tags": ["agentic", "coding", "debug_trace", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-000278", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "typescript", "difficulty": "medium", "title": "Review an API client", "domain": "creator-tools", "user_story": "The client mixes transport errors and business errors in one catch block.", "repo_context": "Repo: chat-agent. Domain: creator-tools. Frontend app with shared utilities and API wrappers.", "instruction": "Review the implementation for: Review an API client. Comment on correctness, UX, and maintainability.", "reference_solution": "Look for async race conditions, unnecessary re-renders, error handling gaps, and accessibility issues.", "evaluation": {"checks": ["quality issues found", "practical suggestions", "frontend concerns"]}, "tags": ["agentic", "coding", "code_review", "typescript", "creator-tools", "persian-friendly"]}
{"id": "JT3K-001039", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "sql", "difficulty": "easy", "title": "Review a reporting query", "domain": "developer-tools", "user_story": "The query is correct but slow due to an unbounded scan and unnecessary subquery.", "repo_context": "Repo: scheduler. Domain: developer-tools. SQL reporting and ETL jobs.", "instruction": "Review the SQL for: Review a reporting query. Identify correctness and performance risks.", "reference_solution": "Flag full scans, unnecessary CTEs, non-sargable filters, and inconsistent joins.", "evaluation": {"checks": ["performance risk", "logic risk", "clear recommendation"]}, "tags": ["agentic", "coding", "code_review", "sql", "developer-tools", "persian-friendly"]}
{"id": "JT3K-000747", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "typescript", "difficulty": "medium", "title": "Plan verify-edit cycle", "domain": "healthtech", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: analytics-core. Domain: healthtech. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "typescript", "healthtech", "persian-friendly"]}
{"id": "JT3K-000620", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "javascript", "difficulty": "hard", "title": "Test a formatter", "domain": "healthtech", "user_story": "Write tests for a date formatter that must respect locale and fallback rules.", "repo_context": "Repo: dashboard-web. Domain: healthtech. JS testing stack with Vitest/Jest and React Testing Library.", "instruction": "Write test cases for: Test a formatter. Start with the edge cases and expected behavior.", "reference_solution": "Cover input validation, boundary values, async states, and error rendering.", "evaluation": {"checks": ["edge coverage", "async behavior", "clear expectations"]}, "tags": ["agentic", "coding", "tdd", "javascript", "healthtech", "persian-friendly"]}
{"id": "JT3K-002273", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "javascript", "difficulty": "medium", "title": "Plan verify-edit cycle", "domain": "fintech", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: payments-api. Domain: fintech. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "javascript", "fintech", "persian-friendly"]}
{"id": "JT3K-002257", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "javascript", "difficulty": "hard", "title": "Plan verify-edit cycle", "domain": "edtech", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: ml-pipeline. Domain: edtech. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "javascript", "edtech", "persian-friendly"]}
{"id": "JT3K-002333", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "javascript", "difficulty": "easy", "title": "Guard null props", "domain": "healthtech", "user_story": "A component crashes when optional props are missing.", "repo_context": "Repo: dashboard-web. Domain: healthtech. React/Next.js application with hooks and server/client boundaries.", "instruction": "Fix this bug: Guard null props. Keep the change minimal and safe.", "reference_solution": "Guard missing values, avoid stale state, and keep server/client-only logic in the right layer.", "evaluation": {"checks": ["bug fixed", "safe boundary handling", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "javascript", "healthtech", "persian-friendly"]}
{"id": "JT3K-001873", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "javascript", "difficulty": "easy", "title": "Analyze hydration mismatch", "domain": "fintech", "user_story": "A server-rendered page renders different timestamps on client and server.", "repo_context": "Repo: dashboard-web. Domain: fintech. Frontend app with hydration and async data flow.", "instruction": "Diagnose the issue: Analyze hydration mismatch. Explain likely cause and how to verify the fix.", "reference_solution": "Check server/client output mismatches, stale closures, promise handling, and state synchronization.", "evaluation": {"checks": ["root cause", "verification steps", "frontend-specific diagnosis"]}, "tags": ["agentic", "coding", "debug_trace", "javascript", "fintech", "persian-friendly"]}
{"id": "JT3K-002090", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "typescript", "difficulty": "easy", "title": "Plan verify-edit cycle", "domain": "developer-tools", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: dashboard-web. Domain: developer-tools. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "typescript", "developer-tools", "persian-friendly"]}
{"id": "JT3K-000639", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "sql", "difficulty": "medium", "title": "Plan migration checks", "domain": "infra", "user_story": "Inspect tables, compare counts, run a diff query, and validate migration impact.", "repo_context": "Repo: auth-service. Domain: infra. Agent workflow over SQL files and validation queries.", "instruction": "Create a tool-use plan for: Plan migration checks. Read schema, inspect query, run checks, and validate output.", "reference_solution": "Inspect schema first, isolate the grain, run a comparison query, then verify counts after the patch.", "evaluation": {"checks": ["schema first", "comparison query", "verification"]}, "tags": ["agentic", "coding", "tool_use", "sql", "infra", "persian-friendly"]}
{"id": "JT3K-002309", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "medium", "title": "Handle empty payloads", "domain": "creator-tools", "user_story": "A webhook handler crashes when the JSON body is empty. Make it resilient and return a proper error.", "repo_context": "Repo: auth-service. Domain: creator-tools. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Handle empty payloads. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "creator-tools", "persian-friendly"]}
{"id": "JT3K-001472", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "bash", "difficulty": "medium", "title": "Plan release verification", "domain": "infra", "user_story": "Run lint, unit tests, packaging, and smoke checks before publishing.", "repo_context": "Repo: analytics-core. Domain: infra. Release and verification scripts in bash.", "instruction": "Build a tool-use plan for: Plan release verification. Include commands to inspect, run, and validate.", "reference_solution": "List the exact commands in order: inspect files, run targeted checks, then package and smoke test.", "evaluation": {"checks": ["command order", "verification", "safe execution"]}, "tags": ["agentic", "coding", "tool_use", "bash", "infra", "persian-friendly"]}
{"id": "JT3K-002920", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "javascript", "difficulty": "medium", "title": "Test a formatter", "domain": "saas", "user_story": "Write tests for a date formatter that must respect locale and fallback rules.", "repo_context": "Repo: payments-api. Domain: saas. JS testing stack with Vitest/Jest and React Testing Library.", "instruction": "Write test cases for: Test a formatter. Start with the edge cases and expected behavior.", "reference_solution": "Cover input validation, boundary values, async states, and error rendering.", "evaluation": {"checks": ["edge coverage", "async behavior", "clear expectations"]}, "tags": ["agentic", "coding", "tdd", "javascript", "saas", "persian-friendly"]}
{"id": "JT3K-001156", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "javascript", "difficulty": "easy", "title": "Test a formatter", "domain": "edtech", "user_story": "Write tests for a date formatter that must respect locale and fallback rules.", "repo_context": "Repo: mobile-sdk. Domain: edtech. JS testing stack with Vitest/Jest and React Testing Library.", "instruction": "Write test cases for: Test a formatter. Start with the edge cases and expected behavior.", "reference_solution": "Cover input validation, boundary values, async states, and error rendering.", "evaluation": {"checks": ["edge coverage", "async behavior", "clear expectations"]}, "tags": ["agentic", "coding", "tdd", "javascript", "edtech", "persian-friendly"]}
{"id": "JT3K-002690", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "typescript", "difficulty": "medium", "title": "Review an API client", "domain": "ecommerce", "user_story": "The client mixes transport errors and business errors in one catch block.", "repo_context": "Repo: chat-agent. Domain: ecommerce. Frontend app with shared utilities and API wrappers.", "instruction": "Review the implementation for: Review an API client. Comment on correctness, UX, and maintainability.", "reference_solution": "Look for async race conditions, unnecessary re-renders, error handling gaps, and accessibility issues.", "evaluation": {"checks": ["quality issues found", "practical suggestions", "frontend concerns"]}, "tags": ["agentic", "coding", "code_review", "typescript", "ecommerce", "persian-friendly"]}
{"id": "JT3K-002951", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "bash", "difficulty": "easy", "title": "Diagnose build failure", "domain": "healthtech", "user_story": "A CI script fails because one command exits early and hides the root cause.", "repo_context": "Repo: ecommerce-backend. Domain: healthtech. CI scripts, release automation, and shell helpers.", "instruction": "Diagnose this shell failure: Diagnose build failure. Explain the failure point and the safest fix.", "reference_solution": "Check exit codes, quote variables, and make failure handling explicit with set -euo pipefail when appropriate.", "evaluation": {"checks": ["failure point", "shell safety", "clear fix"]}, "tags": ["agentic", "coding", "debug_trace", "bash", "healthtech", "persian-friendly"]}
{"id": "JT3K-000141", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "javascript", "difficulty": "hard", "title": "Locate data fetching boundaries", "domain": "creator-tools", "user_story": "A Next.js repo separates server actions, client components, and shared utilities.", "repo_context": "Repo: dashboard-web. Domain: creator-tools. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Locate data fetching boundaries. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "javascript", "creator-tools", "persian-friendly"]}
{"id": "JT3K-002021", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "python", "difficulty": "medium", "title": "Trace a failing unit test", "domain": "saas", "user_story": "A test intermittently fails due to shared mutable state between cases.", "repo_context": "Repo: chat-agent. Domain: saas. Python service with logs, stack traces, and tests.", "instruction": "Diagnose this failure: Trace a failing unit test. Provide the most likely cause and the verification steps.", "reference_solution": "Read the stack trace, identify the failing boundary, isolate shared state or missing key paths, then add a regression test.", "evaluation": {"checks": ["root cause", "debug steps", "verification plan"]}, "tags": ["agentic", "coding", "debug_trace", "python", "saas", "persian-friendly"]}
{"id": "JT3K-002825", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "sql", "difficulty": "hard", "title": "Trace missing rows", "domain": "infra", "user_story": "A dashboard shows missing records after a timezone conversion.", "repo_context": "Repo: payments-api. Domain: infra. SQL pipeline and dashboard queries.", "instruction": "Trace the data issue: Trace missing rows. Explain likely cause and how to verify the fix.", "reference_solution": "Check timezone conversion, join cardinality, and filters that remove expected rows.", "evaluation": {"checks": ["root cause", "data verification", "metric consistency"]}, "tags": ["agentic", "coding", "debug_trace", "sql", "infra", "persian-friendly"]}
{"id": "JT3K-001749", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "python", "difficulty": "medium", "title": "Split a large function", "domain": "developer-tools", "user_story": "Refactor a 120-line data transformation function into small testable helpers.", "repo_context": "Repo: dashboard-web. Domain: developer-tools. Python codebase with repeated helpers and service objects.", "instruction": "Refactor the following situation: Split a large function. Preserve behavior and reduce complexity.", "reference_solution": "Extract helpers, keep public API stable, add/adjust tests, and avoid mixing business logic with I/O.", "evaluation": {"checks": ["behavior preserved", "complexity reduced", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "python", "developer-tools", "persian-friendly"]}
{"id": "JT3K-001180", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "sql", "difficulty": "medium", "title": "Understand schema relationships", "domain": "creator-tools", "user_story": "A warehouse schema uses fact and dimension tables for subscriptions and events.", "repo_context": "Repo: dashboard-web. Domain: creator-tools. Data warehouse with fact/dimension tables and scheduled jobs.", "instruction": "Explain the data flow for: Understand schema relationships. Identify tables, joins, and transformation steps.", "reference_solution": "Trace source tables to staging, then to marts. Identify key joins, filters, and aggregation points.", "evaluation": {"checks": ["table flow", "join logic", "aggregation path"]}, "tags": ["agentic", "coding", "repo_understanding", "sql", "creator-tools", "persian-friendly"]}
{"id": "JT3K-000345", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "python", "difficulty": "easy", "title": "Write tests before implementation", "domain": "healthtech", "user_story": "Create unit tests for a discount calculator covering edge cases and rounding.", "repo_context": "Repo: ml-pipeline. Domain: healthtech. Python package with pytest and typed functions.", "instruction": "Write a test-first plan for: Write tests before implementation. Include the key test cases before implementation.", "reference_solution": "Cover happy path, invalid input, boundaries, and failure modes. Keep tests deterministic and small.", "evaluation": {"checks": ["edge cases covered", "test names clear", "deterministic behavior"]}, "tags": ["agentic", "coding", "tdd", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-001446", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "typescript", "difficulty": "easy", "title": "Plan verify-edit cycle", "domain": "developer-tools", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: mobile-sdk. Domain: developer-tools. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "typescript", "developer-tools", "persian-friendly"]}
{"id": "JT3K-001567", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "sql", "difficulty": "easy", "title": "Trace missing rows", "domain": "fintech", "user_story": "A dashboard shows missing records after a timezone conversion.", "repo_context": "Repo: ml-pipeline. Domain: fintech. SQL pipeline and dashboard queries.", "instruction": "Trace the data issue: Trace missing rows. Explain likely cause and how to verify the fix.", "reference_solution": "Check timezone conversion, join cardinality, and filters that remove expected rows.", "evaluation": {"checks": ["root cause", "data verification", "metric consistency"]}, "tags": ["agentic", "coding", "debug_trace", "sql", "fintech", "persian-friendly"]}
{"id": "JT3K-001503", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "bash", "difficulty": "medium", "title": "Plan release verification", "domain": "developer-tools", "user_story": "Run lint, unit tests, packaging, and smoke checks before publishing.", "repo_context": "Repo: chat-agent. Domain: developer-tools. Release and verification scripts in bash.", "instruction": "Build a tool-use plan for: Plan release verification. Include commands to inspect, run, and validate.", "reference_solution": "List the exact commands in order: inspect files, run targeted checks, then package and smoke test.", "evaluation": {"checks": ["command order", "verification", "safe execution"]}, "tags": ["agentic", "coding", "tool_use", "bash", "developer-tools", "persian-friendly"]}
{"id": "JT3K-000080", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "sql", "difficulty": "medium", "title": "Trace missing rows", "domain": "creator-tools", "user_story": "A dashboard shows missing records after a timezone conversion.", "repo_context": "Repo: ecommerce-backend. Domain: creator-tools. SQL pipeline and dashboard queries.", "instruction": "Trace the data issue: Trace missing rows. Explain likely cause and how to verify the fix.", "reference_solution": "Check timezone conversion, join cardinality, and filters that remove expected rows.", "evaluation": {"checks": ["root cause", "data verification", "metric consistency"]}, "tags": ["agentic", "coding", "debug_trace", "sql", "creator-tools", "persian-friendly"]}
{"id": "JT3K-000794", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "python", "difficulty": "medium", "title": "Reduce duplication", "domain": "infra", "user_story": "The repo has repeated validation logic in three endpoints; centralize it.", "repo_context": "Repo: cms-platform. Domain: infra. Python codebase with repeated helpers and service objects.", "instruction": "Refactor the following situation: Reduce duplication. Preserve behavior and reduce complexity.", "reference_solution": "Extract helpers, keep public API stable, add/adjust tests, and avoid mixing business logic with I/O.", "evaluation": {"checks": ["behavior preserved", "complexity reduced", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "python", "infra", "persian-friendly"]}
{"id": "JT3K-002906", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "javascript", "difficulty": "easy", "title": "Guard null props", "domain": "infra", "user_story": "A component crashes when optional props are missing.", "repo_context": "Repo: payments-api. Domain: infra. React/Next.js application with hooks and server/client boundaries.", "instruction": "Fix this bug: Guard null props. Keep the change minimal and safe.", "reference_solution": "Guard missing values, avoid stale state, and keep server/client-only logic in the right layer.", "evaluation": {"checks": ["bug fixed", "safe boundary handling", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "javascript", "infra", "persian-friendly"]}
{"id": "JT3K-002753", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "python", "difficulty": "medium", "title": "Write tests before implementation", "domain": "developer-tools", "user_story": "Create unit tests for a discount calculator covering edge cases and rounding.", "repo_context": "Repo: mobile-sdk. Domain: developer-tools. Python package with pytest and typed functions.", "instruction": "Write a test-first plan for: Write tests before implementation. Include the key test cases before implementation.", "reference_solution": "Cover happy path, invalid input, boundaries, and failure modes. Keep tests deterministic and small.", "evaluation": {"checks": ["edge cases covered", "test names clear", "deterministic behavior"]}, "tags": ["agentic", "coding", "tdd", "python", "developer-tools", "persian-friendly"]}
{"id": "JT3K-001838", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "javascript", "difficulty": "medium", "title": "Review an API client", "domain": "edtech", "user_story": "The client mixes transport errors and business errors in one catch block.", "repo_context": "Repo: mobile-sdk. Domain: edtech. Frontend app with shared utilities and API wrappers.", "instruction": "Review the implementation for: Review an API client. Comment on correctness, UX, and maintainability.", "reference_solution": "Look for async race conditions, unnecessary re-renders, error handling gaps, and accessibility issues.", "evaluation": {"checks": ["quality issues found", "practical suggestions", "frontend concerns"]}, "tags": ["agentic", "coding", "code_review", "javascript", "edtech", "persian-friendly"]}
{"id": "JT3K-001329", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "sql", "difficulty": "easy", "title": "Design audit tables", "domain": "fintech", "user_story": "Design tables for tracking API events and change history.", "repo_context": "Repo: mobile-sdk. Domain: fintech. Analytics schema design.", "instruction": "Design the tables or schema for: Design audit tables. Define keys, partitions, and audit columns.", "reference_solution": "Use surrogate keys, timestamps, source metadata, and clear fact/dimension separation.", "evaluation": {"checks": ["schema clarity", "scalability", "auditability"]}, "tags": ["agentic", "coding", "api_design", "sql", "fintech", "persian-friendly"]}
{"id": "JT3K-001957", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "javascript", "difficulty": "medium", "title": "Plan verify-edit cycle", "domain": "healthtech", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: cms-platform. Domain: healthtech. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "javascript", "healthtech", "persian-friendly"]}
{"id": "JT3K-000117", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "javascript", "difficulty": "medium", "title": "Design client methods", "domain": "healthtech", "user_story": "Design a TypeScript client for search, pagination, and retries.", "repo_context": "Repo: cms-platform. Domain: healthtech. TypeScript client or Node API layer.", "instruction": "Design a TypeScript API for: Design client methods. Define method names, inputs, and return types.", "reference_solution": "Prefer explicit types, typed errors, pagination support, and composable options objects.", "evaluation": {"checks": ["typed contract", "clear ergonomics", "extensibility"]}, "tags": ["agentic", "coding", "api_design", "javascript", "healthtech", "persian-friendly"]}
{"id": "JT3K-000401", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "javascript", "difficulty": "easy", "title": "Plan verify-edit cycle", "domain": "developer-tools", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: payments-api. Domain: developer-tools. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "javascript", "developer-tools", "persian-friendly"]}
{"id": "JT3K-001490", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "typescript", "difficulty": "easy", "title": "Review an API client", "domain": "developer-tools", "user_story": "The client mixes transport errors and business errors in one catch block.", "repo_context": "Repo: cms-platform. Domain: developer-tools. Frontend app with shared utilities and API wrappers.", "instruction": "Review the implementation for: Review an API client. Comment on correctness, UX, and maintainability.", "reference_solution": "Look for async race conditions, unnecessary re-renders, error handling gaps, and accessibility issues.", "evaluation": {"checks": ["quality issues found", "practical suggestions", "frontend concerns"]}, "tags": ["agentic", "coding", "code_review", "typescript", "developer-tools", "persian-friendly"]}
{"id": "JT3K-002900", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "sql", "difficulty": "easy", "title": "Design audit tables", "domain": "saas", "user_story": "Design tables for tracking API events and change history.", "repo_context": "Repo: ml-pipeline. Domain: saas. Analytics schema design.", "instruction": "Design the tables or schema for: Design audit tables. Define keys, partitions, and audit columns.", "reference_solution": "Use surrogate keys, timestamps, source metadata, and clear fact/dimension separation.", "evaluation": {"checks": ["schema clarity", "scalability", "auditability"]}, "tags": ["agentic", "coding", "api_design", "sql", "saas", "persian-friendly"]}
{"id": "JT3K-001082", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "python", "difficulty": "easy", "title": "Review an async worker", "domain": "fintech", "user_story": "An async task uses blocking I/O in the event loop; suggest a safer pattern.", "repo_context": "Repo: ml-pipeline. Domain: fintech. Python backend with API handlers and background workers.", "instruction": "Review the code for: Review an async worker. Call out risks and give a practical improvement plan.", "reference_solution": "Flag correctness, performance, maintainability, and security concerns. Recommend the smallest safe refactor.", "evaluation": {"checks": ["risk identified", "actionable fix", "production awareness"]}, "tags": ["agentic", "coding", "code_review", "python", "fintech", "persian-friendly"]}
{"id": "JT3K-001323", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "python", "difficulty": "easy", "title": "Review a serializer", "domain": "ecommerce", "user_story": "The serializer leaks internal IDs in API responses; propose a cleaner contract.", "repo_context": "Repo: ecommerce-backend. Domain: ecommerce. Python backend with API handlers and background workers.", "instruction": "Review the code for: Review a serializer. Call out risks and give a practical improvement plan.", "reference_solution": "Flag correctness, performance, maintainability, and security concerns. Recommend the smallest safe refactor.", "evaluation": {"checks": ["risk identified", "actionable fix", "production awareness"]}, "tags": ["agentic", "coding", "code_review", "python", "ecommerce", "persian-friendly"]}
{"id": "JT3K-002954", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "javascript", "difficulty": "medium", "title": "Plan verify-edit cycle", "domain": "infra", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: scheduler. Domain: infra. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "javascript", "infra", "persian-friendly"]}
{"id": "JT3K-001719", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "sql", "difficulty": "medium", "title": "Write query tests", "domain": "creator-tools", "user_story": "Create validation cases for a revenue query with month boundaries.", "repo_context": "Repo: analytics-core. Domain: creator-tools. SQL validation and BI dashboards.", "instruction": "Define tests/validation cases for: Write query tests. Include edge and boundary conditions.", "reference_solution": "Use small fixture tables and validate row counts, null handling, and date windows.", "evaluation": {"checks": ["fixtures", "edge cases", "row-level expectations"]}, "tags": ["agentic", "coding", "tdd", "sql", "creator-tools", "persian-friendly"]}
{"id": "JT3K-001973", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "javascript", "difficulty": "easy", "title": "Diagnose promise flow", "domain": "saas", "user_story": "A promise chain swallows the original error and returns a misleading success.", "repo_context": "Repo: chat-agent. Domain: saas. Frontend app with hydration and async data flow.", "instruction": "Diagnose the issue: Diagnose promise flow. Explain likely cause and how to verify the fix.", "reference_solution": "Check server/client output mismatches, stale closures, promise handling, and state synchronization.", "evaluation": {"checks": ["root cause", "verification steps", "frontend-specific diagnosis"]}, "tags": ["agentic", "coding", "debug_trace", "javascript", "saas", "persian-friendly"]}
{"id": "JT3K-000066", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "python", "difficulty": "medium", "title": "Plan rate limiting", "domain": "healthtech", "user_story": "Add a rate limiter design that can scale across multiple workers.", "repo_context": "Repo: mobile-sdk. Domain: healthtech. Python backend exposing JSON APIs.", "instruction": "Design the API/contract for: Plan rate limiting. Define inputs, outputs, errors, and extensibility points.", "reference_solution": "Use explicit request/response schemas, stable error shapes, and version-friendly naming.", "evaluation": {"checks": ["clear contract", "consistent errors", "future-proof design"]}, "tags": ["agentic", "coding", "api_design", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-000058", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "typescript", "difficulty": "easy", "title": "Plan verify-edit cycle", "domain": "ecommerce", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: analytics-core. Domain: ecommerce. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "typescript", "ecommerce", "persian-friendly"]}
{"id": "JT3K-000259", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "python", "difficulty": "easy", "title": "Sequence tool calls", "domain": "saas", "user_story": "Construct a tool-use plan for fixing an import error across multiple files.", "repo_context": "Repo: scheduler. Domain: saas. Agentic coding workflow with file reads, tests, and patches.", "instruction": "Create a tool-use plan for: Sequence tool calls. Include the ideal sequence of read, edit, test, and verify actions.", "reference_solution": "Read the relevant file set, inspect failure output, patch the narrowest surface, rerun targeted tests, then broaden checks.", "evaluation": {"checks": ["tool order sensible", "minimal edit strategy", "verification included"]}, "tags": ["agentic", "coding", "tool_use", "python", "saas", "persian-friendly"]}
{"id": "JT3K-000004", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "python", "difficulty": "medium", "title": "Trace a failing unit test", "domain": "developer-tools", "user_story": "A test intermittently fails due to shared mutable state between cases.", "repo_context": "Repo: analytics-core. Domain: developer-tools. Python service with logs, stack traces, and tests.", "instruction": "Diagnose this failure: Trace a failing unit test. Provide the most likely cause and the verification steps.", "reference_solution": "Read the stack trace, identify the failing boundary, isolate shared state or missing key paths, then add a regression test.", "evaluation": {"checks": ["root cause", "debug steps", "verification plan"]}, "tags": ["agentic", "coding", "debug_trace", "python", "developer-tools", "persian-friendly"]}
{"id": "JT3K-000486", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "python", "difficulty": "medium", "title": "Coordinate verification", "domain": "infra", "user_story": "Describe a tool sequence that verifies lint, unit tests, and type checks.", "repo_context": "Repo: mobile-sdk. Domain: infra. Agentic coding workflow with file reads, tests, and patches.", "instruction": "Create a tool-use plan for: Coordinate verification. Include the ideal sequence of read, edit, test, and verify actions.", "reference_solution": "Read the relevant file set, inspect failure output, patch the narrowest surface, rerun targeted tests, then broaden checks.", "evaluation": {"checks": ["tool order sensible", "minimal edit strategy", "verification included"]}, "tags": ["agentic", "coding", "tool_use", "python", "infra", "persian-friendly"]}
{"id": "JT3K-000941", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "javascript", "difficulty": "medium", "title": "Extract reusable hook", "domain": "saas", "user_story": "Refactor repeated fetch logic into a custom React hook.", "repo_context": "Repo: analytics-core. Domain: saas. React components and shared logic in a frontend repo.", "instruction": "Refactor the following: Extract reusable hook. Reduce duplication and keep the public interface stable.", "reference_solution": "Extract a hook or utility, preserve props and return shape, and add tests around behavior.", "evaluation": {"checks": ["duplication removed", "interface stable", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "javascript", "saas", "persian-friendly"]}
{"id": "JT3K-001296", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "javascript", "difficulty": "hard", "title": "Locate data fetching boundaries", "domain": "saas", "user_story": "A Next.js repo separates server actions, client components, and shared utilities.", "repo_context": "Repo: payments-api. Domain: saas. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Locate data fetching boundaries. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "javascript", "saas", "persian-friendly"]}
{"id": "JT3K-002214", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "javascript", "difficulty": "medium", "title": "Extract reusable hook", "domain": "edtech", "user_story": "Refactor repeated fetch logic into a custom React hook.", "repo_context": "Repo: mobile-sdk. Domain: edtech. React components and shared logic in a frontend repo.", "instruction": "Refactor the following: Extract reusable hook. Reduce duplication and keep the public interface stable.", "reference_solution": "Extract a hook or utility, preserve props and return shape, and add tests around behavior.", "evaluation": {"checks": ["duplication removed", "interface stable", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "javascript", "edtech", "persian-friendly"]}
{"id": "JT3K-002600", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "sql", "difficulty": "hard", "title": "Design audit tables", "domain": "ecommerce", "user_story": "Design tables for tracking API events and change history.", "repo_context": "Repo: payments-api. Domain: ecommerce. Analytics schema design.", "instruction": "Design the tables or schema for: Design audit tables. Define keys, partitions, and audit columns.", "reference_solution": "Use surrogate keys, timestamps, source metadata, and clear fact/dimension separation.", "evaluation": {"checks": ["schema clarity", "scalability", "auditability"]}, "tags": ["agentic", "coding", "api_design", "sql", "ecommerce", "persian-friendly"]}
{"id": "JT3K-001438", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "javascript", "difficulty": "easy", "title": "Trace state across components", "domain": "edtech", "user_story": "A React app manages auth state in context, hooks, and a small API layer.", "repo_context": "Repo: cms-platform. Domain: edtech. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Trace state across components. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "javascript", "edtech", "persian-friendly"]}
{"id": "JT3K-001572", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "bash", "difficulty": "easy", "title": "Plan release verification", "domain": "creator-tools", "user_story": "Run lint, unit tests, packaging, and smoke checks before publishing.", "repo_context": "Repo: cms-platform. Domain: creator-tools. Release and verification scripts in bash.", "instruction": "Build a tool-use plan for: Plan release verification. Include commands to inspect, run, and validate.", "reference_solution": "List the exact commands in order: inspect files, run targeted checks, then package and smoke test.", "evaluation": {"checks": ["command order", "verification", "safe execution"]}, "tags": ["agentic", "coding", "tool_use", "bash", "creator-tools", "persian-friendly"]}
{"id": "JT3K-002864", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "typescript", "difficulty": "easy", "title": "Trace state across components", "domain": "ecommerce", "user_story": "A React app manages auth state in context, hooks, and a small API layer.", "repo_context": "Repo: payments-api. Domain: ecommerce. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Trace state across components. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "typescript", "ecommerce", "persian-friendly"]}
{"id": "JT3K-000772", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "bash", "difficulty": "hard", "title": "Diagnose build failure", "domain": "creator-tools", "user_story": "A CI script fails because one command exits early and hides the root cause.", "repo_context": "Repo: mobile-sdk. Domain: creator-tools. CI scripts, release automation, and shell helpers.", "instruction": "Diagnose this shell failure: Diagnose build failure. Explain the failure point and the safest fix.", "reference_solution": "Check exit codes, quote variables, and make failure handling explicit with set -euo pipefail when appropriate.", "evaluation": {"checks": ["failure point", "shell safety", "clear fix"]}, "tags": ["agentic", "coding", "debug_trace", "bash", "creator-tools", "persian-friendly"]}
{"id": "JT3K-002666", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "python", "difficulty": "medium", "title": "Sequence tool calls", "domain": "infra", "user_story": "Construct a tool-use plan for fixing an import error across multiple files.", "repo_context": "Repo: mobile-sdk. Domain: infra. Agentic coding workflow with file reads, tests, and patches.", "instruction": "Create a tool-use plan for: Sequence tool calls. Include the ideal sequence of read, edit, test, and verify actions.", "reference_solution": "Read the relevant file set, inspect failure output, patch the narrowest surface, rerun targeted tests, then broaden checks.", "evaluation": {"checks": ["tool order sensible", "minimal edit strategy", "verification included"]}, "tags": ["agentic", "coding", "tool_use", "python", "infra", "persian-friendly"]}
{"id": "JT3K-002941", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "javascript", "difficulty": "easy", "title": "Trace state across components", "domain": "infra", "user_story": "A React app manages auth state in context, hooks, and a small API layer.", "repo_context": "Repo: ml-pipeline. Domain: infra. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Trace state across components. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "javascript", "infra", "persian-friendly"]}
{"id": "JT3K-001288", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "sql", "difficulty": "medium", "title": "Design audit tables", "domain": "ecommerce", "user_story": "Design tables for tracking API events and change history.", "repo_context": "Repo: dashboard-web. Domain: ecommerce. Analytics schema design.", "instruction": "Design the tables or schema for: Design audit tables. Define keys, partitions, and audit columns.", "reference_solution": "Use surrogate keys, timestamps, source metadata, and clear fact/dimension separation.", "evaluation": {"checks": ["schema clarity", "scalability", "auditability"]}, "tags": ["agentic", "coding", "api_design", "sql", "ecommerce", "persian-friendly"]}
{"id": "JT3K-002326", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "javascript", "difficulty": "easy", "title": "Plan verify-edit cycle", "domain": "fintech", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: dashboard-web. Domain: fintech. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "javascript", "fintech", "persian-friendly"]}
{"id": "JT3K-001612", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "python", "difficulty": "hard", "title": "Reduce duplication", "domain": "healthtech", "user_story": "The repo has repeated validation logic in three endpoints; centralize it.", "repo_context": "Repo: chat-agent. Domain: healthtech. Python codebase with repeated helpers and service objects.", "instruction": "Refactor the following situation: Reduce duplication. Preserve behavior and reduce complexity.", "reference_solution": "Extract helpers, keep public API stable, add/adjust tests, and avoid mixing business logic with I/O.", "evaluation": {"checks": ["behavior preserved", "complexity reduced", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-001445", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "python", "difficulty": "medium", "title": "Review an async worker", "domain": "creator-tools", "user_story": "An async task uses blocking I/O in the event loop; suggest a safer pattern.", "repo_context": "Repo: dashboard-web. Domain: creator-tools. Python backend with API handlers and background workers.", "instruction": "Review the code for: Review an async worker. Call out risks and give a practical improvement plan.", "reference_solution": "Flag correctness, performance, maintainability, and security concerns. Recommend the smallest safe refactor.", "evaluation": {"checks": ["risk identified", "actionable fix", "production awareness"]}, "tags": ["agentic", "coding", "code_review", "python", "creator-tools", "persian-friendly"]}
{"id": "JT3K-001401", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "javascript", "difficulty": "medium", "title": "Fix stale closure", "domain": "fintech", "user_story": "A hook callback reads old state after rapid updates; patch it.", "repo_context": "Repo: analytics-core. Domain: fintech. React/Next.js application with hooks and server/client boundaries.", "instruction": "Fix this bug: Fix stale closure. Keep the change minimal and safe.", "reference_solution": "Guard missing values, avoid stale state, and keep server/client-only logic in the right layer.", "evaluation": {"checks": ["bug fixed", "safe boundary handling", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "javascript", "fintech", "persian-friendly"]}
{"id": "JT3K-000152", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "javascript", "difficulty": "hard", "title": "Guard null props", "domain": "edtech", "user_story": "A component crashes when optional props are missing.", "repo_context": "Repo: ml-pipeline. Domain: edtech. React/Next.js application with hooks and server/client boundaries.", "instruction": "Fix this bug: Guard null props. Keep the change minimal and safe.", "reference_solution": "Guard missing values, avoid stale state, and keep server/client-only logic in the right layer.", "evaluation": {"checks": ["bug fixed", "safe boundary handling", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "javascript", "edtech", "persian-friendly"]}
{"id": "JT3K-001789", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "python", "difficulty": "medium", "title": "Improve readability", "domain": "healthtech", "user_story": "Rename ambiguous variables and simplify nested conditionals in a config loader.", "repo_context": "Repo: auth-service. Domain: healthtech. Python codebase with repeated helpers and service objects.", "instruction": "Refactor the following situation: Improve readability. Preserve behavior and reduce complexity.", "reference_solution": "Extract helpers, keep public API stable, add/adjust tests, and avoid mixing business logic with I/O.", "evaluation": {"checks": ["behavior preserved", "complexity reduced", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-000641", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "javascript", "difficulty": "easy", "title": "Test form validation", "domain": "ecommerce", "user_story": "Create tests for password and email validation in a UI flow.", "repo_context": "Repo: cms-platform. Domain: ecommerce. JS testing stack with Vitest/Jest and React Testing Library.", "instruction": "Write test cases for: Test form validation. Start with the edge cases and expected behavior.", "reference_solution": "Cover input validation, boundary values, async states, and error rendering.", "evaluation": {"checks": ["edge coverage", "async behavior", "clear expectations"]}, "tags": ["agentic", "coding", "tdd", "javascript", "ecommerce", "persian-friendly"]}
{"id": "JT3K-000086", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "python", "difficulty": "easy", "title": "Define response shapes", "domain": "creator-tools", "user_story": "Design consistent API responses for success, validation errors, and server errors.", "repo_context": "Repo: ecommerce-backend. Domain: creator-tools. Python backend exposing JSON APIs.", "instruction": "Design the API/contract for: Define response shapes. Define inputs, outputs, errors, and extensibility points.", "reference_solution": "Use explicit request/response schemas, stable error shapes, and version-friendly naming.", "evaluation": {"checks": ["clear contract", "consistent errors", "future-proof design"]}, "tags": ["agentic", "coding", "api_design", "python", "creator-tools", "persian-friendly"]}
{"id": "JT3K-002108", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "typescript", "difficulty": "easy", "title": "Review a memoized component", "domain": "ecommerce", "user_story": "An over-memoized component hides actual prop changes.", "repo_context": "Repo: chat-agent. Domain: ecommerce. Frontend app with shared utilities and API wrappers.", "instruction": "Review the implementation for: Review a memoized component. Comment on correctness, UX, and maintainability.", "reference_solution": "Look for async race conditions, unnecessary re-renders, error handling gaps, and accessibility issues.", "evaluation": {"checks": ["quality issues found", "practical suggestions", "frontend concerns"]}, "tags": ["agentic", "coding", "code_review", "typescript", "ecommerce", "persian-friendly"]}
{"id": "JT3K-001059", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "medium", "title": "Prevent duplicate inserts", "domain": "ecommerce", "user_story": "A repository method can insert duplicate rows under retries. Add idempotency guards.", "repo_context": "Repo: scheduler. Domain: ecommerce. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Prevent duplicate inserts. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "ecommerce", "persian-friendly"]}
{"id": "JT3K-000375", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "javascript", "difficulty": "medium", "title": "Test a formatter", "domain": "fintech", "user_story": "Write tests for a date formatter that must respect locale and fallback rules.", "repo_context": "Repo: ml-pipeline. Domain: fintech. JS testing stack with Vitest/Jest and React Testing Library.", "instruction": "Write test cases for: Test a formatter. Start with the edge cases and expected behavior.", "reference_solution": "Cover input validation, boundary values, async states, and error rendering.", "evaluation": {"checks": ["edge coverage", "async behavior", "clear expectations"]}, "tags": ["agentic", "coding", "tdd", "javascript", "fintech", "persian-friendly"]}
{"id": "JT3K-002673", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "javascript", "difficulty": "hard", "title": "Design client methods", "domain": "infra", "user_story": "Design a TypeScript client for search, pagination, and retries.", "repo_context": "Repo: ml-pipeline. Domain: infra. TypeScript client or Node API layer.", "instruction": "Design a TypeScript API for: Design client methods. Define method names, inputs, and return types.", "reference_solution": "Prefer explicit types, typed errors, pagination support, and composable options objects.", "evaluation": {"checks": ["typed contract", "clear ergonomics", "extensibility"]}, "tags": ["agentic", "coding", "api_design", "javascript", "infra", "persian-friendly"]}
{"id": "JT3K-001725", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "sql", "difficulty": "hard", "title": "Fix duplicate aggregation", "domain": "creator-tools", "user_story": "A query double-counts rows after joining two one-to-many tables.", "repo_context": "Repo: chat-agent. Domain: creator-tools. SQL analytics pipeline with reporting queries.", "instruction": "Fix the query issue: Fix duplicate aggregation. Explain the bug and provide the corrected pattern.", "reference_solution": "Deduplicate before aggregating, use correct grain, and be careful with time boundaries.", "evaluation": {"checks": ["correct grain", "no double count", "boundary handling"]}, "tags": ["agentic", "coding", "bug_fix", "sql", "creator-tools", "persian-friendly"]}
{"id": "JT3K-002634", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "typescript", "difficulty": "easy", "title": "Locate data fetching boundaries", "domain": "ecommerce", "user_story": "A Next.js repo separates server actions, client components, and shared utilities.", "repo_context": "Repo: auth-service. Domain: ecommerce. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Locate data fetching boundaries. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "typescript", "ecommerce", "persian-friendly"]}
{"id": "JT3K-000820", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "javascript", "difficulty": "easy", "title": "Extract reusable hook", "domain": "creator-tools", "user_story": "Refactor repeated fetch logic into a custom React hook.", "repo_context": "Repo: ecommerce-backend. Domain: creator-tools. React components and shared logic in a frontend repo.", "instruction": "Refactor the following: Extract reusable hook. Reduce duplication and keep the public interface stable.", "reference_solution": "Extract a hook or utility, preserve props and return shape, and add tests around behavior.", "evaluation": {"checks": ["duplication removed", "interface stable", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "javascript", "creator-tools", "persian-friendly"]}
{"id": "JT3K-002022", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "medium", "title": "Fix timezone parsing", "domain": "healthtech", "user_story": "A datetime parser returns naive objects and breaks scheduling. Patch it to handle UTC offsets safely.", "repo_context": "Repo: dashboard-web. Domain: healthtech. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Fix timezone parsing. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-001948", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "typescript", "difficulty": "hard", "title": "Design client methods", "domain": "saas", "user_story": "Design a TypeScript client for search, pagination, and retries.", "repo_context": "Repo: auth-service. Domain: saas. TypeScript client or Node API layer.", "instruction": "Design a TypeScript API for: Design client methods. Define method names, inputs, and return types.", "reference_solution": "Prefer explicit types, typed errors, pagination support, and composable options objects.", "evaluation": {"checks": ["typed contract", "clear ergonomics", "extensibility"]}, "tags": ["agentic", "coding", "api_design", "typescript", "saas", "persian-friendly"]}
{"id": "JT3K-002380", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "sql", "difficulty": "hard", "title": "Review a reporting query", "domain": "fintech", "user_story": "The query is correct but slow due to an unbounded scan and unnecessary subquery.", "repo_context": "Repo: chat-agent. Domain: fintech. SQL reporting and ETL jobs.", "instruction": "Review the SQL for: Review a reporting query. Identify correctness and performance risks.", "reference_solution": "Flag full scans, unnecessary CTEs, non-sargable filters, and inconsistent joins.", "evaluation": {"checks": ["performance risk", "logic risk", "clear recommendation"]}, "tags": ["agentic", "coding", "code_review", "sql", "fintech", "persian-friendly"]}
{"id": "JT3K-001889", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "python", "difficulty": "medium", "title": "Identify the file to change for feature X", "domain": "fintech", "user_story": "A FastAPI repo stores schema logic in one module and persistence in another. Locate where to add a new field safely.", "repo_context": "Repo: ml-pipeline. Domain: fintech. Python service with app/, services/, repositories/, tests/.", "instruction": "Explain how this codebase handles the task: Identify the file to change for feature X. Identify the files and modules most likely involved.", "reference_solution": "1. Start at the route/controller.\n2. Follow service layer calls.\n3. Check repository and schema modules.\n4. Verify tests that cover the flow.", "evaluation": {"checks": ["correct entry point", "correct data flow", "relevant files identified"]}, "tags": ["agentic", "coding", "repo_understanding", "python", "fintech", "persian-friendly"]}
{"id": "JT3K-002253", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "python", "difficulty": "hard", "title": "Diagnose a race condition", "domain": "saas", "user_story": "Two concurrent requests occasionally overwrite each other in a cache-backed flow.", "repo_context": "Repo: cms-platform. Domain: saas. Python service with logs, stack traces, and tests.", "instruction": "Diagnose this failure: Diagnose a race condition. Provide the most likely cause and the verification steps.", "reference_solution": "Read the stack trace, identify the failing boundary, isolate shared state or missing key paths, then add a regression test.", "evaluation": {"checks": ["root cause", "debug steps", "verification plan"]}, "tags": ["agentic", "coding", "debug_trace", "python", "saas", "persian-friendly"]}
{"id": "JT3K-001919", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "sql", "difficulty": "medium", "title": "Review a reporting query", "domain": "ecommerce", "user_story": "The query is correct but slow due to an unbounded scan and unnecessary subquery.", "repo_context": "Repo: analytics-core. Domain: ecommerce. SQL reporting and ETL jobs.", "instruction": "Review the SQL for: Review a reporting query. Identify correctness and performance risks.", "reference_solution": "Flag full scans, unnecessary CTEs, non-sargable filters, and inconsistent joins.", "evaluation": {"checks": ["performance risk", "logic risk", "clear recommendation"]}, "tags": ["agentic", "coding", "code_review", "sql", "ecommerce", "persian-friendly"]}
{"id": "JT3K-001602", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "javascript", "difficulty": "easy", "title": "Diagnose promise flow", "domain": "saas", "user_story": "A promise chain swallows the original error and returns a misleading success.", "repo_context": "Repo: cms-platform. Domain: saas. Frontend app with hydration and async data flow.", "instruction": "Diagnose the issue: Diagnose promise flow. Explain likely cause and how to verify the fix.", "reference_solution": "Check server/client output mismatches, stale closures, promise handling, and state synchronization.", "evaluation": {"checks": ["root cause", "verification steps", "frontend-specific diagnosis"]}, "tags": ["agentic", "coding", "debug_trace", "javascript", "saas", "persian-friendly"]}
{"id": "JT3K-002655", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "python", "difficulty": "easy", "title": "Design a service interface", "domain": "fintech", "user_story": "Propose a clean Python interface for a notification service supporting email and SMS.", "repo_context": "Repo: auth-service. Domain: fintech. Python backend exposing JSON APIs.", "instruction": "Design the API/contract for: Design a service interface. Define inputs, outputs, errors, and extensibility points.", "reference_solution": "Use explicit request/response schemas, stable error shapes, and version-friendly naming.", "evaluation": {"checks": ["clear contract", "consistent errors", "future-proof design"]}, "tags": ["agentic", "coding", "api_design", "python", "fintech", "persian-friendly"]}
{"id": "JT3K-001866", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "easy", "title": "Fix timezone parsing", "domain": "fintech", "user_story": "A datetime parser returns naive objects and breaks scheduling. Patch it to handle UTC offsets safely.", "repo_context": "Repo: auth-service. Domain: fintech. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Fix timezone parsing. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "fintech", "persian-friendly"]}
{"id": "JT3K-002800", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "sql", "difficulty": "hard", "title": "Understand schema relationships", "domain": "fintech", "user_story": "A warehouse schema uses fact and dimension tables for subscriptions and events.", "repo_context": "Repo: auth-service. Domain: fintech. Data warehouse with fact/dimension tables and scheduled jobs.", "instruction": "Explain the data flow for: Understand schema relationships. Identify tables, joins, and transformation steps.", "reference_solution": "Trace source tables to staging, then to marts. Identify key joins, filters, and aggregation points.", "evaluation": {"checks": ["table flow", "join logic", "aggregation path"]}, "tags": ["agentic", "coding", "repo_understanding", "sql", "fintech", "persian-friendly"]}
{"id": "JT3K-002905", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "typescript", "difficulty": "easy", "title": "Diagnose promise flow", "domain": "infra", "user_story": "A promise chain swallows the original error and returns a misleading success.", "repo_context": "Repo: mobile-sdk. Domain: infra. Frontend app with hydration and async data flow.", "instruction": "Diagnose the issue: Diagnose promise flow. Explain likely cause and how to verify the fix.", "reference_solution": "Check server/client output mismatches, stale closures, promise handling, and state synchronization.", "evaluation": {"checks": ["root cause", "verification steps", "frontend-specific diagnosis"]}, "tags": ["agentic", "coding", "debug_trace", "typescript", "infra", "persian-friendly"]}
{"id": "JT3K-002137", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "bash", "difficulty": "medium", "title": "Diagnose build failure", "domain": "fintech", "user_story": "A CI script fails because one command exits early and hides the root cause.", "repo_context": "Repo: auth-service. Domain: fintech. CI scripts, release automation, and shell helpers.", "instruction": "Diagnose this shell failure: Diagnose build failure. Explain the failure point and the safest fix.", "reference_solution": "Check exit codes, quote variables, and make failure handling explicit with set -euo pipefail when appropriate.", "evaluation": {"checks": ["failure point", "shell safety", "clear fix"]}, "tags": ["agentic", "coding", "debug_trace", "bash", "fintech", "persian-friendly"]}
{"id": "JT3K-002682", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "python", "difficulty": "hard", "title": "Explain a stack trace", "domain": "saas", "user_story": "A KeyError appears in production when a field is absent in a nested payload.", "repo_context": "Repo: mobile-sdk. Domain: saas. Python service with logs, stack traces, and tests.", "instruction": "Diagnose this failure: Explain a stack trace. Provide the most likely cause and the verification steps.", "reference_solution": "Read the stack trace, identify the failing boundary, isolate shared state or missing key paths, then add a regression test.", "evaluation": {"checks": ["root cause", "debug steps", "verification plan"]}, "tags": ["agentic", "coding", "debug_trace", "python", "saas", "persian-friendly"]}
{"id": "JT3K-002423", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "python", "difficulty": "easy", "title": "Review a cache helper", "domain": "ecommerce", "user_story": "A cache wrapper forgets to namespace keys and can collide across tenants.", "repo_context": "Repo: dashboard-web. Domain: ecommerce. Python backend with API handlers and background workers.", "instruction": "Review the code for: Review a cache helper. Call out risks and give a practical improvement plan.", "reference_solution": "Flag correctness, performance, maintainability, and security concerns. Recommend the smallest safe refactor.", "evaluation": {"checks": ["risk identified", "actionable fix", "production awareness"]}, "tags": ["agentic", "coding", "code_review", "python", "ecommerce", "persian-friendly"]}
{"id": "JT3K-002826", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "javascript", "difficulty": "easy", "title": "Trace state across components", "domain": "edtech", "user_story": "A React app manages auth state in context, hooks, and a small API layer.", "repo_context": "Repo: ecommerce-backend. Domain: edtech. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Trace state across components. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "javascript", "edtech", "persian-friendly"]}
{"id": "JT3K-001467", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "sql", "difficulty": "medium", "title": "Fix duplicate aggregation", "domain": "healthtech", "user_story": "A query double-counts rows after joining two one-to-many tables.", "repo_context": "Repo: ml-pipeline. Domain: healthtech. SQL analytics pipeline with reporting queries.", "instruction": "Fix the query issue: Fix duplicate aggregation. Explain the bug and provide the corrected pattern.", "reference_solution": "Deduplicate before aggregating, use correct grain, and be careful with time boundaries.", "evaluation": {"checks": ["correct grain", "no double count", "boundary handling"]}, "tags": ["agentic", "coding", "bug_fix", "sql", "healthtech", "persian-friendly"]}
{"id": "JT3K-000320", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "javascript", "difficulty": "easy", "title": "Review an API client", "domain": "healthtech", "user_story": "The client mixes transport errors and business errors in one catch block.", "repo_context": "Repo: dashboard-web. Domain: healthtech. Frontend app with shared utilities and API wrappers.", "instruction": "Review the implementation for: Review an API client. Comment on correctness, UX, and maintainability.", "reference_solution": "Look for async race conditions, unnecessary re-renders, error handling gaps, and accessibility issues.", "evaluation": {"checks": ["quality issues found", "practical suggestions", "frontend concerns"]}, "tags": ["agentic", "coding", "code_review", "javascript", "healthtech", "persian-friendly"]}
{"id": "JT3K-000708", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "typescript", "difficulty": "easy", "title": "Test a formatter", "domain": "ecommerce", "user_story": "Write tests for a date formatter that must respect locale and fallback rules.", "repo_context": "Repo: analytics-core. Domain: ecommerce. JS testing stack with Vitest/Jest and React Testing Library.", "instruction": "Write test cases for: Test a formatter. Start with the edge cases and expected behavior.", "reference_solution": "Cover input validation, boundary values, async states, and error rendering.", "evaluation": {"checks": ["edge coverage", "async behavior", "clear expectations"]}, "tags": ["agentic", "coding", "tdd", "typescript", "ecommerce", "persian-friendly"]}
{"id": "JT3K-000891", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "sql", "difficulty": "medium", "title": "Understand schema relationships", "domain": "developer-tools", "user_story": "A warehouse schema uses fact and dimension tables for subscriptions and events.", "repo_context": "Repo: auth-service. Domain: developer-tools. Data warehouse with fact/dimension tables and scheduled jobs.", "instruction": "Explain the data flow for: Understand schema relationships. Identify tables, joins, and transformation steps.", "reference_solution": "Trace source tables to staging, then to marts. Identify key joins, filters, and aggregation points.", "evaluation": {"checks": ["table flow", "join logic", "aggregation path"]}, "tags": ["agentic", "coding", "repo_understanding", "sql", "developer-tools", "persian-friendly"]}
{"id": "JT3K-000713", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "javascript", "difficulty": "easy", "title": "Review an API client", "domain": "fintech", "user_story": "The client mixes transport errors and business errors in one catch block.", "repo_context": "Repo: ml-pipeline. Domain: fintech. Frontend app with shared utilities and API wrappers.", "instruction": "Review the implementation for: Review an API client. Comment on correctness, UX, and maintainability.", "reference_solution": "Look for async race conditions, unnecessary re-renders, error handling gaps, and accessibility issues.", "evaluation": {"checks": ["quality issues found", "practical suggestions", "frontend concerns"]}, "tags": ["agentic", "coding", "code_review", "javascript", "fintech", "persian-friendly"]}
{"id": "JT3K-001196", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "javascript", "difficulty": "medium", "title": "Test a formatter", "domain": "edtech", "user_story": "Write tests for a date formatter that must respect locale and fallback rules.", "repo_context": "Repo: ml-pipeline. Domain: edtech. JS testing stack with Vitest/Jest and React Testing Library.", "instruction": "Write test cases for: Test a formatter. Start with the edge cases and expected behavior.", "reference_solution": "Cover input validation, boundary values, async states, and error rendering.", "evaluation": {"checks": ["edge coverage", "async behavior", "clear expectations"]}, "tags": ["agentic", "coding", "tdd", "javascript", "edtech", "persian-friendly"]}
{"id": "JT3K-002466", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "python", "difficulty": "easy", "title": "Plan a repair loop", "domain": "edtech", "user_story": "The agent should read files, run tests, inspect failures, patch code, and re-run checks.", "repo_context": "Repo: chat-agent. Domain: edtech. Agentic coding workflow with file reads, tests, and patches.", "instruction": "Create a tool-use plan for: Plan a repair loop. Include the ideal sequence of read, edit, test, and verify actions.", "reference_solution": "Read the relevant file set, inspect failure output, patch the narrowest surface, rerun targeted tests, then broaden checks.", "evaluation": {"checks": ["tool order sensible", "minimal edit strategy", "verification included"]}, "tags": ["agentic", "coding", "tool_use", "python", "edtech", "persian-friendly"]}
{"id": "JT3K-001314", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "javascript", "difficulty": "hard", "title": "Diagnose promise flow", "domain": "creator-tools", "user_story": "A promise chain swallows the original error and returns a misleading success.", "repo_context": "Repo: analytics-core. Domain: creator-tools. Frontend app with hydration and async data flow.", "instruction": "Diagnose the issue: Diagnose promise flow. Explain likely cause and how to verify the fix.", "reference_solution": "Check server/client output mismatches, stale closures, promise handling, and state synchronization.", "evaluation": {"checks": ["root cause", "verification steps", "frontend-specific diagnosis"]}, "tags": ["agentic", "coding", "debug_trace", "javascript", "creator-tools", "persian-friendly"]}
{"id": "JT3K-002111", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "typescript", "difficulty": "hard", "title": "Analyze hydration mismatch", "domain": "fintech", "user_story": "A server-rendered page renders different timestamps on client and server.", "repo_context": "Repo: payments-api. Domain: fintech. Frontend app with hydration and async data flow.", "instruction": "Diagnose the issue: Analyze hydration mismatch. Explain likely cause and how to verify the fix.", "reference_solution": "Check server/client output mismatches, stale closures, promise handling, and state synchronization.", "evaluation": {"checks": ["root cause", "verification steps", "frontend-specific diagnosis"]}, "tags": ["agentic", "coding", "debug_trace", "typescript", "fintech", "persian-friendly"]}
{"id": "JT3K-002058", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "typescript", "difficulty": "medium", "title": "Trace state across components", "domain": "infra", "user_story": "A React app manages auth state in context, hooks, and a small API layer.", "repo_context": "Repo: dashboard-web. Domain: infra. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Trace state across components. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "typescript", "infra", "persian-friendly"]}
{"id": "JT3K-001988", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "sql", "difficulty": "easy", "title": "Understand schema relationships", "domain": "fintech", "user_story": "A warehouse schema uses fact and dimension tables for subscriptions and events.", "repo_context": "Repo: ecommerce-backend. Domain: fintech. Data warehouse with fact/dimension tables and scheduled jobs.", "instruction": "Explain the data flow for: Understand schema relationships. Identify tables, joins, and transformation steps.", "reference_solution": "Trace source tables to staging, then to marts. Identify key joins, filters, and aggregation points.", "evaluation": {"checks": ["table flow", "join logic", "aggregation path"]}, "tags": ["agentic", "coding", "repo_understanding", "sql", "fintech", "persian-friendly"]}
{"id": "JT3K-000091", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "sql", "difficulty": "medium", "title": "Plan migration checks", "domain": "fintech", "user_story": "Inspect tables, compare counts, run a diff query, and validate migration impact.", "repo_context": "Repo: dashboard-web. Domain: fintech. Agent workflow over SQL files and validation queries.", "instruction": "Create a tool-use plan for: Plan migration checks. Read schema, inspect query, run checks, and validate output.", "reference_solution": "Inspect schema first, isolate the grain, run a comparison query, then verify counts after the patch.", "evaluation": {"checks": ["schema first", "comparison query", "verification"]}, "tags": ["agentic", "coding", "tool_use", "sql", "fintech", "persian-friendly"]}
{"id": "JT3K-002806", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "javascript", "difficulty": "easy", "title": "Trace state across components", "domain": "saas", "user_story": "A React app manages auth state in context, hooks, and a small API layer.", "repo_context": "Repo: scheduler. Domain: saas. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Trace state across components. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "javascript", "saas", "persian-friendly"]}
{"id": "JT3K-001709", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "easy", "title": "Prevent duplicate inserts", "domain": "edtech", "user_story": "A repository method can insert duplicate rows under retries. Add idempotency guards.", "repo_context": "Repo: dashboard-web. Domain: edtech. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Prevent duplicate inserts. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "edtech", "persian-friendly"]}
{"id": "JT3K-000763", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "python", "difficulty": "medium", "title": "Plan rate limiting", "domain": "creator-tools", "user_story": "Add a rate limiter design that can scale across multiple workers.", "repo_context": "Repo: ecommerce-backend. Domain: creator-tools. Python backend exposing JSON APIs.", "instruction": "Design the API/contract for: Plan rate limiting. Define inputs, outputs, errors, and extensibility points.", "reference_solution": "Use explicit request/response schemas, stable error shapes, and version-friendly naming.", "evaluation": {"checks": ["clear contract", "consistent errors", "future-proof design"]}, "tags": ["agentic", "coding", "api_design", "python", "creator-tools", "persian-friendly"]}
{"id": "JT3K-002250", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "easy", "title": "Prevent duplicate inserts", "domain": "infra", "user_story": "A repository method can insert duplicate rows under retries. Add idempotency guards.", "repo_context": "Repo: ecommerce-backend. Domain: infra. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Prevent duplicate inserts. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "infra", "persian-friendly"]}
{"id": "JT3K-002276", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "sql", "difficulty": "medium", "title": "Simplify a report query", "domain": "healthtech", "user_story": "Refactor a deeply nested CTE chain for readability.", "repo_context": "Repo: chat-agent. Domain: healthtech. SQL queries with layered CTEs.", "instruction": "Refactor the query for: Simplify a report query. Make it easier to read without changing results.", "reference_solution": "Reduce nested CTE depth, standardize aliases, and keep one clear grain per step.", "evaluation": {"checks": ["readability", "stable results", "clear grain"]}, "tags": ["agentic", "coding", "refactor", "sql", "healthtech", "persian-friendly"]}
{"id": "JT3K-000625", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "sql", "difficulty": "medium", "title": "Review a reporting query", "domain": "healthtech", "user_story": "The query is correct but slow due to an unbounded scan and unnecessary subquery.", "repo_context": "Repo: ecommerce-backend. Domain: healthtech. SQL reporting and ETL jobs.", "instruction": "Review the SQL for: Review a reporting query. Identify correctness and performance risks.", "reference_solution": "Flag full scans, unnecessary CTEs, non-sargable filters, and inconsistent joins.", "evaluation": {"checks": ["performance risk", "logic risk", "clear recommendation"]}, "tags": ["agentic", "coding", "code_review", "sql", "healthtech", "persian-friendly"]}
{"id": "JT3K-000717", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "javascript", "difficulty": "easy", "title": "Plan verify-edit cycle", "domain": "developer-tools", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: dashboard-web. Domain: developer-tools. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "javascript", "developer-tools", "persian-friendly"]}
{"id": "JT3K-000241", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "javascript", "difficulty": "hard", "title": "Locate data fetching boundaries", "domain": "fintech", "user_story": "A Next.js repo separates server actions, client components, and shared utilities.", "repo_context": "Repo: cms-platform. Domain: fintech. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Locate data fetching boundaries. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "javascript", "fintech", "persian-friendly"]}
{"id": "JT3K-000582", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "python", "difficulty": "easy", "title": "Find hidden coupling", "domain": "fintech", "user_story": "Inspect a Python package with utils, handlers, and workers; explain how retry logic affects job processing.", "repo_context": "Repo: analytics-core. Domain: fintech. Python service with app/, services/, repositories/, tests/.", "instruction": "Explain how this codebase handles the task: Find hidden coupling. Identify the files and modules most likely involved.", "reference_solution": "1. Start at the route/controller.\n2. Follow service layer calls.\n3. Check repository and schema modules.\n4. Verify tests that cover the flow.", "evaluation": {"checks": ["correct entry point", "correct data flow", "relevant files identified"]}, "tags": ["agentic", "coding", "repo_understanding", "python", "fintech", "persian-friendly"]}
{"id": "JT3K-002611", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "typescript", "difficulty": "easy", "title": "Analyze hydration mismatch", "domain": "creator-tools", "user_story": "A server-rendered page renders different timestamps on client and server.", "repo_context": "Repo: chat-agent. Domain: creator-tools. Frontend app with hydration and async data flow.", "instruction": "Diagnose the issue: Analyze hydration mismatch. Explain likely cause and how to verify the fix.", "reference_solution": "Check server/client output mismatches, stale closures, promise handling, and state synchronization.", "evaluation": {"checks": ["root cause", "verification steps", "frontend-specific diagnosis"]}, "tags": ["agentic", "coding", "debug_trace", "typescript", "creator-tools", "persian-friendly"]}
{"id": "JT3K-000618", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "bash", "difficulty": "easy", "title": "Diagnose build failure", "domain": "fintech", "user_story": "A CI script fails because one command exits early and hides the root cause.", "repo_context": "Repo: scheduler. Domain: fintech. CI scripts, release automation, and shell helpers.", "instruction": "Diagnose this shell failure: Diagnose build failure. Explain the failure point and the safest fix.", "reference_solution": "Check exit codes, quote variables, and make failure handling explicit with set -euo pipefail when appropriate.", "evaluation": {"checks": ["failure point", "shell safety", "clear fix"]}, "tags": ["agentic", "coding", "debug_trace", "bash", "fintech", "persian-friendly"]}
{"id": "JT3K-001745", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "python", "difficulty": "easy", "title": "Sequence tool calls", "domain": "infra", "user_story": "Construct a tool-use plan for fixing an import error across multiple files.", "repo_context": "Repo: analytics-core. Domain: infra. Agentic coding workflow with file reads, tests, and patches.", "instruction": "Create a tool-use plan for: Sequence tool calls. Include the ideal sequence of read, edit, test, and verify actions.", "reference_solution": "Read the relevant file set, inspect failure output, patch the narrowest surface, rerun targeted tests, then broaden checks.", "evaluation": {"checks": ["tool order sensible", "minimal edit strategy", "verification included"]}, "tags": ["agentic", "coding", "tool_use", "python", "infra", "persian-friendly"]}
{"id": "JT3K-002586", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "python", "difficulty": "easy", "title": "Coordinate verification", "domain": "developer-tools", "user_story": "Describe a tool sequence that verifies lint, unit tests, and type checks.", "repo_context": "Repo: auth-service. Domain: developer-tools. Agentic coding workflow with file reads, tests, and patches.", "instruction": "Create a tool-use plan for: Coordinate verification. Include the ideal sequence of read, edit, test, and verify actions.", "reference_solution": "Read the relevant file set, inspect failure output, patch the narrowest surface, rerun targeted tests, then broaden checks.", "evaluation": {"checks": ["tool order sensible", "minimal edit strategy", "verification included"]}, "tags": ["agentic", "coding", "tool_use", "python", "developer-tools", "persian-friendly"]}
{"id": "JT3K-001409", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "python", "difficulty": "easy", "title": "Review a cache helper", "domain": "healthtech", "user_story": "A cache wrapper forgets to namespace keys and can collide across tenants.", "repo_context": "Repo: payments-api. Domain: healthtech. Python backend with API handlers and background workers.", "instruction": "Review the code for: Review a cache helper. Call out risks and give a practical improvement plan.", "reference_solution": "Flag correctness, performance, maintainability, and security concerns. Recommend the smallest safe refactor.", "evaluation": {"checks": ["risk identified", "actionable fix", "production awareness"]}, "tags": ["agentic", "coding", "code_review", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-001837", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "bash", "difficulty": "hard", "title": "Diagnose build failure", "domain": "infra", "user_story": "A CI script fails because one command exits early and hides the root cause.", "repo_context": "Repo: mobile-sdk. Domain: infra. CI scripts, release automation, and shell helpers.", "instruction": "Diagnose this shell failure: Diagnose build failure. Explain the failure point and the safest fix.", "reference_solution": "Check exit codes, quote variables, and make failure handling explicit with set -euo pipefail when appropriate.", "evaluation": {"checks": ["failure point", "shell safety", "clear fix"]}, "tags": ["agentic", "coding", "debug_trace", "bash", "infra", "persian-friendly"]}
{"id": "JT3K-001949", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "python", "difficulty": "easy", "title": "Coordinate verification", "domain": "edtech", "user_story": "Describe a tool sequence that verifies lint, unit tests, and type checks.", "repo_context": "Repo: analytics-core. Domain: edtech. Agentic coding workflow with file reads, tests, and patches.", "instruction": "Create a tool-use plan for: Coordinate verification. Include the ideal sequence of read, edit, test, and verify actions.", "reference_solution": "Read the relevant file set, inspect failure output, patch the narrowest surface, rerun targeted tests, then broaden checks.", "evaluation": {"checks": ["tool order sensible", "minimal edit strategy", "verification included"]}, "tags": ["agentic", "coding", "tool_use", "python", "edtech", "persian-friendly"]}
{"id": "JT3K-000604", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "python", "difficulty": "hard", "title": "Trace a failing unit test", "domain": "healthtech", "user_story": "A test intermittently fails due to shared mutable state between cases.", "repo_context": "Repo: dashboard-web. Domain: healthtech. Python service with logs, stack traces, and tests.", "instruction": "Diagnose this failure: Trace a failing unit test. Provide the most likely cause and the verification steps.", "reference_solution": "Read the stack trace, identify the failing boundary, isolate shared state or missing key paths, then add a regression test.", "evaluation": {"checks": ["root cause", "debug steps", "verification plan"]}, "tags": ["agentic", "coding", "debug_trace", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-000812", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "python", "difficulty": "medium", "title": "Improve readability", "domain": "saas", "user_story": "Rename ambiguous variables and simplify nested conditionals in a config loader.", "repo_context": "Repo: mobile-sdk. Domain: saas. Python codebase with repeated helpers and service objects.", "instruction": "Refactor the following situation: Improve readability. Preserve behavior and reduce complexity.", "reference_solution": "Extract helpers, keep public API stable, add/adjust tests, and avoid mixing business logic with I/O.", "evaluation": {"checks": ["behavior preserved", "complexity reduced", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "python", "saas", "persian-friendly"]}
{"id": "JT3K-002672", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "bash", "difficulty": "easy", "title": "Plan release verification", "domain": "developer-tools", "user_story": "Run lint, unit tests, packaging, and smoke checks before publishing.", "repo_context": "Repo: ecommerce-backend. Domain: developer-tools. Release and verification scripts in bash.", "instruction": "Build a tool-use plan for: Plan release verification. Include commands to inspect, run, and validate.", "reference_solution": "List the exact commands in order: inspect files, run targeted checks, then package and smoke test.", "evaluation": {"checks": ["command order", "verification", "safe execution"]}, "tags": ["agentic", "coding", "tool_use", "bash", "developer-tools", "persian-friendly"]}
{"id": "JT3K-000914", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "javascript", "difficulty": "easy", "title": "Fix stale closure", "domain": "ecommerce", "user_story": "A hook callback reads old state after rapid updates; patch it.", "repo_context": "Repo: payments-api. Domain: ecommerce. React/Next.js application with hooks and server/client boundaries.", "instruction": "Fix this bug: Fix stale closure. Keep the change minimal and safe.", "reference_solution": "Guard missing values, avoid stale state, and keep server/client-only logic in the right layer.", "evaluation": {"checks": ["bug fixed", "safe boundary handling", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "javascript", "ecommerce", "persian-friendly"]}
{"id": "JT3K-001630", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "python", "difficulty": "easy", "title": "Trace a failing unit test", "domain": "saas", "user_story": "A test intermittently fails due to shared mutable state between cases.", "repo_context": "Repo: ml-pipeline. Domain: saas. Python service with logs, stack traces, and tests.", "instruction": "Diagnose this failure: Trace a failing unit test. Provide the most likely cause and the verification steps.", "reference_solution": "Read the stack trace, identify the failing boundary, isolate shared state or missing key paths, then add a regression test.", "evaluation": {"checks": ["root cause", "debug steps", "verification plan"]}, "tags": ["agentic", "coding", "debug_trace", "python", "saas", "persian-friendly"]}
{"id": "JT3K-002495", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "python", "difficulty": "medium", "title": "Trace a failing unit test", "domain": "developer-tools", "user_story": "A test intermittently fails due to shared mutable state between cases.", "repo_context": "Repo: chat-agent. Domain: developer-tools. Python service with logs, stack traces, and tests.", "instruction": "Diagnose this failure: Trace a failing unit test. Provide the most likely cause and the verification steps.", "reference_solution": "Read the stack trace, identify the failing boundary, isolate shared state or missing key paths, then add a regression test.", "evaluation": {"checks": ["root cause", "debug steps", "verification plan"]}, "tags": ["agentic", "coding", "debug_trace", "python", "developer-tools", "persian-friendly"]}
{"id": "JT3K-002307", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "medium", "title": "Fix timezone parsing", "domain": "healthtech", "user_story": "A datetime parser returns naive objects and breaks scheduling. Patch it to handle UTC offsets safely.", "repo_context": "Repo: auth-service. Domain: healthtech. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Fix timezone parsing. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-000049", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "python", "difficulty": "medium", "title": "Improve readability", "domain": "creator-tools", "user_story": "Rename ambiguous variables and simplify nested conditionals in a config loader.", "repo_context": "Repo: dashboard-web. Domain: creator-tools. Python codebase with repeated helpers and service objects.", "instruction": "Refactor the following situation: Improve readability. Preserve behavior and reduce complexity.", "reference_solution": "Extract helpers, keep public API stable, add/adjust tests, and avoid mixing business logic with I/O.", "evaluation": {"checks": ["behavior preserved", "complexity reduced", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "python", "creator-tools", "persian-friendly"]}
{"id": "JT3K-000791", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "sql", "difficulty": "medium", "title": "Trace missing rows", "domain": "saas", "user_story": "A dashboard shows missing records after a timezone conversion.", "repo_context": "Repo: payments-api. Domain: saas. SQL pipeline and dashboard queries.", "instruction": "Trace the data issue: Trace missing rows. Explain likely cause and how to verify the fix.", "reference_solution": "Check timezone conversion, join cardinality, and filters that remove expected rows.", "evaluation": {"checks": ["root cause", "data verification", "metric consistency"]}, "tags": ["agentic", "coding", "debug_trace", "sql", "saas", "persian-friendly"]}
{"id": "JT3K-000605", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "typescript", "difficulty": "easy", "title": "Extract reusable hook", "domain": "creator-tools", "user_story": "Refactor repeated fetch logic into a custom React hook.", "repo_context": "Repo: chat-agent. Domain: creator-tools. React components and shared logic in a frontend repo.", "instruction": "Refactor the following: Extract reusable hook. Reduce duplication and keep the public interface stable.", "reference_solution": "Extract a hook or utility, preserve props and return shape, and add tests around behavior.", "evaluation": {"checks": ["duplication removed", "interface stable", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "typescript", "creator-tools", "persian-friendly"]}
{"id": "JT3K-002527", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "sql", "difficulty": "easy", "title": "Simplify a report query", "domain": "healthtech", "user_story": "Refactor a deeply nested CTE chain for readability.", "repo_context": "Repo: dashboard-web. Domain: healthtech. SQL queries with layered CTEs.", "instruction": "Refactor the query for: Simplify a report query. Make it easier to read without changing results.", "reference_solution": "Reduce nested CTE depth, standardize aliases, and keep one clear grain per step.", "evaluation": {"checks": ["readability", "stable results", "clear grain"]}, "tags": ["agentic", "coding", "refactor", "sql", "healthtech", "persian-friendly"]}
{"id": "JT3K-000082", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "python", "difficulty": "easy", "title": "Improve readability", "domain": "creator-tools", "user_story": "Rename ambiguous variables and simplify nested conditionals in a config loader.", "repo_context": "Repo: ecommerce-backend. Domain: creator-tools. Python codebase with repeated helpers and service objects.", "instruction": "Refactor the following situation: Improve readability. Preserve behavior and reduce complexity.", "reference_solution": "Extract helpers, keep public API stable, add/adjust tests, and avoid mixing business logic with I/O.", "evaluation": {"checks": ["behavior preserved", "complexity reduced", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "python", "creator-tools", "persian-friendly"]}
{"id": "JT3K-002271", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "easy", "title": "Handle empty payloads", "domain": "infra", "user_story": "A webhook handler crashes when the JSON body is empty. Make it resilient and return a proper error.", "repo_context": "Repo: payments-api. Domain: infra. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Handle empty payloads. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "infra", "persian-friendly"]}
{"id": "JT3K-002500", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "sql", "difficulty": "hard", "title": "Fix duplicate aggregation", "domain": "developer-tools", "user_story": "A query double-counts rows after joining two one-to-many tables.", "repo_context": "Repo: dashboard-web. Domain: developer-tools. SQL analytics pipeline with reporting queries.", "instruction": "Fix the query issue: Fix duplicate aggregation. Explain the bug and provide the corrected pattern.", "reference_solution": "Deduplicate before aggregating, use correct grain, and be careful with time boundaries.", "evaluation": {"checks": ["correct grain", "no double count", "boundary handling"]}, "tags": ["agentic", "coding", "bug_fix", "sql", "developer-tools", "persian-friendly"]}
{"id": "JT3K-002198", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "python", "difficulty": "easy", "title": "Plan a repair loop", "domain": "saas", "user_story": "The agent should read files, run tests, inspect failures, patch code, and re-run checks.", "repo_context": "Repo: auth-service. Domain: saas. Agentic coding workflow with file reads, tests, and patches.", "instruction": "Create a tool-use plan for: Plan a repair loop. Include the ideal sequence of read, edit, test, and verify actions.", "reference_solution": "Read the relevant file set, inspect failure output, patch the narrowest surface, rerun targeted tests, then broaden checks.", "evaluation": {"checks": ["tool order sensible", "minimal edit strategy", "verification included"]}, "tags": ["agentic", "coding", "tool_use", "python", "saas", "persian-friendly"]}
{"id": "JT3K-002528", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "python", "difficulty": "hard", "title": "Review a serializer", "domain": "fintech", "user_story": "The serializer leaks internal IDs in API responses; propose a cleaner contract.", "repo_context": "Repo: mobile-sdk. Domain: fintech. Python backend with API handlers and background workers.", "instruction": "Review the code for: Review a serializer. Call out risks and give a practical improvement plan.", "reference_solution": "Flag correctness, performance, maintainability, and security concerns. Recommend the smallest safe refactor.", "evaluation": {"checks": ["risk identified", "actionable fix", "production awareness"]}, "tags": ["agentic", "coding", "code_review", "python", "fintech", "persian-friendly"]}
{"id": "JT3K-000489", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "python", "difficulty": "hard", "title": "Identify the file to change for feature X", "domain": "infra", "user_story": "A FastAPI repo stores schema logic in one module and persistence in another. Locate where to add a new field safely.", "repo_context": "Repo: mobile-sdk. Domain: infra. Python service with app/, services/, repositories/, tests/.", "instruction": "Explain how this codebase handles the task: Identify the file to change for feature X. Identify the files and modules most likely involved.", "reference_solution": "1. Start at the route/controller.\n2. Follow service layer calls.\n3. Check repository and schema modules.\n4. Verify tests that cover the flow.", "evaluation": {"checks": ["correct entry point", "correct data flow", "relevant files identified"]}, "tags": ["agentic", "coding", "repo_understanding", "python", "infra", "persian-friendly"]}
{"id": "JT3K-001417", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "javascript", "difficulty": "easy", "title": "Review a memoized component", "domain": "fintech", "user_story": "An over-memoized component hides actual prop changes.", "repo_context": "Repo: ml-pipeline. Domain: fintech. Frontend app with shared utilities and API wrappers.", "instruction": "Review the implementation for: Review a memoized component. Comment on correctness, UX, and maintainability.", "reference_solution": "Look for async race conditions, unnecessary re-renders, error handling gaps, and accessibility issues.", "evaluation": {"checks": ["quality issues found", "practical suggestions", "frontend concerns"]}, "tags": ["agentic", "coding", "code_review", "javascript", "fintech", "persian-friendly"]}
{"id": "JT3K-001836", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "javascript", "difficulty": "medium", "title": "Design client methods", "domain": "healthtech", "user_story": "Design a TypeScript client for search, pagination, and retries.", "repo_context": "Repo: mobile-sdk. Domain: healthtech. TypeScript client or Node API layer.", "instruction": "Design a TypeScript API for: Design client methods. Define method names, inputs, and return types.", "reference_solution": "Prefer explicit types, typed errors, pagination support, and composable options objects.", "evaluation": {"checks": ["typed contract", "clear ergonomics", "extensibility"]}, "tags": ["agentic", "coding", "api_design", "javascript", "healthtech", "persian-friendly"]}
{"id": "JT3K-000045", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "easy", "title": "Prevent duplicate inserts", "domain": "saas", "user_story": "A repository method can insert duplicate rows under retries. Add idempotency guards.", "repo_context": "Repo: ecommerce-backend. Domain: saas. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Prevent duplicate inserts. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "saas", "persian-friendly"]}
{"id": "JT3K-002552", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "javascript", "difficulty": "medium", "title": "Extract reusable hook", "domain": "infra", "user_story": "Refactor repeated fetch logic into a custom React hook.", "repo_context": "Repo: dashboard-web. Domain: infra. React components and shared logic in a frontend repo.", "instruction": "Refactor the following: Extract reusable hook. Reduce duplication and keep the public interface stable.", "reference_solution": "Extract a hook or utility, preserve props and return shape, and add tests around behavior.", "evaluation": {"checks": ["duplication removed", "interface stable", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "javascript", "infra", "persian-friendly"]}
{"id": "JT3K-000362", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "python", "difficulty": "medium", "title": "Define response shapes", "domain": "saas", "user_story": "Design consistent API responses for success, validation errors, and server errors.", "repo_context": "Repo: cms-platform. Domain: saas. Python backend exposing JSON APIs.", "instruction": "Design the API/contract for: Define response shapes. Define inputs, outputs, errors, and extensibility points.", "reference_solution": "Use explicit request/response schemas, stable error shapes, and version-friendly naming.", "evaluation": {"checks": ["clear contract", "consistent errors", "future-proof design"]}, "tags": ["agentic", "coding", "api_design", "python", "saas", "persian-friendly"]}
{"id": "JT3K-001022", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "python", "difficulty": "easy", "title": "Split a large function", "domain": "developer-tools", "user_story": "Refactor a 120-line data transformation function into small testable helpers.", "repo_context": "Repo: dashboard-web. Domain: developer-tools. Python codebase with repeated helpers and service objects.", "instruction": "Refactor the following situation: Split a large function. Preserve behavior and reduce complexity.", "reference_solution": "Extract helpers, keep public API stable, add/adjust tests, and avoid mixing business logic with I/O.", "evaluation": {"checks": ["behavior preserved", "complexity reduced", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "python", "developer-tools", "persian-friendly"]}
{"id": "JT3K-002323", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "python", "difficulty": "hard", "title": "Explain a stack trace", "domain": "developer-tools", "user_story": "A KeyError appears in production when a field is absent in a nested payload.", "repo_context": "Repo: ml-pipeline. Domain: developer-tools. Python service with logs, stack traces, and tests.", "instruction": "Diagnose this failure: Explain a stack trace. Provide the most likely cause and the verification steps.", "reference_solution": "Read the stack trace, identify the failing boundary, isolate shared state or missing key paths, then add a regression test.", "evaluation": {"checks": ["root cause", "debug steps", "verification plan"]}, "tags": ["agentic", "coding", "debug_trace", "python", "developer-tools", "persian-friendly"]}
{"id": "JT3K-001547", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "typescript", "difficulty": "easy", "title": "Plan verify-edit cycle", "domain": "developer-tools", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: mobile-sdk. Domain: developer-tools. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "typescript", "developer-tools", "persian-friendly"]}
{"id": "JT3K-000679", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "sql", "difficulty": "hard", "title": "Write query tests", "domain": "infra", "user_story": "Create validation cases for a revenue query with month boundaries.", "repo_context": "Repo: chat-agent. Domain: infra. SQL validation and BI dashboards.", "instruction": "Define tests/validation cases for: Write query tests. Include edge and boundary conditions.", "reference_solution": "Use small fixture tables and validate row counts, null handling, and date windows.", "evaluation": {"checks": ["fixtures", "edge cases", "row-level expectations"]}, "tags": ["agentic", "coding", "tdd", "sql", "infra", "persian-friendly"]}
{"id": "JT3K-001926", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "javascript", "difficulty": "easy", "title": "Plan verify-edit cycle", "domain": "edtech", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: dashboard-web. Domain: edtech. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "javascript", "edtech", "persian-friendly"]}
{"id": "JT3K-000326", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "javascript", "difficulty": "easy", "title": "Design client methods", "domain": "saas", "user_story": "Design a TypeScript client for search, pagination, and retries.", "repo_context": "Repo: dashboard-web. Domain: saas. TypeScript client or Node API layer.", "instruction": "Design a TypeScript API for: Design client methods. Define method names, inputs, and return types.", "reference_solution": "Prefer explicit types, typed errors, pagination support, and composable options objects.", "evaluation": {"checks": ["typed contract", "clear ergonomics", "extensibility"]}, "tags": ["agentic", "coding", "api_design", "javascript", "saas", "persian-friendly"]}
{"id": "JT3K-002607", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "python", "difficulty": "hard", "title": "Review a serializer", "domain": "fintech", "user_story": "The serializer leaks internal IDs in API responses; propose a cleaner contract.", "repo_context": "Repo: dashboard-web. Domain: fintech. Python backend with API handlers and background workers.", "instruction": "Review the code for: Review a serializer. Call out risks and give a practical improvement plan.", "reference_solution": "Flag correctness, performance, maintainability, and security concerns. Recommend the smallest safe refactor.", "evaluation": {"checks": ["risk identified", "actionable fix", "production awareness"]}, "tags": ["agentic", "coding", "code_review", "python", "fintech", "persian-friendly"]}
{"id": "JT3K-002303", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "bash", "difficulty": "easy", "title": "Plan release verification", "domain": "fintech", "user_story": "Run lint, unit tests, packaging, and smoke checks before publishing.", "repo_context": "Repo: chat-agent. Domain: fintech. Release and verification scripts in bash.", "instruction": "Build a tool-use plan for: Plan release verification. Include commands to inspect, run, and validate.", "reference_solution": "List the exact commands in order: inspect files, run targeted checks, then package and smoke test.", "evaluation": {"checks": ["command order", "verification", "safe execution"]}, "tags": ["agentic", "coding", "tool_use", "bash", "fintech", "persian-friendly"]}
{"id": "JT3K-000764", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "typescript", "difficulty": "medium", "title": "Diagnose promise flow", "domain": "ecommerce", "user_story": "A promise chain swallows the original error and returns a misleading success.", "repo_context": "Repo: ecommerce-backend. Domain: ecommerce. Frontend app with hydration and async data flow.", "instruction": "Diagnose the issue: Diagnose promise flow. Explain likely cause and how to verify the fix.", "reference_solution": "Check server/client output mismatches, stale closures, promise handling, and state synchronization.", "evaluation": {"checks": ["root cause", "verification steps", "frontend-specific diagnosis"]}, "tags": ["agentic", "coding", "debug_trace", "typescript", "ecommerce", "persian-friendly"]}
{"id": "JT3K-001465", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "sql", "difficulty": "medium", "title": "Plan migration checks", "domain": "edtech", "user_story": "Inspect tables, compare counts, run a diff query, and validate migration impact.", "repo_context": "Repo: scheduler. Domain: edtech. Agent workflow over SQL files and validation queries.", "instruction": "Create a tool-use plan for: Plan migration checks. Read schema, inspect query, run checks, and validate output.", "reference_solution": "Inspect schema first, isolate the grain, run a comparison query, then verify counts after the patch.", "evaluation": {"checks": ["schema first", "comparison query", "verification"]}, "tags": ["agentic", "coding", "tool_use", "sql", "edtech", "persian-friendly"]}
{"id": "JT3K-001603", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "bash", "difficulty": "easy", "title": "Diagnose build failure", "domain": "edtech", "user_story": "A CI script fails because one command exits early and hides the root cause.", "repo_context": "Repo: dashboard-web. Domain: edtech. CI scripts, release automation, and shell helpers.", "instruction": "Diagnose this shell failure: Diagnose build failure. Explain the failure point and the safest fix.", "reference_solution": "Check exit codes, quote variables, and make failure handling explicit with set -euo pipefail when appropriate.", "evaluation": {"checks": ["failure point", "shell safety", "clear fix"]}, "tags": ["agentic", "coding", "debug_trace", "bash", "edtech", "persian-friendly"]}
{"id": "JT3K-001579", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "sql", "difficulty": "medium", "title": "Plan migration checks", "domain": "healthtech", "user_story": "Inspect tables, compare counts, run a diff query, and validate migration impact.", "repo_context": "Repo: mobile-sdk. Domain: healthtech. Agent workflow over SQL files and validation queries.", "instruction": "Create a tool-use plan for: Plan migration checks. Read schema, inspect query, run checks, and validate output.", "reference_solution": "Inspect schema first, isolate the grain, run a comparison query, then verify counts after the patch.", "evaluation": {"checks": ["schema first", "comparison query", "verification"]}, "tags": ["agentic", "coding", "tool_use", "sql", "healthtech", "persian-friendly"]}
{"id": "JT3K-000963", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "medium", "title": "Handle empty payloads", "domain": "developer-tools", "user_story": "A webhook handler crashes when the JSON body is empty. Make it resilient and return a proper error.", "repo_context": "Repo: auth-service. Domain: developer-tools. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Handle empty payloads. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "developer-tools", "persian-friendly"]}
{"id": "JT3K-002539", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "sql", "difficulty": "medium", "title": "Fix duplicate aggregation", "domain": "saas", "user_story": "A query double-counts rows after joining two one-to-many tables.", "repo_context": "Repo: ecommerce-backend. Domain: saas. SQL analytics pipeline with reporting queries.", "instruction": "Fix the query issue: Fix duplicate aggregation. Explain the bug and provide the corrected pattern.", "reference_solution": "Deduplicate before aggregating, use correct grain, and be careful with time boundaries.", "evaluation": {"checks": ["correct grain", "no double count", "boundary handling"]}, "tags": ["agentic", "coding", "bug_fix", "sql", "saas", "persian-friendly"]}
{"id": "JT3K-000882", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "python", "difficulty": "medium", "title": "Write tests before implementation", "domain": "developer-tools", "user_story": "Create unit tests for a discount calculator covering edge cases and rounding.", "repo_context": "Repo: scheduler. Domain: developer-tools. Python package with pytest and typed functions.", "instruction": "Write a test-first plan for: Write tests before implementation. Include the key test cases before implementation.", "reference_solution": "Cover happy path, invalid input, boundaries, and failure modes. Keep tests deterministic and small.", "evaluation": {"checks": ["edge cases covered", "test names clear", "deterministic behavior"]}, "tags": ["agentic", "coding", "tdd", "python", "developer-tools", "persian-friendly"]}
{"id": "JT3K-001358", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "typescript", "difficulty": "hard", "title": "Diagnose promise flow", "domain": "edtech", "user_story": "A promise chain swallows the original error and returns a misleading success.", "repo_context": "Repo: analytics-core. Domain: edtech. Frontend app with hydration and async data flow.", "instruction": "Diagnose the issue: Diagnose promise flow. Explain likely cause and how to verify the fix.", "reference_solution": "Check server/client output mismatches, stale closures, promise handling, and state synchronization.", "evaluation": {"checks": ["root cause", "verification steps", "frontend-specific diagnosis"]}, "tags": ["agentic", "coding", "debug_trace", "typescript", "edtech", "persian-friendly"]}
{"id": "JT3K-002675", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "javascript", "difficulty": "medium", "title": "Design client methods", "domain": "infra", "user_story": "Design a TypeScript client for search, pagination, and retries.", "repo_context": "Repo: analytics-core. Domain: infra. TypeScript client or Node API layer.", "instruction": "Design a TypeScript API for: Design client methods. Define method names, inputs, and return types.", "reference_solution": "Prefer explicit types, typed errors, pagination support, and composable options objects.", "evaluation": {"checks": ["typed contract", "clear ergonomics", "extensibility"]}, "tags": ["agentic", "coding", "api_design", "javascript", "infra", "persian-friendly"]}
{"id": "JT3K-001395", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "python", "difficulty": "medium", "title": "Map the flow of a request through the service", "domain": "healthtech", "user_story": "Given a Python Flask service with routes, services, and a DB layer, explain the request path for creating an order.", "repo_context": "Repo: auth-service. Domain: healthtech. Python service with app/, services/, repositories/, tests/.", "instruction": "Explain how this codebase handles the task: Map the flow of a request through the service. Identify the files and modules most likely involved.", "reference_solution": "1. Start at the route/controller.\n2. Follow service layer calls.\n3. Check repository and schema modules.\n4. Verify tests that cover the flow.", "evaluation": {"checks": ["correct entry point", "correct data flow", "relevant files identified"]}, "tags": ["agentic", "coding", "repo_understanding", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-000019", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "sql", "difficulty": "easy", "title": "Design audit tables", "domain": "developer-tools", "user_story": "Design tables for tracking API events and change history.", "repo_context": "Repo: auth-service. Domain: developer-tools. Analytics schema design.", "instruction": "Design the tables or schema for: Design audit tables. Define keys, partitions, and audit columns.", "reference_solution": "Use surrogate keys, timestamps, source metadata, and clear fact/dimension separation.", "evaluation": {"checks": ["schema clarity", "scalability", "auditability"]}, "tags": ["agentic", "coding", "api_design", "sql", "developer-tools", "persian-friendly"]}
{"id": "JT3K-002430", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "python", "difficulty": "hard", "title": "Define response shapes", "domain": "healthtech", "user_story": "Design consistent API responses for success, validation errors, and server errors.", "repo_context": "Repo: ecommerce-backend. Domain: healthtech. Python backend exposing JSON APIs.", "instruction": "Design the API/contract for: Define response shapes. Define inputs, outputs, errors, and extensibility points.", "reference_solution": "Use explicit request/response schemas, stable error shapes, and version-friendly naming.", "evaluation": {"checks": ["clear contract", "consistent errors", "future-proof design"]}, "tags": ["agentic", "coding", "api_design", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-000549", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "medium", "title": "Handle empty payloads", "domain": "edtech", "user_story": "A webhook handler crashes when the JSON body is empty. Make it resilient and return a proper error.", "repo_context": "Repo: dashboard-web. Domain: edtech. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Handle empty payloads. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "edtech", "persian-friendly"]}
{"id": "JT3K-002435", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "typescript", "difficulty": "easy", "title": "Trace state across components", "domain": "developer-tools", "user_story": "A React app manages auth state in context, hooks, and a small API layer.", "repo_context": "Repo: ecommerce-backend. Domain: developer-tools. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Trace state across components. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "typescript", "developer-tools", "persian-friendly"]}
{"id": "JT3K-001013", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "javascript", "difficulty": "medium", "title": "Test a formatter", "domain": "ecommerce", "user_story": "Write tests for a date formatter that must respect locale and fallback rules.", "repo_context": "Repo: ecommerce-backend. Domain: ecommerce. JS testing stack with Vitest/Jest and React Testing Library.", "instruction": "Write test cases for: Test a formatter. Start with the edge cases and expected behavior.", "reference_solution": "Cover input validation, boundary values, async states, and error rendering.", "evaluation": {"checks": ["edge coverage", "async behavior", "clear expectations"]}, "tags": ["agentic", "coding", "tdd", "javascript", "ecommerce", "persian-friendly"]}
{"id": "JT3K-001913", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "javascript", "difficulty": "medium", "title": "Diagnose promise flow", "domain": "developer-tools", "user_story": "A promise chain swallows the original error and returns a misleading success.", "repo_context": "Repo: ecommerce-backend. Domain: developer-tools. Frontend app with hydration and async data flow.", "instruction": "Diagnose the issue: Diagnose promise flow. Explain likely cause and how to verify the fix.", "reference_solution": "Check server/client output mismatches, stale closures, promise handling, and state synchronization.", "evaluation": {"checks": ["root cause", "verification steps", "frontend-specific diagnosis"]}, "tags": ["agentic", "coding", "debug_trace", "javascript", "developer-tools", "persian-friendly"]}
{"id": "JT3K-001407", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "python", "difficulty": "easy", "title": "Sequence tool calls", "domain": "edtech", "user_story": "Construct a tool-use plan for fixing an import error across multiple files.", "repo_context": "Repo: cms-platform. Domain: edtech. Agentic coding workflow with file reads, tests, and patches.", "instruction": "Create a tool-use plan for: Sequence tool calls. Include the ideal sequence of read, edit, test, and verify actions.", "reference_solution": "Read the relevant file set, inspect failure output, patch the narrowest surface, rerun targeted tests, then broaden checks.", "evaluation": {"checks": ["tool order sensible", "minimal edit strategy", "verification included"]}, "tags": ["agentic", "coding", "tool_use", "python", "edtech", "persian-friendly"]}
{"id": "JT3K-001935", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "typescript", "difficulty": "hard", "title": "Design client methods", "domain": "healthtech", "user_story": "Design a TypeScript client for search, pagination, and retries.", "repo_context": "Repo: payments-api. Domain: healthtech. TypeScript client or Node API layer.", "instruction": "Design a TypeScript API for: Design client methods. Define method names, inputs, and return types.", "reference_solution": "Prefer explicit types, typed errors, pagination support, and composable options objects.", "evaluation": {"checks": ["typed contract", "clear ergonomics", "extensibility"]}, "tags": ["agentic", "coding", "api_design", "typescript", "healthtech", "persian-friendly"]}
{"id": "JT3K-000578", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "typescript", "difficulty": "medium", "title": "Design client methods", "domain": "ecommerce", "user_story": "Design a TypeScript client for search, pagination, and retries.", "repo_context": "Repo: ml-pipeline. Domain: ecommerce. TypeScript client or Node API layer.", "instruction": "Design a TypeScript API for: Design client methods. Define method names, inputs, and return types.", "reference_solution": "Prefer explicit types, typed errors, pagination support, and composable options objects.", "evaluation": {"checks": ["typed contract", "clear ergonomics", "extensibility"]}, "tags": ["agentic", "coding", "api_design", "typescript", "ecommerce", "persian-friendly"]}
{"id": "JT3K-002979", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "sql", "difficulty": "easy", "title": "Fix duplicate aggregation", "domain": "edtech", "user_story": "A query double-counts rows after joining two one-to-many tables.", "repo_context": "Repo: payments-api. Domain: edtech. SQL analytics pipeline with reporting queries.", "instruction": "Fix the query issue: Fix duplicate aggregation. Explain the bug and provide the corrected pattern.", "reference_solution": "Deduplicate before aggregating, use correct grain, and be careful with time boundaries.", "evaluation": {"checks": ["correct grain", "no double count", "boundary handling"]}, "tags": ["agentic", "coding", "bug_fix", "sql", "edtech", "persian-friendly"]}
{"id": "JT3K-000695", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "python", "difficulty": "easy", "title": "Reduce duplication", "domain": "healthtech", "user_story": "The repo has repeated validation logic in three endpoints; centralize it.", "repo_context": "Repo: cms-platform. Domain: healthtech. Python codebase with repeated helpers and service objects.", "instruction": "Refactor the following situation: Reduce duplication. Preserve behavior and reduce complexity.", "reference_solution": "Extract helpers, keep public API stable, add/adjust tests, and avoid mixing business logic with I/O.", "evaluation": {"checks": ["behavior preserved", "complexity reduced", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-002892", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "python", "difficulty": "hard", "title": "Trace a failing unit test", "domain": "ecommerce", "user_story": "A test intermittently fails due to shared mutable state between cases.", "repo_context": "Repo: ml-pipeline. Domain: ecommerce. Python service with logs, stack traces, and tests.", "instruction": "Diagnose this failure: Trace a failing unit test. Provide the most likely cause and the verification steps.", "reference_solution": "Read the stack trace, identify the failing boundary, isolate shared state or missing key paths, then add a regression test.", "evaluation": {"checks": ["root cause", "debug steps", "verification plan"]}, "tags": ["agentic", "coding", "debug_trace", "python", "ecommerce", "persian-friendly"]}
{"id": "JT3K-001324", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "bash", "difficulty": "medium", "title": "Plan release verification", "domain": "ecommerce", "user_story": "Run lint, unit tests, packaging, and smoke checks before publishing.", "repo_context": "Repo: dashboard-web. Domain: ecommerce. Release and verification scripts in bash.", "instruction": "Build a tool-use plan for: Plan release verification. Include commands to inspect, run, and validate.", "reference_solution": "List the exact commands in order: inspect files, run targeted checks, then package and smoke test.", "evaluation": {"checks": ["command order", "verification", "safe execution"]}, "tags": ["agentic", "coding", "tool_use", "bash", "ecommerce", "persian-friendly"]}
{"id": "JT3K-002145", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "python", "difficulty": "easy", "title": "Review an async worker", "domain": "fintech", "user_story": "An async task uses blocking I/O in the event loop; suggest a safer pattern.", "repo_context": "Repo: chat-agent. Domain: fintech. Python backend with API handlers and background workers.", "instruction": "Review the code for: Review an async worker. Call out risks and give a practical improvement plan.", "reference_solution": "Flag correctness, performance, maintainability, and security concerns. Recommend the smallest safe refactor.", "evaluation": {"checks": ["risk identified", "actionable fix", "production awareness"]}, "tags": ["agentic", "coding", "code_review", "python", "fintech", "persian-friendly"]}
{"id": "JT3K-002994", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "easy", "title": "Handle empty payloads", "domain": "healthtech", "user_story": "A webhook handler crashes when the JSON body is empty. Make it resilient and return a proper error.", "repo_context": "Repo: chat-agent. Domain: healthtech. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Handle empty payloads. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-000885", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "bash", "difficulty": "medium", "title": "Simplify script logic", "domain": "fintech", "user_story": "Refactor a long shell script into functions with clearer error handling.", "repo_context": "Repo: scheduler. Domain: fintech. Bash automation with functions and environment variables.", "instruction": "Refactor the script for: Simplify script logic. Reduce repetition and improve error handling.", "reference_solution": "Extract functions, use strict mode carefully, and keep reusable command pieces in variables.", "evaluation": {"checks": ["repetition reduced", "error handling", "maintainability"]}, "tags": ["agentic", "coding", "refactor", "bash", "fintech", "persian-friendly"]}
{"id": "JT3K-000525", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "sql", "difficulty": "medium", "title": "Review a reporting query", "domain": "ecommerce", "user_story": "The query is correct but slow due to an unbounded scan and unnecessary subquery.", "repo_context": "Repo: analytics-core. Domain: ecommerce. SQL reporting and ETL jobs.", "instruction": "Review the SQL for: Review a reporting query. Identify correctness and performance risks.", "reference_solution": "Flag full scans, unnecessary CTEs, non-sargable filters, and inconsistent joins.", "evaluation": {"checks": ["performance risk", "logic risk", "clear recommendation"]}, "tags": ["agentic", "coding", "code_review", "sql", "ecommerce", "persian-friendly"]}
{"id": "JT3K-001215", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "typescript", "difficulty": "easy", "title": "Test form validation", "domain": "creator-tools", "user_story": "Create tests for password and email validation in a UI flow.", "repo_context": "Repo: mobile-sdk. Domain: creator-tools. JS testing stack with Vitest/Jest and React Testing Library.", "instruction": "Write test cases for: Test form validation. Start with the edge cases and expected behavior.", "reference_solution": "Cover input validation, boundary values, async states, and error rendering.", "evaluation": {"checks": ["edge coverage", "async behavior", "clear expectations"]}, "tags": ["agentic", "coding", "tdd", "typescript", "creator-tools", "persian-friendly"]}
{"id": "JT3K-002990", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "typescript", "difficulty": "medium", "title": "Diagnose promise flow", "domain": "developer-tools", "user_story": "A promise chain swallows the original error and returns a misleading success.", "repo_context": "Repo: cms-platform. Domain: developer-tools. Frontend app with hydration and async data flow.", "instruction": "Diagnose the issue: Diagnose promise flow. Explain likely cause and how to verify the fix.", "reference_solution": "Check server/client output mismatches, stale closures, promise handling, and state synchronization.", "evaluation": {"checks": ["root cause", "verification steps", "frontend-specific diagnosis"]}, "tags": ["agentic", "coding", "debug_trace", "typescript", "developer-tools", "persian-friendly"]}
{"id": "JT3K-001980", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "sql", "difficulty": "easy", "title": "Fix duplicate aggregation", "domain": "saas", "user_story": "A query double-counts rows after joining two one-to-many tables.", "repo_context": "Repo: payments-api. Domain: saas. SQL analytics pipeline with reporting queries.", "instruction": "Fix the query issue: Fix duplicate aggregation. Explain the bug and provide the corrected pattern.", "reference_solution": "Deduplicate before aggregating, use correct grain, and be careful with time boundaries.", "evaluation": {"checks": ["correct grain", "no double count", "boundary handling"]}, "tags": ["agentic", "coding", "bug_fix", "sql", "saas", "persian-friendly"]}
{"id": "JT3K-002824", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "bash", "difficulty": "medium", "title": "Simplify script logic", "domain": "saas", "user_story": "Refactor a long shell script into functions with clearer error handling.", "repo_context": "Repo: dashboard-web. Domain: saas. Bash automation with functions and environment variables.", "instruction": "Refactor the script for: Simplify script logic. Reduce repetition and improve error handling.", "reference_solution": "Extract functions, use strict mode carefully, and keep reusable command pieces in variables.", "evaluation": {"checks": ["repetition reduced", "error handling", "maintainability"]}, "tags": ["agentic", "coding", "refactor", "bash", "saas", "persian-friendly"]}
{"id": "JT3K-002040", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "python", "difficulty": "easy", "title": "Trace a failing unit test", "domain": "ecommerce", "user_story": "A test intermittently fails due to shared mutable state between cases.", "repo_context": "Repo: ml-pipeline. Domain: ecommerce. Python service with logs, stack traces, and tests.", "instruction": "Diagnose this failure: Trace a failing unit test. Provide the most likely cause and the verification steps.", "reference_solution": "Read the stack trace, identify the failing boundary, isolate shared state or missing key paths, then add a regression test.", "evaluation": {"checks": ["root cause", "debug steps", "verification plan"]}, "tags": ["agentic", "coding", "debug_trace", "python", "ecommerce", "persian-friendly"]}
{"id": "JT3K-001404", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "python", "difficulty": "medium", "title": "Review a cache helper", "domain": "creator-tools", "user_story": "A cache wrapper forgets to namespace keys and can collide across tenants.", "repo_context": "Repo: analytics-core. Domain: creator-tools. Python backend with API handlers and background workers.", "instruction": "Review the code for: Review a cache helper. Call out risks and give a practical improvement plan.", "reference_solution": "Flag correctness, performance, maintainability, and security concerns. Recommend the smallest safe refactor.", "evaluation": {"checks": ["risk identified", "actionable fix", "production awareness"]}, "tags": ["agentic", "coding", "code_review", "python", "creator-tools", "persian-friendly"]}
{"id": "JT3K-000719", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "sql", "difficulty": "easy", "title": "Fix duplicate aggregation", "domain": "edtech", "user_story": "A query double-counts rows after joining two one-to-many tables.", "repo_context": "Repo: ecommerce-backend. Domain: edtech. SQL analytics pipeline with reporting queries.", "instruction": "Fix the query issue: Fix duplicate aggregation. Explain the bug and provide the corrected pattern.", "reference_solution": "Deduplicate before aggregating, use correct grain, and be careful with time boundaries.", "evaluation": {"checks": ["correct grain", "no double count", "boundary handling"]}, "tags": ["agentic", "coding", "bug_fix", "sql", "edtech", "persian-friendly"]}
{"id": "JT3K-002067", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "sql", "difficulty": "medium", "title": "Review a reporting query", "domain": "saas", "user_story": "The query is correct but slow due to an unbounded scan and unnecessary subquery.", "repo_context": "Repo: scheduler. Domain: saas. SQL reporting and ETL jobs.", "instruction": "Review the SQL for: Review a reporting query. Identify correctness and performance risks.", "reference_solution": "Flag full scans, unnecessary CTEs, non-sargable filters, and inconsistent joins.", "evaluation": {"checks": ["performance risk", "logic risk", "clear recommendation"]}, "tags": ["agentic", "coding", "code_review", "sql", "saas", "persian-friendly"]}
{"id": "JT3K-000889", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "python", "difficulty": "hard", "title": "Identify the file to change for feature X", "domain": "healthtech", "user_story": "A FastAPI repo stores schema logic in one module and persistence in another. Locate where to add a new field safely.", "repo_context": "Repo: ecommerce-backend. Domain: healthtech. Python service with app/, services/, repositories/, tests/.", "instruction": "Explain how this codebase handles the task: Identify the file to change for feature X. Identify the files and modules most likely involved.", "reference_solution": "1. Start at the route/controller.\n2. Follow service layer calls.\n3. Check repository and schema modules.\n4. Verify tests that cover the flow.", "evaluation": {"checks": ["correct entry point", "correct data flow", "relevant files identified"]}, "tags": ["agentic", "coding", "repo_understanding", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-001435", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "typescript", "difficulty": "hard", "title": "Plan verify-edit cycle", "domain": "ecommerce", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: mobile-sdk. Domain: ecommerce. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "typescript", "ecommerce", "persian-friendly"]}
{"id": "JT3K-002991", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "sql", "difficulty": "medium", "title": "Write query tests", "domain": "creator-tools", "user_story": "Create validation cases for a revenue query with month boundaries.", "repo_context": "Repo: ml-pipeline. Domain: creator-tools. SQL validation and BI dashboards.", "instruction": "Define tests/validation cases for: Write query tests. Include edge and boundary conditions.", "reference_solution": "Use small fixture tables and validate row counts, null handling, and date windows.", "evaluation": {"checks": ["fixtures", "edge cases", "row-level expectations"]}, "tags": ["agentic", "coding", "tdd", "sql", "creator-tools", "persian-friendly"]}
{"id": "JT3K-002121", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "python", "difficulty": "easy", "title": "Define response shapes", "domain": "edtech", "user_story": "Design consistent API responses for success, validation errors, and server errors.", "repo_context": "Repo: cms-platform. Domain: edtech. Python backend exposing JSON APIs.", "instruction": "Design the API/contract for: Define response shapes. Define inputs, outputs, errors, and extensibility points.", "reference_solution": "Use explicit request/response schemas, stable error shapes, and version-friendly naming.", "evaluation": {"checks": ["clear contract", "consistent errors", "future-proof design"]}, "tags": ["agentic", "coding", "api_design", "python", "edtech", "persian-friendly"]}
{"id": "JT3K-001882", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "python", "difficulty": "easy", "title": "Trace a failing unit test", "domain": "creator-tools", "user_story": "A test intermittently fails due to shared mutable state between cases.", "repo_context": "Repo: scheduler. Domain: creator-tools. Python service with logs, stack traces, and tests.", "instruction": "Diagnose this failure: Trace a failing unit test. Provide the most likely cause and the verification steps.", "reference_solution": "Read the stack trace, identify the failing boundary, isolate shared state or missing key paths, then add a regression test.", "evaluation": {"checks": ["root cause", "debug steps", "verification plan"]}, "tags": ["agentic", "coding", "debug_trace", "python", "creator-tools", "persian-friendly"]}
{"id": "JT3K-002231", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "python", "difficulty": "medium", "title": "Define response shapes", "domain": "saas", "user_story": "Design consistent API responses for success, validation errors, and server errors.", "repo_context": "Repo: scheduler. Domain: saas. Python backend exposing JSON APIs.", "instruction": "Design the API/contract for: Define response shapes. Define inputs, outputs, errors, and extensibility points.", "reference_solution": "Use explicit request/response schemas, stable error shapes, and version-friendly naming.", "evaluation": {"checks": ["clear contract", "consistent errors", "future-proof design"]}, "tags": ["agentic", "coding", "api_design", "python", "saas", "persian-friendly"]}
{"id": "JT3K-002592", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "python", "difficulty": "easy", "title": "Explain a stack trace", "domain": "healthtech", "user_story": "A KeyError appears in production when a field is absent in a nested payload.", "repo_context": "Repo: chat-agent. Domain: healthtech. Python service with logs, stack traces, and tests.", "instruction": "Diagnose this failure: Explain a stack trace. Provide the most likely cause and the verification steps.", "reference_solution": "Read the stack trace, identify the failing boundary, isolate shared state or missing key paths, then add a regression test.", "evaluation": {"checks": ["root cause", "debug steps", "verification plan"]}, "tags": ["agentic", "coding", "debug_trace", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-000974", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "bash", "difficulty": "easy", "title": "Simplify script logic", "domain": "healthtech", "user_story": "Refactor a long shell script into functions with clearer error handling.", "repo_context": "Repo: chat-agent. Domain: healthtech. Bash automation with functions and environment variables.", "instruction": "Refactor the script for: Simplify script logic. Reduce repetition and improve error handling.", "reference_solution": "Extract functions, use strict mode carefully, and keep reusable command pieces in variables.", "evaluation": {"checks": ["repetition reduced", "error handling", "maintainability"]}, "tags": ["agentic", "coding", "refactor", "bash", "healthtech", "persian-friendly"]}
{"id": "JT3K-002901", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "javascript", "difficulty": "medium", "title": "Locate data fetching boundaries", "domain": "edtech", "user_story": "A Next.js repo separates server actions, client components, and shared utilities.", "repo_context": "Repo: dashboard-web. Domain: edtech. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Locate data fetching boundaries. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "javascript", "edtech", "persian-friendly"]}
{"id": "JT3K-001449", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "python", "difficulty": "medium", "title": "Trace a failing unit test", "domain": "creator-tools", "user_story": "A test intermittently fails due to shared mutable state between cases.", "repo_context": "Repo: scheduler. Domain: creator-tools. Python service with logs, stack traces, and tests.", "instruction": "Diagnose this failure: Trace a failing unit test. Provide the most likely cause and the verification steps.", "reference_solution": "Read the stack trace, identify the failing boundary, isolate shared state or missing key paths, then add a regression test.", "evaluation": {"checks": ["root cause", "debug steps", "verification plan"]}, "tags": ["agentic", "coding", "debug_trace", "python", "creator-tools", "persian-friendly"]}
{"id": "JT3K-000233", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "javascript", "difficulty": "medium", "title": "Locate data fetching boundaries", "domain": "ecommerce", "user_story": "A Next.js repo separates server actions, client components, and shared utilities.", "repo_context": "Repo: payments-api. Domain: ecommerce. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Locate data fetching boundaries. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "javascript", "ecommerce", "persian-friendly"]}
{"id": "JT3K-002210", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "sql", "difficulty": "medium", "title": "Design audit tables", "domain": "saas", "user_story": "Design tables for tracking API events and change history.", "repo_context": "Repo: dashboard-web. Domain: saas. Analytics schema design.", "instruction": "Design the tables or schema for: Design audit tables. Define keys, partitions, and audit columns.", "reference_solution": "Use surrogate keys, timestamps, source metadata, and clear fact/dimension separation.", "evaluation": {"checks": ["schema clarity", "scalability", "auditability"]}, "tags": ["agentic", "coding", "api_design", "sql", "saas", "persian-friendly"]}
{"id": "JT3K-002799", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "python", "difficulty": "easy", "title": "Improve readability", "domain": "edtech", "user_story": "Rename ambiguous variables and simplify nested conditionals in a config loader.", "repo_context": "Repo: ml-pipeline. Domain: edtech. Python codebase with repeated helpers and service objects.", "instruction": "Refactor the following situation: Improve readability. Preserve behavior and reduce complexity.", "reference_solution": "Extract helpers, keep public API stable, add/adjust tests, and avoid mixing business logic with I/O.", "evaluation": {"checks": ["behavior preserved", "complexity reduced", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "python", "edtech", "persian-friendly"]}
{"id": "JT3K-001783", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "python", "difficulty": "easy", "title": "Write tests before implementation", "domain": "edtech", "user_story": "Create unit tests for a discount calculator covering edge cases and rounding.", "repo_context": "Repo: cms-platform. Domain: edtech. Python package with pytest and typed functions.", "instruction": "Write a test-first plan for: Write tests before implementation. Include the key test cases before implementation.", "reference_solution": "Cover happy path, invalid input, boundaries, and failure modes. Keep tests deterministic and small.", "evaluation": {"checks": ["edge cases covered", "test names clear", "deterministic behavior"]}, "tags": ["agentic", "coding", "tdd", "python", "edtech", "persian-friendly"]}
{"id": "JT3K-000169", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "bash", "difficulty": "easy", "title": "Simplify script logic", "domain": "developer-tools", "user_story": "Refactor a long shell script into functions with clearer error handling.", "repo_context": "Repo: chat-agent. Domain: developer-tools. Bash automation with functions and environment variables.", "instruction": "Refactor the script for: Simplify script logic. Reduce repetition and improve error handling.", "reference_solution": "Extract functions, use strict mode carefully, and keep reusable command pieces in variables.", "evaluation": {"checks": ["repetition reduced", "error handling", "maintainability"]}, "tags": ["agentic", "coding", "refactor", "bash", "developer-tools", "persian-friendly"]}
{"id": "JT3K-001130", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "easy", "title": "Handle empty payloads", "domain": "ecommerce", "user_story": "A webhook handler crashes when the JSON body is empty. Make it resilient and return a proper error.", "repo_context": "Repo: ecommerce-backend. Domain: ecommerce. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Handle empty payloads. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "ecommerce", "persian-friendly"]}
{"id": "JT3K-002048", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "typescript", "difficulty": "easy", "title": "Test a formatter", "domain": "healthtech", "user_story": "Write tests for a date formatter that must respect locale and fallback rules.", "repo_context": "Repo: dashboard-web. Domain: healthtech. JS testing stack with Vitest/Jest and React Testing Library.", "instruction": "Write test cases for: Test a formatter. Start with the edge cases and expected behavior.", "reference_solution": "Cover input validation, boundary values, async states, and error rendering.", "evaluation": {"checks": ["edge coverage", "async behavior", "clear expectations"]}, "tags": ["agentic", "coding", "tdd", "typescript", "healthtech", "persian-friendly"]}
{"id": "JT3K-002773", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "javascript", "difficulty": "hard", "title": "Trace state across components", "domain": "creator-tools", "user_story": "A React app manages auth state in context, hooks, and a small API layer.", "repo_context": "Repo: dashboard-web. Domain: creator-tools. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Trace state across components. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "javascript", "creator-tools", "persian-friendly"]}
{"id": "JT3K-000648", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "typescript", "difficulty": "easy", "title": "Diagnose promise flow", "domain": "saas", "user_story": "A promise chain swallows the original error and returns a misleading success.", "repo_context": "Repo: analytics-core. Domain: saas. Frontend app with hydration and async data flow.", "instruction": "Diagnose the issue: Diagnose promise flow. Explain likely cause and how to verify the fix.", "reference_solution": "Check server/client output mismatches, stale closures, promise handling, and state synchronization.", "evaluation": {"checks": ["root cause", "verification steps", "frontend-specific diagnosis"]}, "tags": ["agentic", "coding", "debug_trace", "typescript", "saas", "persian-friendly"]}
{"id": "JT3K-000015", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "typescript", "difficulty": "medium", "title": "Fix stale closure", "domain": "saas", "user_story": "A hook callback reads old state after rapid updates; patch it.", "repo_context": "Repo: ecommerce-backend. Domain: saas. React/Next.js application with hooks and server/client boundaries.", "instruction": "Fix this bug: Fix stale closure. Keep the change minimal and safe.", "reference_solution": "Guard missing values, avoid stale state, and keep server/client-only logic in the right layer.", "evaluation": {"checks": ["bug fixed", "safe boundary handling", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "typescript", "saas", "persian-friendly"]}
{"id": "JT3K-002232", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "javascript", "difficulty": "medium", "title": "Simplify reducer logic", "domain": "creator-tools", "user_story": "Reduce branching in a Redux reducer without changing behavior.", "repo_context": "Repo: scheduler. Domain: creator-tools. React components and shared logic in a frontend repo.", "instruction": "Refactor the following: Simplify reducer logic. Reduce duplication and keep the public interface stable.", "reference_solution": "Extract a hook or utility, preserve props and return shape, and add tests around behavior.", "evaluation": {"checks": ["duplication removed", "interface stable", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "javascript", "creator-tools", "persian-friendly"]}
{"id": "JT3K-002129", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "sql", "difficulty": "medium", "title": "Simplify a report query", "domain": "fintech", "user_story": "Refactor a deeply nested CTE chain for readability.", "repo_context": "Repo: cms-platform. Domain: fintech. SQL queries with layered CTEs.", "instruction": "Refactor the query for: Simplify a report query. Make it easier to read without changing results.", "reference_solution": "Reduce nested CTE depth, standardize aliases, and keep one clear grain per step.", "evaluation": {"checks": ["readability", "stable results", "clear grain"]}, "tags": ["agentic", "coding", "refactor", "sql", "fintech", "persian-friendly"]}
{"id": "JT3K-001436", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "javascript", "difficulty": "medium", "title": "Review an API client", "domain": "saas", "user_story": "The client mixes transport errors and business errors in one catch block.", "repo_context": "Repo: ml-pipeline. Domain: saas. Frontend app with shared utilities and API wrappers.", "instruction": "Review the implementation for: Review an API client. Comment on correctness, UX, and maintainability.", "reference_solution": "Look for async race conditions, unnecessary re-renders, error handling gaps, and accessibility issues.", "evaluation": {"checks": ["quality issues found", "practical suggestions", "frontend concerns"]}, "tags": ["agentic", "coding", "code_review", "javascript", "saas", "persian-friendly"]}
{"id": "JT3K-002393", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "python", "difficulty": "easy", "title": "Map the flow of a request through the service", "domain": "creator-tools", "user_story": "Given a Python Flask service with routes, services, and a DB layer, explain the request path for creating an order.", "repo_context": "Repo: chat-agent. Domain: creator-tools. Python service with app/, services/, repositories/, tests/.", "instruction": "Explain how this codebase handles the task: Map the flow of a request through the service. Identify the files and modules most likely involved.", "reference_solution": "1. Start at the route/controller.\n2. Follow service layer calls.\n3. Check repository and schema modules.\n4. Verify tests that cover the flow.", "evaluation": {"checks": ["correct entry point", "correct data flow", "relevant files identified"]}, "tags": ["agentic", "coding", "repo_understanding", "python", "creator-tools", "persian-friendly"]}
{"id": "JT3K-002444", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "javascript", "difficulty": "easy", "title": "Plan verify-edit cycle", "domain": "creator-tools", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: payments-api. Domain: creator-tools. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "javascript", "creator-tools", "persian-friendly"]}
{"id": "JT3K-001811", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "typescript", "difficulty": "easy", "title": "Design client methods", "domain": "saas", "user_story": "Design a TypeScript client for search, pagination, and retries.", "repo_context": "Repo: cms-platform. Domain: saas. TypeScript client or Node API layer.", "instruction": "Design a TypeScript API for: Design client methods. Define method names, inputs, and return types.", "reference_solution": "Prefer explicit types, typed errors, pagination support, and composable options objects.", "evaluation": {"checks": ["typed contract", "clear ergonomics", "extensibility"]}, "tags": ["agentic", "coding", "api_design", "typescript", "saas", "persian-friendly"]}
{"id": "JT3K-001406", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "javascript", "difficulty": "easy", "title": "Review a memoized component", "domain": "creator-tools", "user_story": "An over-memoized component hides actual prop changes.", "repo_context": "Repo: auth-service. Domain: creator-tools. Frontend app with shared utilities and API wrappers.", "instruction": "Review the implementation for: Review a memoized component. Comment on correctness, UX, and maintainability.", "reference_solution": "Look for async race conditions, unnecessary re-renders, error handling gaps, and accessibility issues.", "evaluation": {"checks": ["quality issues found", "practical suggestions", "frontend concerns"]}, "tags": ["agentic", "coding", "code_review", "javascript", "creator-tools", "persian-friendly"]}
{"id": "JT3K-000508", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "typescript", "difficulty": "easy", "title": "Fix stale closure", "domain": "ecommerce", "user_story": "A hook callback reads old state after rapid updates; patch it.", "repo_context": "Repo: ml-pipeline. Domain: ecommerce. React/Next.js application with hooks and server/client boundaries.", "instruction": "Fix this bug: Fix stale closure. Keep the change minimal and safe.", "reference_solution": "Guard missing values, avoid stale state, and keep server/client-only logic in the right layer.", "evaluation": {"checks": ["bug fixed", "safe boundary handling", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "typescript", "ecommerce", "persian-friendly"]}
{"id": "JT3K-002352", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "javascript", "difficulty": "medium", "title": "Trace state across components", "domain": "infra", "user_story": "A React app manages auth state in context, hooks, and a small API layer.", "repo_context": "Repo: mobile-sdk. Domain: infra. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Trace state across components. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "javascript", "infra", "persian-friendly"]}
{"id": "JT3K-001508", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "typescript", "difficulty": "easy", "title": "Design client methods", "domain": "infra", "user_story": "Design a TypeScript client for search, pagination, and retries.", "repo_context": "Repo: payments-api. Domain: infra. TypeScript client or Node API layer.", "instruction": "Design a TypeScript API for: Design client methods. Define method names, inputs, and return types.", "reference_solution": "Prefer explicit types, typed errors, pagination support, and composable options objects.", "evaluation": {"checks": ["typed contract", "clear ergonomics", "extensibility"]}, "tags": ["agentic", "coding", "api_design", "typescript", "infra", "persian-friendly"]}
{"id": "JT3K-001428", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "hard", "title": "Fix timezone parsing", "domain": "ecommerce", "user_story": "A datetime parser returns naive objects and breaks scheduling. Patch it to handle UTC offsets safely.", "repo_context": "Repo: scheduler. Domain: ecommerce. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Fix timezone parsing. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "ecommerce", "persian-friendly"]}
{"id": "JT3K-001721", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "python", "difficulty": "medium", "title": "Map the flow of a request through the service", "domain": "creator-tools", "user_story": "Given a Python Flask service with routes, services, and a DB layer, explain the request path for creating an order.", "repo_context": "Repo: analytics-core. Domain: creator-tools. Python service with app/, services/, repositories/, tests/.", "instruction": "Explain how this codebase handles the task: Map the flow of a request through the service. Identify the files and modules most likely involved.", "reference_solution": "1. Start at the route/controller.\n2. Follow service layer calls.\n3. Check repository and schema modules.\n4. Verify tests that cover the flow.", "evaluation": {"checks": ["correct entry point", "correct data flow", "relevant files identified"]}, "tags": ["agentic", "coding", "repo_understanding", "python", "creator-tools", "persian-friendly"]}
{"id": "JT3K-000976", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "sql", "difficulty": "hard", "title": "Review a reporting query", "domain": "fintech", "user_story": "The query is correct but slow due to an unbounded scan and unnecessary subquery.", "repo_context": "Repo: chat-agent. Domain: fintech. SQL reporting and ETL jobs.", "instruction": "Review the SQL for: Review a reporting query. Identify correctness and performance risks.", "reference_solution": "Flag full scans, unnecessary CTEs, non-sargable filters, and inconsistent joins.", "evaluation": {"checks": ["performance risk", "logic risk", "clear recommendation"]}, "tags": ["agentic", "coding", "code_review", "sql", "fintech", "persian-friendly"]}
{"id": "JT3K-001213", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "javascript", "difficulty": "medium", "title": "Extract reusable hook", "domain": "saas", "user_story": "Refactor repeated fetch logic into a custom React hook.", "repo_context": "Repo: mobile-sdk. Domain: saas. React components and shared logic in a frontend repo.", "instruction": "Refactor the following: Extract reusable hook. Reduce duplication and keep the public interface stable.", "reference_solution": "Extract a hook or utility, preserve props and return shape, and add tests around behavior.", "evaluation": {"checks": ["duplication removed", "interface stable", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "javascript", "saas", "persian-friendly"]}
{"id": "JT3K-001642", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "python", "difficulty": "easy", "title": "Design a service interface", "domain": "creator-tools", "user_story": "Propose a clean Python interface for a notification service supporting email and SMS.", "repo_context": "Repo: cms-platform. Domain: creator-tools. Python backend exposing JSON APIs.", "instruction": "Design the API/contract for: Design a service interface. Define inputs, outputs, errors, and extensibility points.", "reference_solution": "Use explicit request/response schemas, stable error shapes, and version-friendly naming.", "evaluation": {"checks": ["clear contract", "consistent errors", "future-proof design"]}, "tags": ["agentic", "coding", "api_design", "python", "creator-tools", "persian-friendly"]}
{"id": "JT3K-001726", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "javascript", "difficulty": "easy", "title": "Fix stale closure", "domain": "developer-tools", "user_story": "A hook callback reads old state after rapid updates; patch it.", "repo_context": "Repo: scheduler. Domain: developer-tools. React/Next.js application with hooks and server/client boundaries.", "instruction": "Fix this bug: Fix stale closure. Keep the change minimal and safe.", "reference_solution": "Guard missing values, avoid stale state, and keep server/client-only logic in the right layer.", "evaluation": {"checks": ["bug fixed", "safe boundary handling", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "javascript", "developer-tools", "persian-friendly"]}
{"id": "JT3K-002548", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "typescript", "difficulty": "medium", "title": "Extract reusable hook", "domain": "edtech", "user_story": "Refactor repeated fetch logic into a custom React hook.", "repo_context": "Repo: ml-pipeline. Domain: edtech. React components and shared logic in a frontend repo.", "instruction": "Refactor the following: Extract reusable hook. Reduce duplication and keep the public interface stable.", "reference_solution": "Extract a hook or utility, preserve props and return shape, and add tests around behavior.", "evaluation": {"checks": ["duplication removed", "interface stable", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "typescript", "edtech", "persian-friendly"]}
{"id": "JT3K-002567", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "sql", "difficulty": "medium", "title": "Write query tests", "domain": "infra", "user_story": "Create validation cases for a revenue query with month boundaries.", "repo_context": "Repo: analytics-core. Domain: infra. SQL validation and BI dashboards.", "instruction": "Define tests/validation cases for: Write query tests. Include edge and boundary conditions.", "reference_solution": "Use small fixture tables and validate row counts, null handling, and date windows.", "evaluation": {"checks": ["fixtures", "edge cases", "row-level expectations"]}, "tags": ["agentic", "coding", "tdd", "sql", "infra", "persian-friendly"]}
{"id": "JT3K-001247", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "typescript", "difficulty": "easy", "title": "Locate data fetching boundaries", "domain": "saas", "user_story": "A Next.js repo separates server actions, client components, and shared utilities.", "repo_context": "Repo: analytics-core. Domain: saas. React/Node codebase with components, hooks, and API clients.", "instruction": "Explain the code path for: Locate data fetching boundaries. Point out likely components, hooks, or services.", "reference_solution": "Follow the UI event, state update, API call, and render path. Identify shared utilities and tests.", "evaluation": {"checks": ["UI path identified", "state flow understood", "files mapped"]}, "tags": ["agentic", "coding", "repo_understanding", "typescript", "saas", "persian-friendly"]}
{"id": "JT3K-002136", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "javascript", "difficulty": "medium", "title": "Design client methods", "domain": "creator-tools", "user_story": "Design a TypeScript client for search, pagination, and retries.", "repo_context": "Repo: cms-platform. Domain: creator-tools. TypeScript client or Node API layer.", "instruction": "Design a TypeScript API for: Design client methods. Define method names, inputs, and return types.", "reference_solution": "Prefer explicit types, typed errors, pagination support, and composable options objects.", "evaluation": {"checks": ["typed contract", "clear ergonomics", "extensibility"]}, "tags": ["agentic", "coding", "api_design", "javascript", "creator-tools", "persian-friendly"]}
{"id": "JT3K-002081", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "typescript", "difficulty": "medium", "title": "Design client methods", "domain": "healthtech", "user_story": "Design a TypeScript client for search, pagination, and retries.", "repo_context": "Repo: scheduler. Domain: healthtech. TypeScript client or Node API layer.", "instruction": "Design a TypeScript API for: Design client methods. Define method names, inputs, and return types.", "reference_solution": "Prefer explicit types, typed errors, pagination support, and composable options objects.", "evaluation": {"checks": ["typed contract", "clear ergonomics", "extensibility"]}, "tags": ["agentic", "coding", "api_design", "typescript", "healthtech", "persian-friendly"]}
{"id": "JT3K-001262", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "hard", "title": "Handle empty payloads", "domain": "creator-tools", "user_story": "A webhook handler crashes when the JSON body is empty. Make it resilient and return a proper error.", "repo_context": "Repo: ml-pipeline. Domain: creator-tools. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Handle empty payloads. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "creator-tools", "persian-friendly"]}
{"id": "JT3K-001493", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "python", "difficulty": "easy", "title": "Review an async worker", "domain": "developer-tools", "user_story": "An async task uses blocking I/O in the event loop; suggest a safer pattern.", "repo_context": "Repo: scheduler. Domain: developer-tools. Python backend with API handlers and background workers.", "instruction": "Review the code for: Review an async worker. Call out risks and give a practical improvement plan.", "reference_solution": "Flag correctness, performance, maintainability, and security concerns. Recommend the smallest safe refactor.", "evaluation": {"checks": ["risk identified", "actionable fix", "production awareness"]}, "tags": ["agentic", "coding", "code_review", "python", "developer-tools", "persian-friendly"]}
{"id": "JT3K-002338", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "javascript", "difficulty": "easy", "title": "Extract reusable hook", "domain": "infra", "user_story": "Refactor repeated fetch logic into a custom React hook.", "repo_context": "Repo: mobile-sdk. Domain: infra. React components and shared logic in a frontend repo.", "instruction": "Refactor the following: Extract reusable hook. Reduce duplication and keep the public interface stable.", "reference_solution": "Extract a hook or utility, preserve props and return shape, and add tests around behavior.", "evaluation": {"checks": ["duplication removed", "interface stable", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "javascript", "infra", "persian-friendly"]}
{"id": "JT3K-000324", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "bash", "difficulty": "easy", "title": "Simplify script logic", "domain": "infra", "user_story": "Refactor a long shell script into functions with clearer error handling.", "repo_context": "Repo: cms-platform. Domain: infra. Bash automation with functions and environment variables.", "instruction": "Refactor the script for: Simplify script logic. Reduce repetition and improve error handling.", "reference_solution": "Extract functions, use strict mode carefully, and keep reusable command pieces in variables.", "evaluation": {"checks": ["repetition reduced", "error handling", "maintainability"]}, "tags": ["agentic", "coding", "refactor", "bash", "infra", "persian-friendly"]}
{"id": "JT3K-000386", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "python", "difficulty": "easy", "title": "Plan a repair loop", "domain": "developer-tools", "user_story": "The agent should read files, run tests, inspect failures, patch code, and re-run checks.", "repo_context": "Repo: chat-agent. Domain: developer-tools. Agentic coding workflow with file reads, tests, and patches.", "instruction": "Create a tool-use plan for: Plan a repair loop. Include the ideal sequence of read, edit, test, and verify actions.", "reference_solution": "Read the relevant file set, inspect failure output, patch the narrowest surface, rerun targeted tests, then broaden checks.", "evaluation": {"checks": ["tool order sensible", "minimal edit strategy", "verification included"]}, "tags": ["agentic", "coding", "tool_use", "python", "developer-tools", "persian-friendly"]}
{"id": "JT3K-000915", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "typescript", "difficulty": "medium", "title": "Analyze hydration mismatch", "domain": "developer-tools", "user_story": "A server-rendered page renders different timestamps on client and server.", "repo_context": "Repo: chat-agent. Domain: developer-tools. Frontend app with hydration and async data flow.", "instruction": "Diagnose the issue: Analyze hydration mismatch. Explain likely cause and how to verify the fix.", "reference_solution": "Check server/client output mismatches, stale closures, promise handling, and state synchronization.", "evaluation": {"checks": ["root cause", "verification steps", "frontend-specific diagnosis"]}, "tags": ["agentic", "coding", "debug_trace", "typescript", "developer-tools", "persian-friendly"]}
{"id": "JT3K-001413", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "javascript", "difficulty": "easy", "title": "Design client methods", "domain": "ecommerce", "user_story": "Design a TypeScript client for search, pagination, and retries.", "repo_context": "Repo: auth-service. Domain: ecommerce. TypeScript client or Node API layer.", "instruction": "Design a TypeScript API for: Design client methods. Define method names, inputs, and return types.", "reference_solution": "Prefer explicit types, typed errors, pagination support, and composable options objects.", "evaluation": {"checks": ["typed contract", "clear ergonomics", "extensibility"]}, "tags": ["agentic", "coding", "api_design", "javascript", "ecommerce", "persian-friendly"]}
{"id": "JT3K-001986", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "repo_understanding", "language": "python", "difficulty": "medium", "title": "Identify the file to change for feature X", "domain": "healthtech", "user_story": "A FastAPI repo stores schema logic in one module and persistence in another. Locate where to add a new field safely.", "repo_context": "Repo: payments-api. Domain: healthtech. Python service with app/, services/, repositories/, tests/.", "instruction": "Explain how this codebase handles the task: Identify the file to change for feature X. Identify the files and modules most likely involved.", "reference_solution": "1. Start at the route/controller.\n2. Follow service layer calls.\n3. Check repository and schema modules.\n4. Verify tests that cover the flow.", "evaluation": {"checks": ["correct entry point", "correct data flow", "relevant files identified"]}, "tags": ["agentic", "coding", "repo_understanding", "python", "healthtech", "persian-friendly"]}
{"id": "JT3K-002543", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "javascript", "difficulty": "medium", "title": "Review an API client", "domain": "saas", "user_story": "The client mixes transport errors and business errors in one catch block.", "repo_context": "Repo: analytics-core. Domain: saas. Frontend app with shared utilities and API wrappers.", "instruction": "Review the implementation for: Review an API client. Comment on correctness, UX, and maintainability.", "reference_solution": "Look for async race conditions, unnecessary re-renders, error handling gaps, and accessibility issues.", "evaluation": {"checks": ["quality issues found", "practical suggestions", "frontend concerns"]}, "tags": ["agentic", "coding", "code_review", "javascript", "saas", "persian-friendly"]}
{"id": "JT3K-002665", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "sql", "difficulty": "easy", "title": "Fix duplicate aggregation", "domain": "fintech", "user_story": "A query double-counts rows after joining two one-to-many tables.", "repo_context": "Repo: ml-pipeline. Domain: fintech. SQL analytics pipeline with reporting queries.", "instruction": "Fix the query issue: Fix duplicate aggregation. Explain the bug and provide the corrected pattern.", "reference_solution": "Deduplicate before aggregating, use correct grain, and be careful with time boundaries.", "evaluation": {"checks": ["correct grain", "no double count", "boundary handling"]}, "tags": ["agentic", "coding", "bug_fix", "sql", "fintech", "persian-friendly"]}
{"id": "JT3K-001284", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "sql", "difficulty": "medium", "title": "Review a reporting query", "domain": "developer-tools", "user_story": "The query is correct but slow due to an unbounded scan and unnecessary subquery.", "repo_context": "Repo: payments-api. Domain: developer-tools. SQL reporting and ETL jobs.", "instruction": "Review the SQL for: Review a reporting query. Identify correctness and performance risks.", "reference_solution": "Flag full scans, unnecessary CTEs, non-sargable filters, and inconsistent joins.", "evaluation": {"checks": ["performance risk", "logic risk", "clear recommendation"]}, "tags": ["agentic", "coding", "code_review", "sql", "developer-tools", "persian-friendly"]}
{"id": "JT3K-000910", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "sql", "difficulty": "hard", "title": "Plan migration checks", "domain": "saas", "user_story": "Inspect tables, compare counts, run a diff query, and validate migration impact.", "repo_context": "Repo: cms-platform. Domain: saas. Agent workflow over SQL files and validation queries.", "instruction": "Create a tool-use plan for: Plan migration checks. Read schema, inspect query, run checks, and validate output.", "reference_solution": "Inspect schema first, isolate the grain, run a comparison query, then verify counts after the patch.", "evaluation": {"checks": ["schema first", "comparison query", "verification"]}, "tags": ["agentic", "coding", "tool_use", "sql", "saas", "persian-friendly"]}
{"id": "JT3K-000761", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "python", "difficulty": "medium", "title": "Improve readability", "domain": "saas", "user_story": "Rename ambiguous variables and simplify nested conditionals in a config loader.", "repo_context": "Repo: analytics-core. Domain: saas. Python codebase with repeated helpers and service objects.", "instruction": "Refactor the following situation: Improve readability. Preserve behavior and reduce complexity.", "reference_solution": "Extract helpers, keep public API stable, add/adjust tests, and avoid mixing business logic with I/O.", "evaluation": {"checks": ["behavior preserved", "complexity reduced", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "python", "saas", "persian-friendly"]}
{"id": "JT3K-002726", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "javascript", "difficulty": "medium", "title": "Extract reusable hook", "domain": "developer-tools", "user_story": "Refactor repeated fetch logic into a custom React hook.", "repo_context": "Repo: analytics-core. Domain: developer-tools. React components and shared logic in a frontend repo.", "instruction": "Refactor the following: Extract reusable hook. Reduce duplication and keep the public interface stable.", "reference_solution": "Extract a hook or utility, preserve props and return shape, and add tests around behavior.", "evaluation": {"checks": ["duplication removed", "interface stable", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "javascript", "developer-tools", "persian-friendly"]}
{"id": "JT3K-002625", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "sql", "difficulty": "medium", "title": "Review a reporting query", "domain": "creator-tools", "user_story": "The query is correct but slow due to an unbounded scan and unnecessary subquery.", "repo_context": "Repo: ml-pipeline. Domain: creator-tools. SQL reporting and ETL jobs.", "instruction": "Review the SQL for: Review a reporting query. Identify correctness and performance risks.", "reference_solution": "Flag full scans, unnecessary CTEs, non-sargable filters, and inconsistent joins.", "evaluation": {"checks": ["performance risk", "logic risk", "clear recommendation"]}, "tags": ["agentic", "coding", "code_review", "sql", "creator-tools", "persian-friendly"]}
{"id": "JT3K-001822", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "python", "difficulty": "hard", "title": "Review a cache helper", "domain": "ecommerce", "user_story": "A cache wrapper forgets to namespace keys and can collide across tenants.", "repo_context": "Repo: ml-pipeline. Domain: ecommerce. Python backend with API handlers and background workers.", "instruction": "Review the code for: Review a cache helper. Call out risks and give a practical improvement plan.", "reference_solution": "Flag correctness, performance, maintainability, and security concerns. Recommend the smallest safe refactor.", "evaluation": {"checks": ["risk identified", "actionable fix", "production awareness"]}, "tags": ["agentic", "coding", "code_review", "python", "ecommerce", "persian-friendly"]}
{"id": "JT3K-000277", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "easy", "title": "Prevent duplicate inserts", "domain": "edtech", "user_story": "A repository method can insert duplicate rows under retries. Add idempotency guards.", "repo_context": "Repo: ecommerce-backend. Domain: edtech. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Prevent duplicate inserts. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "edtech", "persian-friendly"]}
{"id": "JT3K-001000", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "sql", "difficulty": "medium", "title": "Review a reporting query", "domain": "fintech", "user_story": "The query is correct but slow due to an unbounded scan and unnecessary subquery.", "repo_context": "Repo: chat-agent. Domain: fintech. SQL reporting and ETL jobs.", "instruction": "Review the SQL for: Review a reporting query. Identify correctness and performance risks.", "reference_solution": "Flag full scans, unnecessary CTEs, non-sargable filters, and inconsistent joins.", "evaluation": {"checks": ["performance risk", "logic risk", "clear recommendation"]}, "tags": ["agentic", "coding", "code_review", "sql", "fintech", "persian-friendly"]}
{"id": "JT3K-002523", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "python", "difficulty": "easy", "title": "Lock down behavior", "domain": "developer-tools", "user_story": "Write tests for a password strength checker with explicit boundary conditions.", "repo_context": "Repo: dashboard-web. Domain: developer-tools. Python package with pytest and typed functions.", "instruction": "Write a test-first plan for: Lock down behavior. Include the key test cases before implementation.", "reference_solution": "Cover happy path, invalid input, boundaries, and failure modes. Keep tests deterministic and small.", "evaluation": {"checks": ["edge cases covered", "test names clear", "deterministic behavior"]}, "tags": ["agentic", "coding", "tdd", "python", "developer-tools", "persian-friendly"]}
{"id": "JT3K-000322", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "python", "difficulty": "hard", "title": "Plan rate limiting", "domain": "saas", "user_story": "Add a rate limiter design that can scale across multiple workers.", "repo_context": "Repo: scheduler. Domain: saas. Python backend exposing JSON APIs.", "instruction": "Design the API/contract for: Plan rate limiting. Define inputs, outputs, errors, and extensibility points.", "reference_solution": "Use explicit request/response schemas, stable error shapes, and version-friendly naming.", "evaluation": {"checks": ["clear contract", "consistent errors", "future-proof design"]}, "tags": ["agentic", "coding", "api_design", "python", "saas", "persian-friendly"]}
{"id": "JT3K-002379", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "sql", "difficulty": "medium", "title": "Plan migration checks", "domain": "infra", "user_story": "Inspect tables, compare counts, run a diff query, and validate migration impact.", "repo_context": "Repo: ml-pipeline. Domain: infra. Agent workflow over SQL files and validation queries.", "instruction": "Create a tool-use plan for: Plan migration checks. Read schema, inspect query, run checks, and validate output.", "reference_solution": "Inspect schema first, isolate the grain, run a comparison query, then verify counts after the patch.", "evaluation": {"checks": ["schema first", "comparison query", "verification"]}, "tags": ["agentic", "coding", "tool_use", "sql", "infra", "persian-friendly"]}
{"id": "JT3K-001825", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "sql", "difficulty": "easy", "title": "Review a reporting query", "domain": "edtech", "user_story": "The query is correct but slow due to an unbounded scan and unnecessary subquery.", "repo_context": "Repo: ml-pipeline. Domain: edtech. SQL reporting and ETL jobs.", "instruction": "Review the SQL for: Review a reporting query. Identify correctness and performance risks.", "reference_solution": "Flag full scans, unnecessary CTEs, non-sargable filters, and inconsistent joins.", "evaluation": {"checks": ["performance risk", "logic risk", "clear recommendation"]}, "tags": ["agentic", "coding", "code_review", "sql", "edtech", "persian-friendly"]}
{"id": "JT3K-000051", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "bash", "difficulty": "easy", "title": "Simplify script logic", "domain": "healthtech", "user_story": "Refactor a long shell script into functions with clearer error handling.", "repo_context": "Repo: auth-service. Domain: healthtech. Bash automation with functions and environment variables.", "instruction": "Refactor the script for: Simplify script logic. Reduce repetition and improve error handling.", "reference_solution": "Extract functions, use strict mode carefully, and keep reusable command pieces in variables.", "evaluation": {"checks": ["repetition reduced", "error handling", "maintainability"]}, "tags": ["agentic", "coding", "refactor", "bash", "healthtech", "persian-friendly"]}
{"id": "JT3K-002433", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "javascript", "difficulty": "medium", "title": "Test form validation", "domain": "saas", "user_story": "Create tests for password and email validation in a UI flow.", "repo_context": "Repo: ml-pipeline. Domain: saas. JS testing stack with Vitest/Jest and React Testing Library.", "instruction": "Write test cases for: Test form validation. Start with the edge cases and expected behavior.", "reference_solution": "Cover input validation, boundary values, async states, and error rendering.", "evaluation": {"checks": ["edge coverage", "async behavior", "clear expectations"]}, "tags": ["agentic", "coding", "tdd", "javascript", "saas", "persian-friendly"]}
{"id": "JT3K-001707", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "bug_fix", "language": "python", "difficulty": "easy", "title": "Fix timezone parsing", "domain": "developer-tools", "user_story": "A datetime parser returns naive objects and breaks scheduling. Patch it to handle UTC offsets safely.", "repo_context": "Repo: scheduler. Domain: developer-tools. Python service uses Pydantic models, DB repositories, and unit tests.", "instruction": "Patch the bug described here: Fix timezone parsing. Return a minimal safe fix and mention any tests to run.", "reference_solution": "Add input validation, preserve backward-compatible behavior, and include a regression test for the broken edge case.", "evaluation": {"checks": ["bug removed", "edge case handled", "regression test suggested"]}, "tags": ["agentic", "coding", "bug_fix", "python", "developer-tools", "persian-friendly"]}
{"id": "JT3K-002866", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tdd", "language": "python", "difficulty": "easy", "title": "Define a failing test suite", "domain": "creator-tools", "user_story": "Produce tests for a CSV importer that must reject malformed rows.", "repo_context": "Repo: scheduler. Domain: creator-tools. Python package with pytest and typed functions.", "instruction": "Write a test-first plan for: Define a failing test suite. Include the key test cases before implementation.", "reference_solution": "Cover happy path, invalid input, boundaries, and failure modes. Keep tests deterministic and small.", "evaluation": {"checks": ["edge cases covered", "test names clear", "deterministic behavior"]}, "tags": ["agentic", "coding", "tdd", "python", "creator-tools", "persian-friendly"]}
{"id": "JT3K-001517", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "javascript", "difficulty": "easy", "title": "Design client methods", "domain": "ecommerce", "user_story": "Design a TypeScript client for search, pagination, and retries.", "repo_context": "Repo: cms-platform. Domain: ecommerce. TypeScript client or Node API layer.", "instruction": "Design a TypeScript API for: Design client methods. Define method names, inputs, and return types.", "reference_solution": "Prefer explicit types, typed errors, pagination support, and composable options objects.", "evaluation": {"checks": ["typed contract", "clear ergonomics", "extensibility"]}, "tags": ["agentic", "coding", "api_design", "javascript", "ecommerce", "persian-friendly"]}
{"id": "JT3K-000555", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "python", "difficulty": "easy", "title": "Plan rate limiting", "domain": "infra", "user_story": "Add a rate limiter design that can scale across multiple workers.", "repo_context": "Repo: dashboard-web. Domain: infra. Python backend exposing JSON APIs.", "instruction": "Design the API/contract for: Plan rate limiting. Define inputs, outputs, errors, and extensibility points.", "reference_solution": "Use explicit request/response schemas, stable error shapes, and version-friendly naming.", "evaluation": {"checks": ["clear contract", "consistent errors", "future-proof design"]}, "tags": ["agentic", "coding", "api_design", "python", "infra", "persian-friendly"]}
{"id": "JT3K-001113", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "javascript", "difficulty": "easy", "title": "Plan verify-edit cycle", "domain": "developer-tools", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: ml-pipeline. Domain: developer-tools. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "javascript", "developer-tools", "persian-friendly"]}
{"id": "JT3K-001902", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "javascript", "difficulty": "medium", "title": "Analyze hydration mismatch", "domain": "ecommerce", "user_story": "A server-rendered page renders different timestamps on client and server.", "repo_context": "Repo: dashboard-web. Domain: ecommerce. Frontend app with hydration and async data flow.", "instruction": "Diagnose the issue: Analyze hydration mismatch. Explain likely cause and how to verify the fix.", "reference_solution": "Check server/client output mismatches, stale closures, promise handling, and state synchronization.", "evaluation": {"checks": ["root cause", "verification steps", "frontend-specific diagnosis"]}, "tags": ["agentic", "coding", "debug_trace", "javascript", "ecommerce", "persian-friendly"]}
{"id": "JT3K-002820", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "javascript", "difficulty": "medium", "title": "Analyze hydration mismatch", "domain": "infra", "user_story": "A server-rendered page renders different timestamps on client and server.", "repo_context": "Repo: analytics-core. Domain: infra. Frontend app with hydration and async data flow.", "instruction": "Diagnose the issue: Analyze hydration mismatch. Explain likely cause and how to verify the fix.", "reference_solution": "Check server/client output mismatches, stale closures, promise handling, and state synchronization.", "evaluation": {"checks": ["root cause", "verification steps", "frontend-specific diagnosis"]}, "tags": ["agentic", "coding", "debug_trace", "javascript", "infra", "persian-friendly"]}
{"id": "JT3K-002708", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "typescript", "difficulty": "medium", "title": "Simplify reducer logic", "domain": "healthtech", "user_story": "Reduce branching in a Redux reducer without changing behavior.", "repo_context": "Repo: cms-platform. Domain: healthtech. React components and shared logic in a frontend repo.", "instruction": "Refactor the following: Simplify reducer logic. Reduce duplication and keep the public interface stable.", "reference_solution": "Extract a hook or utility, preserve props and return shape, and add tests around behavior.", "evaluation": {"checks": ["duplication removed", "interface stable", "tests updated"]}, "tags": ["agentic", "coding", "refactor", "typescript", "healthtech", "persian-friendly"]}
{"id": "JT3K-001650", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "python", "difficulty": "easy", "title": "Review a serializer", "domain": "fintech", "user_story": "The serializer leaks internal IDs in API responses; propose a cleaner contract.", "repo_context": "Repo: payments-api. Domain: fintech. Python backend with API handlers and background workers.", "instruction": "Review the code for: Review a serializer. Call out risks and give a practical improvement plan.", "reference_solution": "Flag correctness, performance, maintainability, and security concerns. Recommend the smallest safe refactor.", "evaluation": {"checks": ["risk identified", "actionable fix", "production awareness"]}, "tags": ["agentic", "coding", "code_review", "python", "fintech", "persian-friendly"]}
{"id": "JT3K-000418", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "refactor", "language": "bash", "difficulty": "medium", "title": "Simplify script logic", "domain": "fintech", "user_story": "Refactor a long shell script into functions with clearer error handling.", "repo_context": "Repo: ml-pipeline. Domain: fintech. Bash automation with functions and environment variables.", "instruction": "Refactor the script for: Simplify script logic. Reduce repetition and improve error handling.", "reference_solution": "Extract functions, use strict mode carefully, and keep reusable command pieces in variables.", "evaluation": {"checks": ["repetition reduced", "error handling", "maintainability"]}, "tags": ["agentic", "coding", "refactor", "bash", "fintech", "persian-friendly"]}
{"id": "JT3K-001545", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "python", "difficulty": "medium", "title": "Plan rate limiting", "domain": "edtech", "user_story": "Add a rate limiter design that can scale across multiple workers.", "repo_context": "Repo: ml-pipeline. Domain: edtech. Python backend exposing JSON APIs.", "instruction": "Design the API/contract for: Plan rate limiting. Define inputs, outputs, errors, and extensibility points.", "reference_solution": "Use explicit request/response schemas, stable error shapes, and version-friendly naming.", "evaluation": {"checks": ["clear contract", "consistent errors", "future-proof design"]}, "tags": ["agentic", "coding", "api_design", "python", "edtech", "persian-friendly"]}
{"id": "JT3K-001308", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "typescript", "difficulty": "hard", "title": "Plan verify-edit cycle", "domain": "saas", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: analytics-core. Domain: saas. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "typescript", "saas", "persian-friendly"]}
{"id": "JT3K-002359", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "python", "difficulty": "medium", "title": "Review a serializer", "domain": "infra", "user_story": "The serializer leaks internal IDs in API responses; propose a cleaner contract.", "repo_context": "Repo: mobile-sdk. Domain: infra. Python backend with API handlers and background workers.", "instruction": "Review the code for: Review a serializer. Call out risks and give a practical improvement plan.", "reference_solution": "Flag correctness, performance, maintainability, and security concerns. Recommend the smallest safe refactor.", "evaluation": {"checks": ["risk identified", "actionable fix", "production awareness"]}, "tags": ["agentic", "coding", "code_review", "python", "infra", "persian-friendly"]}
{"id": "JT3K-000609", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "api_design", "language": "python", "difficulty": "hard", "title": "Define response shapes", "domain": "creator-tools", "user_story": "Design consistent API responses for success, validation errors, and server errors.", "repo_context": "Repo: scheduler. Domain: creator-tools. Python backend exposing JSON APIs.", "instruction": "Design the API/contract for: Define response shapes. Define inputs, outputs, errors, and extensibility points.", "reference_solution": "Use explicit request/response schemas, stable error shapes, and version-friendly naming.", "evaluation": {"checks": ["clear contract", "consistent errors", "future-proof design"]}, "tags": ["agentic", "coding", "api_design", "python", "creator-tools", "persian-friendly"]}
{"id": "JT3K-001632", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "javascript", "difficulty": "hard", "title": "Plan verify-edit cycle", "domain": "developer-tools", "user_story": "Read component files, edit the smallest surface, run tests, and summarize outcome.", "repo_context": "Repo: chat-agent. Domain: developer-tools. Agent workflow over frontend files, tests, and build output.", "instruction": "Create a tool plan for: Plan verify-edit cycle. Include inspect, edit, test, and verify steps.", "reference_solution": "Open the minimal file set, edit the affected component or hook, run targeted tests, then confirm build health.", "evaluation": {"checks": ["tool sequence", "minimal edit", "verification"]}, "tags": ["agentic", "coding", "tool_use", "javascript", "developer-tools", "persian-friendly"]}
{"id": "JT3K-001101", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "javascript", "difficulty": "medium", "title": "Review an API client", "domain": "healthtech", "user_story": "The client mixes transport errors and business errors in one catch block.", "repo_context": "Repo: chat-agent. Domain: healthtech. Frontend app with shared utilities and API wrappers.", "instruction": "Review the implementation for: Review an API client. Comment on correctness, UX, and maintainability.", "reference_solution": "Look for async race conditions, unnecessary re-renders, error handling gaps, and accessibility issues.", "evaluation": {"checks": ["quality issues found", "practical suggestions", "frontend concerns"]}, "tags": ["agentic", "coding", "code_review", "javascript", "healthtech", "persian-friendly"]}
{"id": "JT3K-001674", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "tool_use", "language": "bash", "difficulty": "medium", "title": "Plan release verification", "domain": "developer-tools", "user_story": "Run lint, unit tests, packaging, and smoke checks before publishing.", "repo_context": "Repo: analytics-core. Domain: developer-tools. Release and verification scripts in bash.", "instruction": "Build a tool-use plan for: Plan release verification. Include commands to inspect, run, and validate.", "reference_solution": "List the exact commands in order: inspect files, run targeted checks, then package and smoke test.", "evaluation": {"checks": ["command order", "verification", "safe execution"]}, "tags": ["agentic", "coding", "tool_use", "bash", "developer-tools", "persian-friendly"]}
{"id": "JT3K-000257", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "code_review", "language": "javascript", "difficulty": "medium", "title": "Review an API client", "domain": "creator-tools", "user_story": "The client mixes transport errors and business errors in one catch block.", "repo_context": "Repo: auth-service. Domain: creator-tools. Frontend app with shared utilities and API wrappers.", "instruction": "Review the implementation for: Review an API client. Comment on correctness, UX, and maintainability.", "reference_solution": "Look for async race conditions, unnecessary re-renders, error handling gaps, and accessibility issues.", "evaluation": {"checks": ["quality issues found", "practical suggestions", "frontend concerns"]}, "tags": ["agentic", "coding", "code_review", "javascript", "creator-tools", "persian-friendly"]}
{"id": "JT3K-000353", "dataset": "JumpTrace-3K", "brand": "JumpTrace-3K — Agentic Coding Traces for Modern Software Engineering", "task_type": "debug_trace", "language": "python", "difficulty": "easy", "title": "Trace a failing unit test", "domain": "fintech", "user_story": "A test intermittently fails due to shared mutable state between cases.", "repo_context": "Repo: ecommerce-backend. Domain: fintech. Python service with logs, stack traces, and tests.", "instruction": "Diagnose this failure: Trace a failing unit test. Provide the most likely cause and the verification steps.", "reference_solution": "Read the stack trace, identify the failing boundary, isolate shared state or missing key paths, then add a regression test.", "evaluation": {"checks": ["root cause", "debug steps", "verification plan"]}, "tags": ["agentic", "coding", "debug_trace", "python", "fintech", "persian-friendly"]}