{
  "meta": {
    "benchmark": "SkillRetBench",
    "generator": "app.sefo.benchmark.query_generator",
    "corpus_meta": {
      "generated_at": "2026-03-22T04:54:41.727090+00:00",
      "skills_root": "/Users/hanhyojung/work/thakicloud/ai-model-event-stock-analytics/.cursor/skills",
      "output_path": "/Users/hanhyojung/work/thakicloud/ai-model-event-stock-analytics/backend/app/sefo/benchmark/data/skill_corpus.json",
      "skill_file_count": 501
    },
    "total_queries": 1250,
    "counts_by_setting": {
      "single_skill": 400,
      "multi_skill_composition": 200,
      "distractor": 300,
      "outdated_redundant": 150,
      "budget_constrained": 200
    },
    "rng_seed": 42,
    "annotation_pipeline": "app.sefo.benchmark.annotation_pipeline",
    "judge_mode": "_judge",
    "low_confidence_count": 539,
    "confidence_threshold_review": 0.72
  },
  "queries": [
    {
      "query_id": "SS-0001",
      "setting": "single_skill",
      "query": "업무 맥락에서 '종합 리서치'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?",
      "gold_skills": [
        "deep-research-pipeline"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n업무 맥락에서 '종합 리서치'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?\n\nProposed gold_skills (ordered if composition):\n[\"deep-research-pipeline\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0002",
      "setting": "single_skill",
      "query": "Can you help me the role-dispatcher invokes this skill with a topic? Pick the most specific matching skill.",
      "gold_skills": [
        "role-ceo"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me the role-dispatcher invokes this skill with a topic? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"role-ceo\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0003",
      "setting": "single_skill",
      "query": "I'm trying to references agency-sprint-prioritizer; route me to the right capability.",
      "gold_skills": [
        "agency-sprint-prioritizer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to references agency-sprint-prioritizer; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"agency-sprint-prioritizer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0004",
      "setting": "single_skill",
      "query": "User request: approaching context limits. What single skill applies?",
      "gold_skills": [
        "ecc-strategic-compact"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: approaching context limits. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"ecc-strategic-compact\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0005",
      "setting": "single_skill",
      "query": "Looking for automation that covers: references agency-tool-evaluator.",
      "gold_skills": [
        "agency-tool-evaluator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: references agency-tool-evaluator.\n\nProposed gold_skills (ordered if composition):\n[\"agency-tool-evaluator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0006",
      "setting": "single_skill",
      "query": "Match this intent to one skill: design an AI workflow.",
      "gold_skills": [
        "ai-workflow-integrator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: design an AI workflow.\n\nProposed gold_skills (ordered if composition):\n[\"ai-workflow-integrator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0007",
      "setting": "single_skill",
      "query": "Task description: check for upstream KWP changes. Identify the canonical skill id.",
      "gold_skills": [
        "kwp-sync"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: check for upstream KWP changes. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-sync\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0008",
      "setting": "single_skill",
      "query": "I need to run critical review—which skill should handle this?",
      "gold_skills": [
        "critical-review"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to run critical review—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"critical-review\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0009",
      "setting": "single_skill",
      "query": "Can you help me sentiment? Pick the most specific matching skill.",
      "gold_skills": [
        "alphaear-deepear-lite"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me sentiment? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-deepear-lite\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0010",
      "setting": "single_skill",
      "query": "I'm trying to create NLM video; route me to the right capability.",
      "gold_skills": [
        "nlm-video"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to create NLM video; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"nlm-video\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0011",
      "setting": "single_skill",
      "query": "한국어로 '리포트' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "kwp-product-management-stakeholder-comms"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '리포트' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-product-management-stakeholder-comms\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0012",
      "setting": "single_skill",
      "query": "Looking for automation that covers: detecting conflicts with installed skills.",
      "gold_skills": [
        "skill-seekers"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: detecting conflicts with installed skills.\n\nProposed gold_skills (ordered if composition):\n[\"skill-seekers\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0013",
      "setting": "single_skill",
      "query": "Match this intent to one skill: produce one.",
      "gold_skills": [
        "anthropic-pdf"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: produce one.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-pdf\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0014",
      "setting": "single_skill",
      "query": "Task description: upload a trained model. Identify the canonical skill id.",
      "gold_skills": [
        "hf-models"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: upload a trained model. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"hf-models\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0015",
      "setting": "single_skill",
      "query": "I need to 에이전트 승인—which skill should handle this?",
      "gold_skills": [
        "paperclip-control"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to 에이전트 승인—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"paperclip-control\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0016",
      "setting": "single_skill",
      "query": "업무 맥락에서 '스킬'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?",
      "gold_skills": [
        "agency-visionos-spatial-engineer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n업무 맥락에서 '스킬'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?\n\nProposed gold_skills (ordered if composition):\n[\"agency-visionos-spatial-engineer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0017",
      "setting": "single_skill",
      "query": "I'm trying to 메일 답장; route me to the right capability.",
      "gold_skills": [
        "gws-email-reply"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to 메일 답장; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"gws-email-reply\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0018",
      "setting": "single_skill",
      "query": "User request: 相場環境. What single skill applies?",
      "gold_skills": [
        "trading-market-environment-analysis"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: 相場環境. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"trading-market-environment-analysis\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0019",
      "setting": "single_skill",
      "query": "Looking for automation that covers: 의도 정렬 점수.",
      "gold_skills": [
        "intent-alignment-tracker"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: 의도 정렬 점수.\n\nProposed gold_skills (ordered if composition):\n[\"intent-alignment-tracker\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0020",
      "setting": "single_skill",
      "query": "Match this intent to one skill: when updating existing KB articles.",
      "gold_skills": [
        "kwp-customer-support-knowledge-management"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: when updating existing KB articles.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-customer-support-knowledge-management\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0021",
      "setting": "single_skill",
      "query": "다음 요청과 맞는 워크플로우를 찾아줘: 권한 감사 관련.",
      "gold_skills": [
        "gws-recipe-audit-sharing"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n다음 요청과 맞는 워크플로우를 찾아줘: 권한 감사 관련.\n\nProposed gold_skills (ordered if composition):\n[\"gws-recipe-audit-sharing\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0022",
      "setting": "single_skill",
      "query": "I need to building an MCP server—which skill should handle this?",
      "gold_skills": [
        "anthropic-mcp-builder"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to building an MCP server—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-mcp-builder\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0023",
      "setting": "single_skill",
      "query": "Can you help me GenAI 피처 생성? Pick the most specific matching skill.",
      "gold_skills": [
        "tab-genai-features"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me GenAI 피처 생성? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"tab-genai-features\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0024",
      "setting": "single_skill",
      "query": "I'm trying to subagents lack sufficient context; route me to the right capability.",
      "gold_skills": [
        "ecc-iterative-retrieval"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to subagents lack sufficient context; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"ecc-iterative-retrieval\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0025",
      "setting": "single_skill",
      "query": "User request: references agency-autonomous-optimization-architect. What single skill applies?",
      "gold_skills": [
        "agency-autonomous-optimization-architect"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: references agency-autonomous-optimization-architect. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"agency-autonomous-optimization-architect\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0026",
      "setting": "single_skill",
      "query": "업무 맥락에서 '구글 워크스페이스'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?",
      "gold_skills": [
        "gws-workspace"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n업무 맥락에서 '구글 워크스페이스'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?\n\nProposed gold_skills (ordered if composition):\n[\"gws-workspace\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0027",
      "setting": "single_skill",
      "query": "Match this intent to one skill: explore graph.",
      "gold_skills": [
        "mirofish-graph-explorer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: explore graph.\n\nProposed gold_skills (ordered if composition):\n[\"mirofish-graph-explorer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0028",
      "setting": "single_skill",
      "query": "Task description: an issue needs to go beyond support. Identify the canonical skill id.",
      "gold_skills": [
        "kwp-customer-support-escalation"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: an issue needs to go beyond support. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-customer-support-escalation\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0029",
      "setting": "single_skill",
      "query": "I need to manage branches—which skill should handle this?",
      "gold_skills": [
        "hf-repos"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to manage branches—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"hf-repos\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0030",
      "setting": "single_skill",
      "query": "Can you help me issue-to-code? Pick the most specific matching skill.",
      "gold_skills": [
        "issue-to-code"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me issue-to-code? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"issue-to-code\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0031",
      "setting": "single_skill",
      "query": "다음 요청과 맞는 워크플로우를 찾아줘: 프롬프트 관련.",
      "gold_skills": [
        "prompt-transformer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n다음 요청과 맞는 워크플로우를 찾아줘: 프롬프트 관련.\n\nProposed gold_skills (ordered if composition):\n[\"prompt-transformer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0032",
      "setting": "single_skill",
      "query": "User request: Figma → TDS Code Generator. What single skill applies?",
      "gold_skills": [
        "figma-to-tds"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: Figma → TDS Code Generator. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"figma-to-tds\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0033",
      "setting": "single_skill",
      "query": "Looking for automation that covers: Requirements for Outputs.",
      "gold_skills": [
        "anthropic-xlsx"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: Requirements for Outputs.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-xlsx\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0034",
      "setting": "single_skill",
      "query": "Match this intent to one skill: references agency-security-engineer.",
      "gold_skills": [
        "agency-security-engineer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: references agency-security-engineer.\n\nProposed gold_skills (ordered if composition):\n[\"agency-security-engineer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0035",
      "setting": "single_skill",
      "query": "Task description: Design Handoff. Identify the canonical skill id.",
      "gold_skills": [
        "kwp-design-design-handoff"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: Design Handoff. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-design-design-handoff\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0036",
      "setting": "single_skill",
      "query": "업무 맥락에서 '스킬'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?",
      "gold_skills": [
        "agency-feedback-synthesizer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n업무 맥락에서 '스킬'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?\n\nProposed gold_skills (ordered if composition):\n[\"agency-feedback-synthesizer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0037",
      "setting": "single_skill",
      "query": "Can you help me customer question routing? Pick the most specific matching skill.",
      "gold_skills": [
        "tech-support-brain"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me customer question routing? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"tech-support-brain\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0038",
      "setting": "single_skill",
      "query": "I'm trying to references agency-zhihu-strategist; route me to the right capability.",
      "gold_skills": [
        "agency-zhihu-strategist"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to references agency-zhihu-strategist; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"agency-zhihu-strategist\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0039",
      "setting": "single_skill",
      "query": "User request: batch schedule from a table. What single skill applies?",
      "gold_skills": [
        "gws-recipe-events-from-sheet"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: batch schedule from a table. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"gws-recipe-events-from-sheet\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0040",
      "setting": "single_skill",
      "query": "Looking for automation that covers: PR 영향 예측.",
      "gold_skills": [
        "mirofish-opinion-sim"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: PR 영향 예측.\n\nProposed gold_skills (ordered if composition):\n[\"mirofish-opinion-sim\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0041",
      "setting": "single_skill",
      "query": "Match this intent to one skill: Call Prep.",
      "gold_skills": [
        "kwp-common-room-call-prep"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: Call Prep.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-common-room-call-prep\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0042",
      "setting": "single_skill",
      "query": "Task description: 위협 탐지. Identify the canonical skill id.",
      "gold_skills": [
        "sefo-governance"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: 위협 탐지. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"sefo-governance\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0043",
      "setting": "single_skill",
      "query": "I need to email auto-reply—which skill should handle this?",
      "gold_skills": [
        "email-auto-reply"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to email auto-reply—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"email-auto-reply\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0044",
      "setting": "single_skill",
      "query": "Can you help me write to sheets? Pick the most specific matching skill.",
      "gold_skills": [
        "gws-sheets"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me write to sheets? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"gws-sheets\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0045",
      "setting": "single_skill",
      "query": "I'm trying to decision docs; route me to the right capability.",
      "gold_skills": [
        "anthropic-doc-coauthoring"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to decision docs; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-doc-coauthoring\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0046",
      "setting": "single_skill",
      "query": "User request: Screen Implementation Orchestrator. What single skill applies?",
      "gold_skills": [
        "implement-screen"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: Screen Implementation Orchestrator. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"implement-screen\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0047",
      "setting": "single_skill",
      "query": "Looking for automation that covers: check health.",
      "gold_skills": [
        "service-health-doctor"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: check health.\n\nProposed gold_skills (ordered if composition):\n[\"service-health-doctor\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0048",
      "setting": "single_skill",
      "query": "Match this intent to one skill: Recruiting Pipeline.",
      "gold_skills": [
        "kwp-human-resources-recruiting-pipeline"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: Recruiting Pipeline.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-human-resources-recruiting-pipeline\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0049",
      "setting": "single_skill",
      "query": "Task description: detecting events. Identify the canonical skill id.",
      "gold_skills": [
        "tab-event-detect"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: detecting events. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"tab-event-detect\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0050",
      "setting": "single_skill",
      "query": "I need to find free time—which skill should handle this?",
      "gold_skills": [
        "gws-calendar"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to find free time—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"gws-calendar\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0051",
      "setting": "single_skill",
      "query": "한국어로 '워드 문서' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "anthropic-docx"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '워드 문서' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-docx\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0052",
      "setting": "single_skill",
      "query": "I'm trying to check environment; route me to the right capability.",
      "gold_skills": [
        "setup-doctor"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to check environment; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"setup-doctor\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0053",
      "setting": "single_skill",
      "query": "User request: running analysis. What single skill applies?",
      "gold_skills": [
        "tab-analysis-run"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: running analysis. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"tab-analysis-run\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0054",
      "setting": "single_skill",
      "query": "Looking for automation that covers: Account Research.",
      "gold_skills": [
        "kwp-sales-account-research"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: Account Research.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-sales-account-research\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0055",
      "setting": "single_skill",
      "query": "Match this intent to one skill: references agency-terminal-integration-specialist.",
      "gold_skills": [
        "agency-terminal-integration-specialist"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: references agency-terminal-integration-specialist.\n\nProposed gold_skills (ordered if composition):\n[\"agency-terminal-integration-specialist\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0056",
      "setting": "single_skill",
      "query": "업무 맥락에서 '데이터'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?",
      "gold_skills": [
        "kwp-data-data-visualization"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n업무 맥락에서 '데이터'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-data-data-visualization\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0057",
      "setting": "single_skill",
      "query": "I need to Compliance Tracking—which skill should handle this?",
      "gold_skills": [
        "kwp-operations-compliance-tracking"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to Compliance Tracking—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-operations-compliance-tracking\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0058",
      "setting": "single_skill",
      "query": "Can you help me 변경사항 데모? Pick the most specific matching skill.",
      "gold_skills": [
        "demo-forge"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me 변경사항 데모? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"demo-forge\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0059",
      "setting": "single_skill",
      "query": "I'm trying to comparing experiment results; route me to the right capability.",
      "gold_skills": [
        "hf-trackio"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to comparing experiment results; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"hf-trackio\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0060",
      "setting": "single_skill",
      "query": "User request: wants automated PR-issue relationship management. What single skill applies?",
      "gold_skills": [
        "pr-to-issue-linker"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: wants automated PR-issue relationship management. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"pr-to-issue-linker\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0061",
      "setting": "single_skill",
      "query": "한국어로 'E2E 오버홀' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "e2e-overhaul"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 'E2E 오버홀' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"e2e-overhaul\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0062",
      "setting": "single_skill",
      "query": "Match this intent to one skill: 매크로 포지셔닝.",
      "gold_skills": [
        "trading-stanley-druckenmiller-investment"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: 매크로 포지셔닝.\n\nProposed gold_skills (ordered if composition):\n[\"trading-stanley-druckenmiller-investment\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0063",
      "setting": "single_skill",
      "query": "Task description: references agency-reddit-community-builder. Identify the canonical skill id.",
      "gold_skills": [
        "agency-reddit-community-builder"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: references agency-reddit-community-builder. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"agency-reddit-community-builder\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0064",
      "setting": "single_skill",
      "query": "I need to Brainstorming Ideas Into Designs—which skill should handle this?",
      "gold_skills": [
        "sp-brainstorming"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to Brainstorming Ideas Into Designs—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"sp-brainstorming\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0065",
      "setting": "single_skill",
      "query": "Can you help me references agency-analytics-reporter? Pick the most specific matching skill.",
      "gold_skills": [
        "agency-analytics-reporter"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me references agency-analytics-reporter? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"agency-analytics-reporter\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0066",
      "setting": "single_skill",
      "query": "한국어로 '슬라이드' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "anthropic-pptx"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '슬라이드' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-pptx\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0067",
      "setting": "single_skill",
      "query": "User request: 회사 템플릿으로 문서 만들기. What single skill applies?",
      "gold_skills": [
        "office-template-enforcer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: 회사 템플릿으로 문서 만들기. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"office-template-enforcer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0068",
      "setting": "single_skill",
      "query": "Looking for automation that covers: using same resources) ## Real Example from Session **Scenario:** 6 test failures across 3 files after major refactori....",
      "gold_skills": [
        "sp-parallel-agents"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: using same resources) ## Real Example from Session **Scenario:** 6 test failures across 3 files after major refactori....\n\nProposed gold_skills (ordered if composition):\n[\"sp-parallel-agents\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0069",
      "setting": "single_skill",
      "query": "Match this intent to one skill: 포트폴리오 리스크 배분.",
      "gold_skills": [
        "trading-position-sizer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: 포트폴리오 리스크 배분.\n\nProposed gold_skills (ordered if composition):\n[\"trading-position-sizer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0070",
      "setting": "single_skill",
      "query": "Task description: tab-patterns. Identify the canonical skill id.",
      "gold_skills": [
        "tab-patterns"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: tab-patterns. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"tab-patterns\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0071",
      "setting": "single_skill",
      "query": "다음 요청과 맞는 워크플로우를 찾아줘: GIF 관련.",
      "gold_skills": [
        "anthropic-slack-gif-creator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n다음 요청과 맞는 워크플로우를 찾아줘: GIF 관련.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-slack-gif-creator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0072",
      "setting": "single_skill",
      "query": "Can you help me Design System Management? Pick the most specific matching skill.",
      "gold_skills": [
        "kwp-design-design-system-management"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me Design System Management? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-design-design-system-management\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0073",
      "setting": "single_skill",
      "query": "I'm trying to get markdown from website; route me to the right capability.",
      "gold_skills": [
        "defuddle"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to get markdown from website; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"defuddle\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0074",
      "setting": "single_skill",
      "query": "User request: RTK 분석. What single skill applies?",
      "gold_skills": [
        "rtk"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: RTK 분석. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"rtk\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0075",
      "setting": "single_skill",
      "query": "Looking for automation that covers: build strategy notebook.",
      "gold_skills": [
        "nlm-strategy-hub"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: build strategy notebook.\n\nProposed gold_skills (ordered if composition):\n[\"nlm-strategy-hub\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0076",
      "setting": "single_skill",
      "query": "한국어로 '데이터셋 탐색' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "hf-dataset-viewer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '데이터셋 탐색' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"hf-dataset-viewer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0077",
      "setting": "single_skill",
      "query": "Task description: managing HF repos. Identify the canonical skill id.",
      "gold_skills": [
        "hf-cli"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: managing HF repos. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"hf-cli\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0078",
      "setting": "single_skill",
      "query": "I need to frame this problem—which skill should handle this?",
      "gold_skills": [
        "problem-definition"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to frame this problem—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"problem-definition\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0079",
      "setting": "single_skill",
      "query": "Can you help me validate K8s manifests? Pick the most specific matching skill.",
      "gold_skills": [
        "k8s-manifest-validator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me validate K8s manifests? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"k8s-manifest-validator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0080",
      "setting": "single_skill",
      "query": "I'm trying to strategic tactical market outlook; route me to the right capability.",
      "gold_skills": [
        "trading-breadth-chart-analyst"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to strategic tactical market outlook; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"trading-breadth-chart-analyst\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0081",
      "setting": "single_skill",
      "query": "다음 요청과 맞는 워크플로우를 찾아줘: 생성 관련.",
      "gold_skills": [
        "agency-content-creator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n다음 요청과 맞는 워크플로우를 찾아줘: 생성 관련.\n\nProposed gold_skills (ordered if composition):\n[\"agency-content-creator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0082",
      "setting": "single_skill",
      "query": "Looking for automation that covers: activate the App Store Optimizer agent persona.",
      "gold_skills": [
        "agency-app-store-optimizer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: activate the App Store Optimizer agent persona.\n\nProposed gold_skills (ordered if composition):\n[\"agency-app-store-optimizer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0083",
      "setting": "single_skill",
      "query": "Match this intent to one skill: Resource Planning.",
      "gold_skills": [
        "kwp-operations-resource-planning"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: Resource Planning.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-operations-resource-planning\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0084",
      "setting": "single_skill",
      "query": "Task description: UX Writing. Identify the canonical skill id.",
      "gold_skills": [
        "kwp-design-ux-writing"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: UX Writing. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-design-ux-writing\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0085",
      "setting": "single_skill",
      "query": "I need to Process Optimization—which skill should handle this?",
      "gold_skills": [
        "kwp-operations-process-optimization"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to Process Optimization—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-operations-process-optimization\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0086",
      "setting": "single_skill",
      "query": "한국어로 '버킷 동기화' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "hf-buckets"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '버킷 동기화' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"hf-buckets\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0087",
      "setting": "single_skill",
      "query": "I'm trying to wants a single pipeline for all external information sources; route me to the right capability.",
      "gold_skills": [
        "unified-intel-intake"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to wants a single pipeline for all external information sources; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"unified-intel-intake\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0088",
      "setting": "single_skill",
      "query": "User request: post meeting. What single skill applies?",
      "gold_skills": [
        "sales-meeting-pipeline"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: post meeting. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"sales-meeting-pipeline\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0089",
      "setting": "single_skill",
      "query": "Looking for automation that covers: incident lifecycle.",
      "gold_skills": [
        "incident-lifecycle-manager"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: incident lifecycle.\n\nProposed gold_skills (ordered if composition):\n[\"incident-lifecycle-manager\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0090",
      "setting": "single_skill",
      "query": "Match this intent to one skill: approach. High risk.",
      "gold_skills": [
        "kwp-product-management-competitive-analysis"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: approach. High risk.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-product-management-competitive-analysis\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0091",
      "setting": "single_skill",
      "query": "업무 맥락에서 '프레젠테이션'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?",
      "gold_skills": [
        "agency-inclusive-visuals-specialist"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n업무 맥락에서 '프레젠테이션'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?\n\nProposed gold_skills (ordered if composition):\n[\"agency-inclusive-visuals-specialist\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0092",
      "setting": "single_skill",
      "query": "I need to news-informed forecast adjustments—which skill should handle this?",
      "gold_skills": [
        "alphaear-predictor"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to news-informed forecast adjustments—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-predictor\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0093",
      "setting": "single_skill",
      "query": "Can you help me infra-runbook-generator? Pick the most specific matching skill.",
      "gold_skills": [
        "infra-runbook-generator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me infra-runbook-generator? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"infra-runbook-generator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0094",
      "setting": "single_skill",
      "query": "I'm trying to Testing Strategy; route me to the right capability.",
      "gold_skills": [
        "kwp-engineering-testing-strategy"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to Testing Strategy; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-engineering-testing-strategy\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0095",
      "setting": "single_skill",
      "query": "User request: 인시던트. What single skill applies?",
      "gold_skills": [
        "incident-to-improvement"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: 인시던트. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"incident-to-improvement\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0096",
      "setting": "single_skill",
      "query": "한국어로 '영업 분석' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "role-sales"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '영업 분석' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"role-sales\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0097",
      "setting": "single_skill",
      "query": "Match this intent to one skill: running evals to test a skill.",
      "gold_skills": [
        "anthropic-skill-creator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: running evals to test a skill.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-skill-creator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0098",
      "setting": "single_skill",
      "query": "Task description: research a topic in NotebookLM. Identify the canonical skill id.",
      "gold_skills": [
        "notebooklm-research"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: research a topic in NotebookLM. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"notebooklm-research\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0099",
      "setting": "single_skill",
      "query": "I need to find items needing discussion—which skill should handle this?",
      "gold_skills": [
        "feedback-meeting-scheduler"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to find items needing discussion—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"feedback-meeting-scheduler\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0100",
      "setting": "single_skill",
      "query": "Can you help me convert knowledge to paper? Pick the most specific matching skill.",
      "gold_skills": [
        "knowledge-to-paper"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me convert knowledge to paper? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"knowledge-to-paper\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0101",
      "setting": "single_skill",
      "query": "I'm trying to wants to automatically extract research topics from incoming emails and distribute findings; route me to the right capability.",
      "gold_skills": [
        "email-research-dispatcher"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to wants to automatically extract research topics from incoming emails and distribute findings; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"email-research-dispatcher\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0102",
      "setting": "single_skill",
      "query": "User request: weekly sales report. What single skill applies?",
      "gold_skills": [
        "sales-weekly-dashboard"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: weekly sales report. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"sales-weekly-dashboard\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0103",
      "setting": "single_skill",
      "query": "Looking for automation that covers: calls to action.",
      "gold_skills": [
        "kwp-marketing-content-creation"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: calls to action.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-marketing-content-creation\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0104",
      "setting": "single_skill",
      "query": "Match this intent to one skill: reviewing an analysis for errors.",
      "gold_skills": [
        "kwp-data-data-validation"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: reviewing an analysis for errors.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-data-data-validation\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0105",
      "setting": "single_skill",
      "query": "Task description: architect and plan. Identify the canonical skill id.",
      "gold_skills": [
        "plans"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: architect and plan. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"plans\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0106",
      "setting": "single_skill",
      "query": "한국어로 'HF Jobs' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "hf-jobs"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 'HF Jobs' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"hf-jobs\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0107",
      "setting": "single_skill",
      "query": "Can you help me brand colors? Pick the most specific matching skill.",
      "gold_skills": [
        "anthropic-brand-guidelines"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me brand colors? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-brand-guidelines\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0108",
      "setting": "single_skill",
      "query": "I'm trying to quality audit; route me to the right capability.",
      "gold_skills": [
        "mission-control"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to quality audit; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"mission-control\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0109",
      "setting": "single_skill",
      "query": "User request: query optimization. What single skill applies?",
      "gold_skills": [
        "db-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: query optimization. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"db-expert\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0110",
      "setting": "single_skill",
      "query": "Looking for automation that covers: 보안 엔지니어 분석.",
      "gold_skills": [
        "role-security-engineer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: 보안 엔지니어 분석.\n\nProposed gold_skills (ordered if composition):\n[\"role-security-engineer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0111",
      "setting": "single_skill",
      "query": "업무 맥락에서 '체크'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?",
      "gold_skills": [
        "ecc-verification-loop"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n업무 맥락에서 '체크'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?\n\nProposed gold_skills (ordered if composition):\n[\"ecc-verification-loop\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0112",
      "setting": "single_skill",
      "query": "Task description: run the project locally. Identify the canonical skill id.",
      "gold_skills": [
        "local-dev-runner"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: run the project locally. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"local-dev-runner\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0113",
      "setting": "single_skill",
      "query": "I need to activate the Report Distribution Agent agent persona—which skill should handle this?",
      "gold_skills": [
        "agency-report-distribution-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to activate the Report Distribution Agent agent persona—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"agency-report-distribution-agent\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0114",
      "setting": "single_skill",
      "query": "Can you help me customize plugin command? Pick the most specific matching skill.",
      "gold_skills": [
        "kwp-cowork-plugin-customizer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me customize plugin command? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-cowork-plugin-customizer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0115",
      "setting": "single_skill",
      "query": "I'm trying to iOS mobile testing; route me to the right capability.",
      "gold_skills": [
        "agent-browser"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to iOS mobile testing; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"agent-browser\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0116",
      "setting": "single_skill",
      "query": "User request: evaluate extracted skill. What single skill applies?",
      "gold_skills": [
        "autoskill-judge"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: evaluate extracted skill. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"autoskill-judge\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0117",
      "setting": "single_skill",
      "query": "Looking for automation that covers: resolve fork.",
      "gold_skills": [
        "sefo-federation"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: resolve fork.\n\nProposed gold_skills (ordered if composition):\n[\"sefo-federation\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0118",
      "setting": "single_skill",
      "query": "Match this intent to one skill: threat model.",
      "gold_skills": [
        "security-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: threat model.\n\nProposed gold_skills (ordered if composition):\n[\"security-expert\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0119",
      "setting": "single_skill",
      "query": "Task description: weekly pulse. Identify the canonical skill id.",
      "gold_skills": [
        "planning-weekly-pulse"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: weekly pulse. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"planning-weekly-pulse\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0120",
      "setting": "single_skill",
      "query": "I need to identifying what's working and what needs improvement—which skill should handle this?",
      "gold_skills": [
        "kwp-marketing-performance-analytics"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to identifying what's working and what needs improvement—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-marketing-performance-analytics\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0121",
      "setting": "single_skill",
      "query": "한국어로 '빌드' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "agency-macos-spatial-metal-engineer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '빌드' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"agency-macos-spatial-metal-engineer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0122",
      "setting": "single_skill",
      "query": "I'm trying to product impact; route me to the right capability.",
      "gold_skills": [
        "role-pm"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to product impact; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"role-pm\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0123",
      "setting": "single_skill",
      "query": "User request: references agency-backend-architect. What single skill applies?",
      "gold_skills": [
        "agency-backend-architect"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: references agency-backend-architect. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"agency-backend-architect\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0124",
      "setting": "single_skill",
      "query": "Looking for automation that covers: tracking close progress.",
      "gold_skills": [
        "kwp-finance-close-management"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: tracking close progress.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-finance-close-management\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0125",
      "setting": "single_skill",
      "query": "Match this intent to one skill: analyze AAPL.",
      "gold_skills": [
        "trading-us-stock-analysis"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: analyze AAPL.\n\nProposed gold_skills (ordered if composition):\n[\"trading-us-stock-analysis\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0126",
      "setting": "single_skill",
      "query": "Task description: Swagger API Documentation Generator. Identify the canonical skill id.",
      "gold_skills": [
        "swagger-api-doc-generator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: Swagger API Documentation Generator. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"swagger-api-doc-generator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0127",
      "setting": "single_skill",
      "query": "I need to Tech Debt Management—which skill should handle this?",
      "gold_skills": [
        "kwp-engineering-tech-debt"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to Tech Debt Management—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-engineering-tech-debt\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0128",
      "setting": "single_skill",
      "query": "Can you help me study the docs? Pick the most specific matching skill.",
      "gold_skills": [
        "docs-tutor"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me study the docs? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"docs-tutor\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0129",
      "setting": "single_skill",
      "query": "I'm trying to operational readiness; route me to the right capability.",
      "gold_skills": [
        "sre-devops-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to operational readiness; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"sre-devops-expert\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0130",
      "setting": "single_skill",
      "query": "User request: US Market Bubble Detection Skill (Revised v2.1). What single skill applies?",
      "gold_skills": [
        "trading-us-market-bubble-detector"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: US Market Bubble Detection Skill (Revised v2.1). What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"trading-us-market-bubble-detector\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0131",
      "setting": "single_skill",
      "query": "Looking for automation that covers: what if I change.",
      "gold_skills": [
        "refactor-simulator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: what if I change.\n\nProposed gold_skills (ordered if composition):\n[\"refactor-simulator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0132",
      "setting": "single_skill",
      "query": "Match this intent to one skill: bundle optimization.",
      "gold_skills": [
        "frontend-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: bundle optimization.\n\nProposed gold_skills (ordered if composition):\n[\"frontend-expert\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0133",
      "setting": "single_skill",
      "query": "Task description: references agency-agents-orchestrator. Identify the canonical skill id.",
      "gold_skills": [
        "agency-agents-orchestrator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: references agency-agents-orchestrator. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"agency-agents-orchestrator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0134",
      "setting": "single_skill",
      "query": "I need to references agency-evidence-collector—which skill should handle this?",
      "gold_skills": [
        "agency-evidence-collector"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to references agency-evidence-collector—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"agency-evidence-collector\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0135",
      "setting": "single_skill",
      "query": "Can you help me needs template-compliant DOCX output that preserves corporate branding and structure? Pick the most specific matching skill.",
      "gold_skills": [
        "docx-template-engine"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me needs template-compliant DOCX output that preserves corporate branding and structure? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"docx-template-engine\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0136",
      "setting": "single_skill",
      "query": "한국어로 '데이터' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "kwp-human-resources-people-analytics"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '데이터' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-human-resources-people-analytics\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0137",
      "setting": "single_skill",
      "query": "User request: Polymarket finance prediction data. What single skill applies?",
      "gold_skills": [
        "alphaear-news"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: Polymarket finance prediction data. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-news\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0138",
      "setting": "single_skill",
      "query": "Looking for automation that covers: set up ralph loop.",
      "gold_skills": [
        "ralph-loop"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: set up ralph loop.\n\nProposed gold_skills (ordered if composition):\n[\"ralph-loop\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0139",
      "setting": "single_skill",
      "query": "Match this intent to one skill: Local Development Setup.",
      "gold_skills": [
        "local-dev-setup"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: Local Development Setup.\n\nProposed gold_skills (ordered if composition):\n[\"local-dev-setup\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0140",
      "setting": "single_skill",
      "query": "Task description: 12개 직무 관점. Identify the canonical skill id.",
      "gold_skills": [
        "role-dispatcher"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: 12개 직무 관점. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"role-dispatcher\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0141",
      "setting": "single_skill",
      "query": "한국어로 '가치 제안' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "pm-product-strategy"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '가치 제안' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"pm-product-strategy\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0142",
      "setting": "single_skill",
      "query": "Can you help me Accessibility Review? Pick the most specific matching skill.",
      "gold_skills": [
        "kwp-design-accessibility-review"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me Accessibility Review? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-design-accessibility-review\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0143",
      "setting": "single_skill",
      "query": "I'm trying to running LLM agents; route me to the right capability.",
      "gold_skills": [
        "tab-llm-agents"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to running LLM agents; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"tab-llm-agents\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0144",
      "setting": "single_skill",
      "query": "User request: references agency-xr-immersive-developer. What single skill applies?",
      "gold_skills": [
        "agency-xr-immersive-developer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: references agency-xr-immersive-developer. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"agency-xr-immersive-developer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0145",
      "setting": "single_skill",
      "query": "Looking for automation that covers: Google Daily Automation.",
      "gold_skills": [
        "google-daily"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: Google Daily Automation.\n\nProposed gold_skills (ordered if composition):\n[\"google-daily\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0146",
      "setting": "single_skill",
      "query": "업무 맥락에서 '체크'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?",
      "gold_skills": [
        "dependency-auditor"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n업무 맥락에서 '체크'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?\n\nProposed gold_skills (ordered if composition):\n[\"dependency-auditor\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0147",
      "setting": "single_skill",
      "query": "Task description: automate browser-based testing scenarios. Identify the canonical skill id.",
      "gold_skills": [
        "e2e-testing"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: automate browser-based testing scenarios. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"e2e-testing\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0148",
      "setting": "single_skill",
      "query": "I need to 문장 다듬기—which skill should handle this?",
      "gold_skills": [
        "sentence-polisher"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to 문장 다듬기—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"sentence-polisher\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0149",
      "setting": "single_skill",
      "query": "Can you help me UI 변경 요청 시? Pick the most specific matching skill.",
      "gold_skills": [
        "fe-pipeline"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me UI 변경 요청 시? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"fe-pipeline\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0150",
      "setting": "single_skill",
      "query": "I'm trying to launching a campaign; route me to the right capability.",
      "gold_skills": [
        "kwp-marketing-campaign-planning"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to launching a campaign; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-marketing-campaign-planning\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0151",
      "setting": "single_skill",
      "query": "User request: Slack Messaging Best Practices. What single skill applies?",
      "gold_skills": [
        "kwp-slack-slack-messaging"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: Slack Messaging Best Practices. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-slack-slack-messaging\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0152",
      "setting": "single_skill",
      "query": "Looking for automation that covers: 인프라 드리프트.",
      "gold_skills": [
        "infra-drift-detector"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: 인프라 드리프트.\n\nProposed gold_skills (ordered if composition):\n[\"infra-drift-detector\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0153",
      "setting": "single_skill",
      "query": "Match this intent to one skill: planning agent-delegated work.",
      "gold_skills": [
        "ecc-agentic-engineering"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: planning agent-delegated work.\n\nProposed gold_skills (ordered if composition):\n[\"ecc-agentic-engineering\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0154",
      "setting": "single_skill",
      "query": "Task description: Search Strategy. Identify the canonical skill id.",
      "gold_skills": [
        "kwp-enterprise-search-search-strategy"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: Search Strategy. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-enterprise-search-search-strategy\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0155",
      "setting": "single_skill",
      "query": "I need to Brand Voice Enforcement—which skill should handle this?",
      "gold_skills": [
        "kwp-brand-voice-brand-voice-enforcement"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to Brand Voice Enforcement—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-brand-voice-brand-voice-enforcement\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0156",
      "setting": "single_skill",
      "query": "다음 요청과 맞는 워크플로우를 찾아줘: 파인튜닝 관련.",
      "gold_skills": [
        "hf-model-trainer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n다음 요청과 맞는 워크플로우를 찾아줘: 파인튜닝 관련.\n\nProposed gold_skills (ordered if composition):\n[\"hf-model-trainer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0157",
      "setting": "single_skill",
      "query": "I'm trying to synthetic data; route me to the right capability.",
      "gold_skills": [
        "evals-skills"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to synthetic data; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"evals-skills\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0158",
      "setting": "single_skill",
      "query": "User request: upload markdown to Slack Canvas. What single skill applies?",
      "gold_skills": [
        "md-to-slack-canvas"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: upload markdown to Slack Canvas. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"md-to-slack-canvas\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0159",
      "setting": "single_skill",
      "query": "Looking for automation that covers: activate the Devops Automator agent persona.",
      "gold_skills": [
        "agency-devops-automator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: activate the Devops Automator agent persona.\n\nProposed gold_skills (ordered if composition):\n[\"agency-devops-automator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0160",
      "setting": "single_skill",
      "query": "Match this intent to one skill: Kiwoom API.",
      "gold_skills": [
        "tab-kiwoom"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: Kiwoom API.\n\nProposed gold_skills (ordered if composition):\n[\"tab-kiwoom\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0161",
      "setting": "single_skill",
      "query": "업무 맥락에서 '테스트'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?",
      "gold_skills": [
        "agency-experiment-tracker"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n업무 맥락에서 '테스트'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?\n\nProposed gold_skills (ordered if composition):\n[\"agency-experiment-tracker\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0162",
      "setting": "single_skill",
      "query": "I need to comprehensive code review—which skill should handle this?",
      "gold_skills": [
        "deep-review"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to comprehensive code review—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"deep-review\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0163",
      "setting": "single_skill",
      "query": "Can you help me batch email from sheet data? Pick the most specific matching skill.",
      "gold_skills": [
        "gws-recipe-personalized-emails"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me batch email from sheet data? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"gws-recipe-personalized-emails\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0164",
      "setting": "single_skill",
      "query": "I'm trying to documenting product decisions; route me to the right capability.",
      "gold_skills": [
        "kwp-product-management-feature-spec"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to documenting product decisions; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-product-management-feature-spec\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0165",
      "setting": "single_skill",
      "query": "User request: seed missions. What single skill applies?",
      "gold_skills": [
        "pm-sentinel"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: seed missions. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"pm-sentinel\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0166",
      "setting": "single_skill",
      "query": "Looking for automation that covers: make a slide deck.",
      "gold_skills": [
        "notebooklm-studio"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: make a slide deck.\n\nProposed gold_skills (ordered if composition):\n[\"notebooklm-studio\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0167",
      "setting": "single_skill",
      "query": "Match this intent to one skill: bias prevention.",
      "gold_skills": [
        "trading-backtest-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: bias prevention.\n\nProposed gold_skills (ordered if composition):\n[\"trading-backtest-expert\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0168",
      "setting": "single_skill",
      "query": "Task description: evolving patterns into reusable skills. Identify the canonical skill id.",
      "gold_skills": [
        "ecc-continuous-learning"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: evolving patterns into reusable skills. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"ecc-continuous-learning\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0169",
      "setting": "single_skill",
      "query": "I need to Scientific Problem Selection Skills—which skill should handle this?",
      "gold_skills": [
        "kwp-bio-research-scientific-problem-selection"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to Scientific Problem Selection Skills—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-bio-research-scientific-problem-selection\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0170",
      "setting": "single_skill",
      "query": "Can you help me assumption testing? Pick the most specific matching skill.",
      "gold_skills": [
        "nlm-discovery-lab"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me assumption testing? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"nlm-discovery-lab\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0171",
      "setting": "single_skill",
      "query": "I'm trying to /morning; route me to the right capability.",
      "gold_skills": [
        "morning"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to /morning; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"morning\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0172",
      "setting": "single_skill",
      "query": "User request: 高配当で成長している小型株を探したい. What single skill applies?",
      "gold_skills": [
        "trading-finviz-screener"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: 高配当で成長している小型株を探したい. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"trading-finviz-screener\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0173",
      "setting": "single_skill",
      "query": "Looking for automation that covers: opening hook.",
      "gold_skills": [
        "presentation-strategist"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: opening hook.\n\nProposed gold_skills (ordered if composition):\n[\"presentation-strategist\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0174",
      "setting": "single_skill",
      "query": "Match this intent to one skill: sync all projects.",
      "gold_skills": [
        "sod-ship"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: sync all projects.\n\nProposed gold_skills (ordered if composition):\n[\"sod-ship\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0175",
      "setting": "single_skill",
      "query": "Task description: references agency-data-consolidation-agent. Identify the canonical skill id.",
      "gold_skills": [
        "agency-data-consolidation-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: references agency-data-consolidation-agent. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"agency-data-consolidation-agent\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0176",
      "setting": "single_skill",
      "query": "한국어로 '빌드' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "agency-twitter-engager"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '빌드' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"agency-twitter-engager\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0177",
      "setting": "single_skill",
      "query": "Can you help me complex finance logic flows as diagrams? Pick the most specific matching skill.",
      "gold_skills": [
        "alphaear-logic-visualizer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me complex finance logic flows as diagrams? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-logic-visualizer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0178",
      "setting": "single_skill",
      "query": "I'm trying to Enrich Lead; route me to the right capability.",
      "gold_skills": [
        "kwp-apollo-enrich-lead"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to Enrich Lead; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-apollo-enrich-lead\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0179",
      "setting": "single_skill",
      "query": "User request: agency-roster. What single skill applies?",
      "gold_skills": [
        "agency-roster"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: agency-roster. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"agency-roster\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0180",
      "setting": "single_skill",
      "query": "Looking for automation that covers: developer onboarding.",
      "gold_skills": [
        "onboarding-accelerator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: developer onboarding.\n\nProposed gold_skills (ordered if composition):\n[\"onboarding-accelerator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0181",
      "setting": "single_skill",
      "query": "업무 맥락에서 '아키텍처'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?",
      "gold_skills": [
        "kwp-engineering-system-design"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n업무 맥락에서 '아키텍처'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-engineering-system-design\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0182",
      "setting": "single_skill",
      "query": "Task description: Headline Scenario Analyzer. Identify the canonical skill id.",
      "gold_skills": [
        "trading-scenario-analyzer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: Headline Scenario Analyzer. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"trading-scenario-analyzer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0183",
      "setting": "single_skill",
      "query": "I need to generate a doc with pre-filled content—which skill should handle this?",
      "gold_skills": [
        "gws-recipe-doc-from-template"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to generate a doc with pre-filled content—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"gws-recipe-doc-from-template\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0184",
      "setting": "single_skill",
      "query": "Can you help me convert hypotheses to research tickets? Pick the most specific matching skill.",
      "gold_skills": [
        "trading-edge-candidate-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me convert hypotheses to research tickets? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"trading-edge-candidate-agent\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0185",
      "setting": "single_skill",
      "query": "I'm trying to RFP response; route me to the right capability.",
      "gold_skills": [
        "proposal-factory"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to RFP response; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"proposal-factory\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0186",
      "setting": "single_skill",
      "query": "한국어로 '스킬 시작' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "sp-core"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '스킬 시작' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"sp-core\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0187",
      "setting": "single_skill",
      "query": "Looking for automation that covers: proactive-meeting-scheduler.",
      "gold_skills": [
        "proactive-meeting-scheduler"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: proactive-meeting-scheduler.\n\nProposed gold_skills (ordered if composition):\n[\"proactive-meeting-scheduler\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0188",
      "setting": "single_skill",
      "query": "Match this intent to one skill: Web Application Testing.",
      "gold_skills": [
        "anthropic-webapp-testing"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: Web Application Testing.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-webapp-testing\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0189",
      "setting": "single_skill",
      "query": "Task description: refreshing DualMA. Identify the canonical skill id.",
      "gold_skills": [
        "tab-dualma-refresh"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: refreshing DualMA. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"tab-dualma-refresh\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0190",
      "setting": "single_skill",
      "query": "I need to Theme Factory Skill—which skill should handle this?",
      "gold_skills": [
        "anthropic-theme-factory"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to Theme Factory Skill—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-theme-factory\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0191",
      "setting": "single_skill",
      "query": "Can you help me Overlay 패턴 가이드? Pick the most specific matching skill.",
      "gold_skills": [
        "overlay-layout-patterns"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me Overlay 패턴 가이드? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"overlay-layout-patterns\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0192",
      "setting": "single_skill",
      "query": "I'm trying to Brand Discovery; route me to the right capability.",
      "gold_skills": [
        "kwp-brand-voice-discover-brand"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to Brand Discovery; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-brand-voice-discover-brand\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0193",
      "setting": "single_skill",
      "query": "User request: terraform-reviewer. What single skill applies?",
      "gold_skills": [
        "terraform-reviewer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: terraform-reviewer. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"terraform-reviewer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0194",
      "setting": "single_skill",
      "query": "Looking for automation that covers: 회의 다이제스트.",
      "gold_skills": [
        "meeting-digest"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: 회의 다이제스트.\n\nProposed gold_skills (ordered if composition):\n[\"meeting-digest\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0195",
      "setting": "single_skill",
      "query": "Match this intent to one skill: before executing implementation plans - creates isolated git worktrees with smart directory selection and safety veri....",
      "gold_skills": [
        "sp-git-worktrees"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: before executing implementation plans - creates isolated git worktrees with smart directory selection and safety veri....\n\nProposed gold_skills (ordered if composition):\n[\"sp-git-worktrees\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0196",
      "setting": "single_skill",
      "query": "업무 맥락에서 '주식'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?",
      "gold_skills": [
        "daily-stock-check"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n업무 맥락에서 '주식'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?\n\nProposed gold_skills (ordered if composition):\n[\"daily-stock-check\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0197",
      "setting": "single_skill",
      "query": "I need to What domain or topic needs more context? What queries are failing or producing wrong results?—which skill should handle this?",
      "gold_skills": [
        "kwp-data-data-context-extractor"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to What domain or topic needs more context? What queries are failing or producing wrong results?—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-data-data-context-extractor\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0198",
      "setting": "single_skill",
      "query": "Can you help me activate the Senior Developer agent persona? Pick the most specific matching skill.",
      "gold_skills": [
        "agency-senior-developer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me activate the Senior Developer agent persona? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"agency-senior-developer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0199",
      "setting": "single_skill",
      "query": "I'm trying to activate the Visual Storyteller agent persona; route me to the right capability.",
      "gold_skills": [
        "agency-visual-storyteller"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to activate the Visual Storyteller agent persona; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"agency-visual-storyteller\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0200",
      "setting": "single_skill",
      "query": "User request: create a PR on an HF repo. What single skill applies?",
      "gold_skills": [
        "hf-discussions"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: create a PR on an HF repo. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"hf-discussions\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0201",
      "setting": "single_skill",
      "query": "다음 요청과 맞는 워크플로우를 찾아줘: 계획 실행 관련.",
      "gold_skills": [
        "sp-executing-plans"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n다음 요청과 맞는 워크플로우를 찾아줘: 계획 실행 관련.\n\nProposed gold_skills (ordered if composition):\n[\"sp-executing-plans\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0202",
      "setting": "single_skill",
      "query": "Match this intent to one skill: generating testing workpapers.",
      "gold_skills": [
        "kwp-finance-audit-support"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: generating testing workpapers.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-finance-audit-support\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0203",
      "setting": "single_skill",
      "query": "Task description: 인사 분석. Identify the canonical skill id.",
      "gold_skills": [
        "role-hr"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: 인사 분석. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"role-hr\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0204",
      "setting": "single_skill",
      "query": "I need to when \"CSO perspective—which skill should handle this?",
      "gold_skills": [
        "role-cso"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to when \"CSO perspective—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"role-cso\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0205",
      "setting": "single_skill",
      "query": "Can you help me 동시 실행? Pick the most specific matching skill.",
      "gold_skills": [
        "workflow-parallel"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me 동시 실행? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"workflow-parallel\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0206",
      "setting": "single_skill",
      "query": "다음 요청과 맞는 워크플로우를 찾아줘: 빌드 관련.",
      "gold_skills": [
        "kwp-brand-voice-guideline-generation"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n다음 요청과 맞는 워크플로우를 찾아줘: 빌드 관련.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-brand-voice-guideline-generation\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0207",
      "setting": "single_skill",
      "query": "User request: 품질 반복. What single skill applies?",
      "gold_skills": [
        "workflow-eval-opt"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: 품질 반복. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"workflow-eval-opt\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0208",
      "setting": "single_skill",
      "query": "Looking for automation that covers: encountering a new dataset.",
      "gold_skills": [
        "kwp-data-data-exploration"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: encountering a new dataset.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-data-data-exploration\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0209",
      "setting": "single_skill",
      "query": "Match this intent to one skill: code-review-all.",
      "gold_skills": [
        "code-review-all"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: code-review-all.\n\nProposed gold_skills (ordered if composition):\n[\"code-review-all\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0210",
      "setting": "single_skill",
      "query": "Task description: references agency-executive-summary-generator. Identify the canonical skill id.",
      "gold_skills": [
        "agency-executive-summary-generator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: references agency-executive-summary-generator. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"agency-executive-summary-generator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0211",
      "setting": "single_skill",
      "query": "I need to checking market environment—which skill should handle this?",
      "gold_skills": [
        "tab-market-environment"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to checking market environment—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"tab-market-environment\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0212",
      "setting": "single_skill",
      "query": "Can you help me compliance audits? Pick the most specific matching skill.",
      "gold_skills": [
        "compliance-governance"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me compliance audits? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"compliance-governance\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0213",
      "setting": "single_skill",
      "query": "I'm trying to portfolio report; route me to the right capability.",
      "gold_skills": [
        "portfolio-report-generator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to portfolio report; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"portfolio-report-generator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0214",
      "setting": "single_skill",
      "query": "User request: automate this process. What single skill applies?",
      "gold_skills": [
        "pipeline-builder"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: automate this process. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"pipeline-builder\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0215",
      "setting": "single_skill",
      "query": "Looking for automation that covers: design a plugin. This skill requires Cowork mode with access to the outputs directory for delivering the final .plugi....",
      "gold_skills": [
        "kwp-cowork-create-plugin"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: design a plugin. This skill requires Cowork mode with access to the outputs directory for delivering the final .plugi....\n\nProposed gold_skills (ordered if composition):\n[\"kwp-cowork-create-plugin\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0216",
      "setting": "single_skill",
      "query": "한국어로 '모델' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "kwp-operations-risk-assessment"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '모델' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-operations-risk-assessment\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0217",
      "setting": "single_skill",
      "query": "Task description: generate weekly report. Identify the canonical skill id.",
      "gold_skills": [
        "weekly-status-report"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: generate weekly report. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"weekly-status-report\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0218",
      "setting": "single_skill",
      "query": "I need to frontend quality review—which skill should handle this?",
      "gold_skills": [
        "ui-suite"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to frontend quality review—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"ui-suite\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0219",
      "setting": "single_skill",
      "query": "Can you help me beachhead segment? Pick the most specific matching skill.",
      "gold_skills": [
        "pm-go-to-market"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me beachhead segment? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"pm-go-to-market\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0220",
      "setting": "single_skill",
      "query": "I'm trying to managing glossary; route me to the right capability.",
      "gold_skills": [
        "tab-glossary"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to managing glossary; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"tab-glossary\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0221",
      "setting": "single_skill",
      "query": "User request: wants automated standup reports. What single skill applies?",
      "gold_skills": [
        "standup-digest"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: wants automated standup reports. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"standup-digest\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0222",
      "setting": "single_skill",
      "query": "Looking for automation that covers: references agency-infrastructure-maintainer.",
      "gold_skills": [
        "agency-infrastructure-maintainer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: references agency-infrastructure-maintainer.\n\nProposed gold_skills (ordered if composition):\n[\"agency-infrastructure-maintainer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0223",
      "setting": "single_skill",
      "query": "Match this intent to one skill: Sequence Load.",
      "gold_skills": [
        "kwp-apollo-sequence-load"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: Sequence Load.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-apollo-sequence-load\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0224",
      "setting": "single_skill",
      "query": "Task description: assessing cross-border data transfer requirements. Identify the canonical skill id.",
      "gold_skills": [
        "kwp-legal-compliance"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: assessing cross-border data transfer requirements. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-legal-compliance\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0225",
      "setting": "single_skill",
      "query": "I need to piece of art—which skill should handle this?",
      "gold_skills": [
        "anthropic-canvas-design"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to piece of art—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-canvas-design\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0226",
      "setting": "single_skill",
      "query": "업무 맥락에서 '빌드'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?",
      "gold_skills": [
        "kwp-common-room-prospect"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n업무 맥락에서 '빌드'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-common-room-prospect\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0227",
      "setting": "single_skill",
      "query": "I'm trying to automation strategy; route me to the right capability.",
      "gold_skills": [
        "automation-strategist"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to automation strategy; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"automation-strategist\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0228",
      "setting": "single_skill",
      "query": "User request: hot or cold themes. What single skill applies?",
      "gold_skills": [
        "trading-theme-detector"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: hot or cold themes. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"trading-theme-detector\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0229",
      "setting": "single_skill",
      "query": "Looking for automation that covers: activate the Image Prompt Engineer agent persona.",
      "gold_skills": [
        "agency-image-prompt-engineer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: activate the Image Prompt Engineer agent persona.\n\nProposed gold_skills (ordered if composition):\n[\"agency-image-prompt-engineer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0230",
      "setting": "single_skill",
      "query": "Match this intent to one skill: Create an Asset.",
      "gold_skills": [
        "kwp-sales-create-an-asset"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: Create an Asset.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-sales-create-an-asset\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0231",
      "setting": "single_skill",
      "query": "Task description: sprint digest. Identify the canonical skill id.",
      "gold_skills": [
        "github-sprint-digest"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: sprint digest. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"github-sprint-digest\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0232",
      "setting": "single_skill",
      "query": "I need to refreshing breadth—which skill should handle this?",
      "gold_skills": [
        "tab-market-breadth"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to refreshing breadth—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"tab-market-breadth\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0233",
      "setting": "single_skill",
      "query": "Can you help me 세타 매매? Pick the most specific matching skill.",
      "gold_skills": [
        "trading-options-theta"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me 세타 매매? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"trading-options-theta\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0234",
      "setting": "single_skill",
      "query": "I'm trying to clean up my code; route me to the right capability.",
      "gold_skills": [
        "simplify"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to clean up my code; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"simplify\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0235",
      "setting": "single_skill",
      "query": "User request: Instrument Data to Allotrope Converter. What single skill applies?",
      "gold_skills": [
        "kwp-bio-research-instrument-data-to-allotrope"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: Instrument Data to Allotrope Converter. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-bio-research-instrument-data-to-allotrope\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0236",
      "setting": "single_skill",
      "query": "다음 요청과 맞는 워크플로우를 찾아줘: 슬랙 앱 관련.",
      "gold_skills": [
        "slack-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n다음 요청과 맞는 워크플로우를 찾아줘: 슬랙 앱 관련.\n\nProposed gold_skills (ordered if composition):\n[\"slack-agent\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0237",
      "setting": "single_skill",
      "query": "Match this intent to one skill: 세션에서 스킬 추출.",
      "gold_skills": [
        "autoskill-extractor"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: 세션에서 스킬 추출.\n\nProposed gold_skills (ordered if composition):\n[\"autoskill-extractor\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0238",
      "setting": "single_skill",
      "query": "Task description: when deciding whether an NDA needs full counsel review. Identify the canonical skill id.",
      "gold_skills": [
        "kwp-legal-nda-triage"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: when deciding whether an NDA needs full counsel review. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-legal-nda-triage\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0239",
      "setting": "single_skill",
      "query": "I need to dashboards—which skill should handle this?",
      "gold_skills": [
        "anthropic-frontend-design"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to dashboards—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-frontend-design\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0240",
      "setting": "single_skill",
      "query": "Can you help me fetching news? Pick the most specific matching skill.",
      "gold_skills": [
        "tab-news-fetch"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me fetching news? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"tab-news-fetch\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0241",
      "setting": "single_skill",
      "query": "I'm trying to GitHub 워크플로우 자동화; route me to the right capability.",
      "gold_skills": [
        "github-workflow-automation"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to GitHub 워크플로우 자동화; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"github-workflow-automation\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0242",
      "setting": "single_skill",
      "query": "User request: responding to customer tickets. What single skill applies?",
      "gold_skills": [
        "kwp-customer-support-response-drafting"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: responding to customer tickets. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-customer-support-response-drafting\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0243",
      "setting": "single_skill",
      "query": "Looking for automation that covers: when \"CTO perspective.",
      "gold_skills": [
        "role-cto"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: when \"CTO perspective.\n\nProposed gold_skills (ordered if composition):\n[\"role-cto\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0244",
      "setting": "single_skill",
      "query": "Match this intent to one skill: activate the Support Responder agent persona.",
      "gold_skills": [
        "agency-support-responder"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: activate the Support Responder agent persona.\n\nProposed gold_skills (ordered if composition):\n[\"agency-support-responder\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0245",
      "setting": "single_skill",
      "query": "Task description: 세션 기반 스킬 진화. Identify the canonical skill id.",
      "gold_skills": [
        "autoskill-evolve"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: 세션 기반 스킬 진화. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"autoskill-evolve\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0246",
      "setting": "single_skill",
      "query": "I need to 인프라 코드 리뷰—which skill should handle this?",
      "gold_skills": [
        "iac-review-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to 인프라 코드 리뷰—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"iac-review-agent\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0247",
      "setting": "single_skill",
      "query": "Can you help me optimizing slow SQL? Pick the most specific matching skill.",
      "gold_skills": [
        "kwp-data-sql-queries"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me optimizing slow SQL? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-data-sql-queries\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0248",
      "setting": "single_skill",
      "query": "I'm trying to find trending AI research; route me to the right capability.",
      "gold_skills": [
        "hf-papers"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to find trending AI research; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"hf-papers\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0249",
      "setting": "single_skill",
      "query": "User request: activate the Whimsy Injector agent persona. What single skill applies?",
      "gold_skills": [
        "agency-whimsy-injector"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: activate the Whimsy Injector agent persona. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"agency-whimsy-injector\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0250",
      "setting": "single_skill",
      "query": "Looking for automation that covers: upload markdown to Notion.",
      "gold_skills": [
        "md-to-notion"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: upload markdown to Notion.\n\nProposed gold_skills (ordered if composition):\n[\"md-to-notion\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0251",
      "setting": "single_skill",
      "query": "한국어로 '코드 스타일' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "ecc-coding-standards"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '코드 스타일' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"ecc-coding-standards\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0252",
      "setting": "single_skill",
      "query": "Task description: recovering from context window limits. Identify the canonical skill id.",
      "gold_skills": [
        "ecc-autonomous-loops"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: recovering from context window limits. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"ecc-autonomous-loops\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0253",
      "setting": "single_skill",
      "query": "I need to assessing data quality—which skill should handle this?",
      "gold_skills": [
        "kwp-bio-research-single-cell-rna-qc"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to assessing data quality—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-bio-research-single-cell-rna-qc\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0254",
      "setting": "single_skill",
      "query": "Can you help me 회의록 동기화? Pick the most specific matching skill.",
      "gold_skills": [
        "notion-meeting-sync"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me 회의록 동기화? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"notion-meeting-sync\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0255",
      "setting": "single_skill",
      "query": "I'm trying to Task Management; route me to the right capability.",
      "gold_skills": [
        "kwp-productivity-task-management"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to Task Management; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-productivity-task-management\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0256",
      "setting": "single_skill",
      "query": "한국어로 '설계' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "agency-xr-interface-architect"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '설계' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"agency-xr-interface-architect\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0257",
      "setting": "single_skill",
      "query": "Looking for automation that covers: error models.",
      "gold_skills": [
        "backend-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: error models.\n\nProposed gold_skills (ordered if composition):\n[\"backend-expert\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0258",
      "setting": "single_skill",
      "query": "Match this intent to one skill: before implementing suggestions.",
      "gold_skills": [
        "sp-receive-review"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: before implementing suggestions.\n\nProposed gold_skills (ordered if composition):\n[\"sp-receive-review\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0259",
      "setting": "single_skill",
      "query": "Task description: check test coverage. Identify the canonical skill id.",
      "gold_skills": [
        "test-suite"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: check test coverage. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"test-suite\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0260",
      "setting": "single_skill",
      "query": "I need to schedule a daily digest—which skill should handle this?",
      "gold_skills": [
        "cursor-automations"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to schedule a daily digest—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"cursor-automations\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0261",
      "setting": "single_skill",
      "query": "업무 맥락에서 '검색'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?",
      "gold_skills": [
        "kwp-enterprise-search-source-management"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n업무 맥락에서 '검색'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-enterprise-search-source-management\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0262",
      "setting": "single_skill",
      "query": "I'm trying to triage email; route me to the right capability.",
      "gold_skills": [
        "gmail-daily-triage"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to triage email; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"gmail-daily-triage\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0263",
      "setting": "single_skill",
      "query": "User request: wants a comprehensive pre-release validation. What single skill applies?",
      "gold_skills": [
        "release-commander"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: wants a comprehensive pre-release validation. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"release-commander\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0264",
      "setting": "single_skill",
      "query": "Looking for automation that covers: activate the Project Shepherd agent persona.",
      "gold_skills": [
        "agency-project-shepherd"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: activate the Project Shepherd agent persona.\n\nProposed gold_skills (ordered if composition):\n[\"agency-project-shepherd\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0265",
      "setting": "single_skill",
      "query": "Match this intent to one skill: running flux analysis.",
      "gold_skills": [
        "kwp-finance-financial-statements"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: running flux analysis.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-finance-financial-statements\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0266",
      "setting": "single_skill",
      "query": "업무 맥락에서 '스킬'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?",
      "gold_skills": [
        "agency-lsp-index-engineer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n업무 맥락에서 '스킬'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?\n\nProposed gold_skills (ordered if composition):\n[\"agency-lsp-index-engineer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0267",
      "setting": "single_skill",
      "query": "I need to Screen Description Generator—which skill should handle this?",
      "gold_skills": [
        "screen-description"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to Screen Description Generator—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"screen-description\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0268",
      "setting": "single_skill",
      "query": "Can you help me write an ADR? Pick the most specific matching skill.",
      "gold_skills": [
        "technical-writer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me write an ADR? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"technical-writer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0269",
      "setting": "single_skill",
      "query": "I'm trying to automating form interactions; route me to the right capability.",
      "gold_skills": [
        "playwright-runner"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to automating form interactions; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"playwright-runner\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0270",
      "setting": "single_skill",
      "query": "User request: push and create PR. What single skill applies?",
      "gold_skills": [
        "release-ship"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: push and create PR. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"release-ship\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0271",
      "setting": "single_skill",
      "query": "업무 맥락에서 '스킬'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?",
      "gold_skills": [
        "agency-brand-guardian"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n업무 맥락에서 '스킬'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?\n\nProposed gold_skills (ordered if composition):\n[\"agency-brand-guardian\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0272",
      "setting": "single_skill",
      "query": "Match this intent to one skill: create a Google Doc.",
      "gold_skills": [
        "gws-docs"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: create a Google Doc.\n\nProposed gold_skills (ordered if composition):\n[\"gws-docs\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0273",
      "setting": "single_skill",
      "query": "Task description: 보안 검사. Identify the canonical skill id.",
      "gold_skills": [
        "semantic-guard"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: 보안 검사. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"semantic-guard\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0274",
      "setting": "single_skill",
      "query": "I need to cognee search—which skill should handle this?",
      "gold_skills": [
        "cognee"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to cognee search—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"cognee\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0275",
      "setting": "single_skill",
      "query": "Can you help me review resume? Pick the most specific matching skill.",
      "gold_skills": [
        "pm-toolkit"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me review resume? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"pm-toolkit\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0276",
      "setting": "single_skill",
      "query": "한국어로 '리뷰' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "agency-wechat-official-account-manager"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '리뷰' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"agency-wechat-official-account-manager\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0277",
      "setting": "single_skill",
      "query": "User request: Compose Outreach. What single skill applies?",
      "gold_skills": [
        "kwp-common-room-compose-outreach"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: Compose Outreach. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-common-room-compose-outreach\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0278",
      "setting": "single_skill",
      "query": "Looking for automation that covers: Knowledge Synthesis.",
      "gold_skills": [
        "kwp-enterprise-search-knowledge-synthesis"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: Knowledge Synthesis.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-enterprise-search-knowledge-synthesis\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0279",
      "setting": "single_skill",
      "query": "Match this intent to one skill: scvi-tools Deep Learning Skill.",
      "gold_skills": [
        "kwp-bio-research-scvi-tools"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: scvi-tools Deep Learning Skill.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-bio-research-scvi-tools\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0280",
      "setting": "single_skill",
      "query": "Task description: lead lifecycle. Identify the canonical skill id.",
      "gold_skills": [
        "lead-lifecycle-manager"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: lead lifecycle. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"lead-lifecycle-manager\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0281",
      "setting": "single_skill",
      "query": "I need to strategy comparison—which skill should handle this?",
      "gold_skills": [
        "tab-strategy-comparison"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to strategy comparison—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"tab-strategy-comparison\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0282",
      "setting": "single_skill",
      "query": "Can you help me cross-role summary? Pick the most specific matching skill.",
      "gold_skills": [
        "executive-briefing"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me cross-role summary? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"executive-briefing\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0283",
      "setting": "single_skill",
      "query": "I'm trying to 논문 슬랙 공유; route me to the right capability.",
      "gold_skills": [
        "auto-research-distribute"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to 논문 슬랙 공유; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"auto-research-distribute\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0284",
      "setting": "single_skill",
      "query": "User request: executing implementation plans with independent tasks in the current session. What single skill applies?",
      "gold_skills": [
        "sp-subagent-dev"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: executing implementation plans with independent tasks in the current session. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"sp-subagent-dev\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0285",
      "setting": "single_skill",
      "query": "Looking for automation that covers: user experience impact.",
      "gold_skills": [
        "role-ux-designer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: user experience impact.\n\nProposed gold_skills (ordered if composition):\n[\"role-ux-designer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0286",
      "setting": "single_skill",
      "query": "Match this intent to one skill: academic paper review.",
      "gold_skills": [
        "paper-review"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: academic paper review.\n\nProposed gold_skills (ordered if composition):\n[\"paper-review\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0287",
      "setting": "single_skill",
      "query": "Task description: encountering any bug. Identify the canonical skill id.",
      "gold_skills": [
        "sp-debugging"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: encountering any bug. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"sp-debugging\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0288",
      "setting": "single_skill",
      "query": "I need to FOMC 영향—which skill should handle this?",
      "gold_skills": [
        "trading-market-news-analyst"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to FOMC 영향—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"trading-market-news-analyst\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0289",
      "setting": "single_skill",
      "query": "Can you help me flow fields? Pick the most specific matching skill.",
      "gold_skills": [
        "anthropic-algorithmic-art"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me flow fields? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-algorithmic-art\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0290",
      "setting": "single_skill",
      "query": "I'm trying to customer agreements; route me to the right capability.",
      "gold_skills": [
        "kwp-legal-contract-review"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to customer agreements; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-legal-contract-review\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0291",
      "setting": "single_skill",
      "query": "User request: when assessing severity. What single skill applies?",
      "gold_skills": [
        "kwp-customer-support-ticket-triage"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: when assessing severity. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-customer-support-ticket-triage\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0292",
      "setting": "single_skill",
      "query": "Looking for automation that covers: references agency-xr-cockpit-interaction-specialist.",
      "gold_skills": [
        "agency-xr-cockpit-interaction-specialist"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: references agency-xr-cockpit-interaction-specialist.\n\nProposed gold_skills (ordered if composition):\n[\"agency-xr-cockpit-interaction-specialist\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0293",
      "setting": "single_skill",
      "query": "Match this intent to one skill: references agency-xiaohongshu-specialist.",
      "gold_skills": [
        "agency-xiaohongshu-specialist"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: references agency-xiaohongshu-specialist.\n\nProposed gold_skills (ordered if composition):\n[\"agency-xiaohongshu-specialist\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0294",
      "setting": "single_skill",
      "query": "Task description: preparing variance commentary for leadership. Identify the canonical skill id.",
      "gold_skills": [
        "kwp-finance-variance-analysis"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: preparing variance commentary for leadership. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-finance-variance-analysis\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0295",
      "setting": "single_skill",
      "query": "I need to creating PRs - requires running verification commands and confirming output before making any success claims; evidenc...—which skill should handle this?",
      "gold_skills": [
        "sp-verification"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to creating PRs - requires running verification commands and confirming output before making any success claims; evidenc...—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"sp-verification\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0296",
      "setting": "single_skill",
      "query": "다음 요청과 맞는 워크플로우를 찾아줘: 빌드 관련.",
      "gold_skills": [
        "kwp-data-interactive-dashboard-builder"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n다음 요청과 맞는 워크플로우를 찾아줘: 빌드 관련.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-data-interactive-dashboard-builder\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0297",
      "setting": "single_skill",
      "query": "I'm trying to activate the Senior Project Manager agent persona; route me to the right capability.",
      "gold_skills": [
        "agency-senior-project-manager"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to activate the Senior Project Manager agent persona; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"agency-senior-project-manager\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0298",
      "setting": "single_skill",
      "query": "User request: 회의 예약. What single skill applies?",
      "gold_skills": [
        "smart-meeting-scheduler"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: 회의 예약. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"smart-meeting-scheduler\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0299",
      "setting": "single_skill",
      "query": "Looking for automation that covers: find an API.",
      "gold_skills": [
        "public-apis"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: find an API.\n\nProposed gold_skills (ordered if composition):\n[\"public-apis\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0300",
      "setting": "single_skill",
      "query": "Match this intent to one skill: 모닝 오케스트레이터.",
      "gold_skills": [
        "daily-am-orchestrator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: 모닝 오케스트레이터.\n\nProposed gold_skills (ordered if composition):\n[\"daily-am-orchestrator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0301",
      "setting": "single_skill",
      "query": "다음 요청과 맞는 워크플로우를 찾아줘: 체크 관련.",
      "gold_skills": [
        "agency-legal-compliance-checker"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n다음 요청과 맞는 워크플로우를 찾아줘: 체크 관련.\n\nProposed gold_skills (ordered if composition):\n[\"agency-legal-compliance-checker\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0302",
      "setting": "single_skill",
      "query": "I need to push cursor config to other repos—which skill should handle this?",
      "gold_skills": [
        "cursor-sync"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to push cursor config to other repos—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"cursor-sync\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0303",
      "setting": "single_skill",
      "query": "Can you help me 하루 마무리? Pick the most specific matching skill.",
      "gold_skills": [
        "eod-ship"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me 하루 마무리? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"eod-ship\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0304",
      "setting": "single_skill",
      "query": "I'm trying to A/B test analysis; route me to the right capability.",
      "gold_skills": [
        "pm-data-analytics"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to A/B test analysis; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"pm-data-analytics\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0305",
      "setting": "single_skill",
      "query": "User request: requirements to PRD. What single skill applies?",
      "gold_skills": [
        "prd-research-factory"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: requirements to PRD. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"prd-research-factory\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0306",
      "setting": "single_skill",
      "query": "한국어로 '리뷰' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "agency-behavioral-nudge-engine"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '리뷰' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"agency-behavioral-nudge-engine\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0307",
      "setting": "single_skill",
      "query": "Match this intent to one skill: detecting anomalies.",
      "gold_skills": [
        "kwp-data-statistical-analysis"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: detecting anomalies.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-data-statistical-analysis\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0308",
      "setting": "single_skill",
      "query": "Task description: AI Platform FE - FSD 개발 및 마이그레이션. Identify the canonical skill id.",
      "gold_skills": [
        "fsd-development"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: AI Platform FE - FSD 개발 및 마이그레이션. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"fsd-development\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0309",
      "setting": "single_skill",
      "query": "I need to Competitive Intelligence—which skill should handle this?",
      "gold_skills": [
        "kwp-sales-competitive-intelligence"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to Competitive Intelligence—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-sales-competitive-intelligence\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0310",
      "setting": "single_skill",
      "query": "Can you help me activate the Accessibility Auditor agent persona? Pick the most specific matching skill.",
      "gold_skills": [
        "agency-accessibility-auditor"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me activate the Accessibility Auditor agent persona? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"agency-accessibility-auditor\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0311",
      "setting": "single_skill",
      "query": "업무 맥락에서 '데이터'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?",
      "gold_skills": [
        "kwp-finance-reconciliation"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n업무 맥락에서 '데이터'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-finance-reconciliation\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0312",
      "setting": "single_skill",
      "query": "User request: Org Planning. What single skill applies?",
      "gold_skills": [
        "kwp-human-resources-org-planning"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: Org Planning. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-human-resources-org-planning\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0313",
      "setting": "single_skill",
      "query": "Looking for automation that covers: recent price changes.",
      "gold_skills": [
        "alphaear-stock"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: recent price changes.\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-stock\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0314",
      "setting": "single_skill",
      "query": "Match this intent to one skill: creating financial training datasets (sentiment labels.",
      "gold_skills": [
        "hf-datasets"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: creating financial training datasets (sentiment labels.\n\nProposed gold_skills (ordered if composition):\n[\"hf-datasets\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0315",
      "setting": "single_skill",
      "query": "Task description: 페이퍼클립 설치. Identify the canonical skill id.",
      "gold_skills": [
        "paperclip-setup"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: 페이퍼클립 설치. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"paperclip-setup\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0316",
      "setting": "single_skill",
      "query": "I need to is the market topping—which skill should handle this?",
      "gold_skills": [
        "trading-market-top-detector"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to is the market topping—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"trading-market-top-detector\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0317",
      "setting": "single_skill",
      "query": "Can you help me scrapling? Pick the most specific matching skill.",
      "gold_skills": [
        "scrapling"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me scrapling? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"scrapling\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0318",
      "setting": "single_skill",
      "query": "I'm trying to 시나리오 예측; route me to the right capability.",
      "gold_skills": [
        "mirofish"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to 시나리오 예측; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"mirofish\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0319",
      "setting": "single_skill",
      "query": "User request: adding a dependency. What single skill applies?",
      "gold_skills": [
        "ecc-search-first"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: adding a dependency. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"ecc-search-first\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0320",
      "setting": "single_skill",
      "query": "Looking for automation that covers: activate the Trend Researcher agent persona.",
      "gold_skills": [
        "agency-trend-researcher"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: activate the Trend Researcher agent persona.\n\nProposed gold_skills (ordered if composition):\n[\"agency-trend-researcher\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0321",
      "setting": "single_skill",
      "query": "한국어로 '문서화' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "kwp-engineering-documentation"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '문서화' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-engineering-documentation\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0322",
      "setting": "single_skill",
      "query": "Task description: Write internal communications using company formats. Use for 3P updates (Progress, Plans, Problems), company newsletters, FAQ responses, status reports, leadership updates, project updates, incident r…. Identify the canonical skill id.",
      "gold_skills": [
        "anthropic-internal-comms"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: Write internal communications using company formats. Use for 3P updates (Progress, Plans, Problems), company newsletters, FAQ responses, status reports, leadership updates, project updates, incident r…. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-internal-comms\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0323",
      "setting": "single_skill",
      "query": "I need to User Research—which skill should handle this?",
      "gold_skills": [
        "kwp-design-user-research"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to User Research—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-design-user-research\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0324",
      "setting": "single_skill",
      "query": "Can you help me analyze feature requests? Pick the most specific matching skill.",
      "gold_skills": [
        "pm-product-discovery"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me analyze feature requests? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"pm-product-discovery\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0325",
      "setting": "single_skill",
      "query": "I'm trying to activate the Agentic Identity Trust Architect agent persona; route me to the right capability.",
      "gold_skills": [
        "agency-agentic-identity-trust-architect"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to activate the Agentic Identity Trust Architect agent persona; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"agency-agentic-identity-trust-architect\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0326",
      "setting": "single_skill",
      "query": "User request: tab-bollinger-refresh. What single skill applies?",
      "gold_skills": [
        "tab-bollinger-refresh"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: tab-bollinger-refresh. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"tab-bollinger-refresh\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0327",
      "setting": "single_skill",
      "query": "Looking for automation that covers: AlphaEar Sentiment.",
      "gold_skills": [
        "alphaear-sentiment"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: AlphaEar Sentiment.\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-sentiment\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0328",
      "setting": "single_skill",
      "query": "Match this intent to one skill: FASTQ analysis.",
      "gold_skills": [
        "kwp-bio-research-nextflow-development"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: FASTQ analysis.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-bio-research-nextflow-development\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0329",
      "setting": "single_skill",
      "query": "Task description: requirements for a multi-step task. Identify the canonical skill id.",
      "gold_skills": [
        "sp-writing-plans"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: requirements for a multi-step task. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"sp-writing-plans\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0330",
      "setting": "single_skill",
      "query": "I need to sequential pipeline—which skill should handle this?",
      "gold_skills": [
        "workflow-sequential"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to sequential pipeline—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"workflow-sequential\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0331",
      "setting": "single_skill",
      "query": "업무 맥락에서 '트위터 타임라인'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?",
      "gold_skills": [
        "twitter-timeline-to-slack"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n업무 맥락에서 '트위터 타임라인'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?\n\nProposed gold_skills (ordered if composition):\n[\"twitter-timeline-to-slack\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0332",
      "setting": "single_skill",
      "query": "I'm trying to arxiv to slides; route me to the right capability.",
      "gold_skills": [
        "nlm-arxiv-slides"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to arxiv to slides; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"nlm-arxiv-slides\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0333",
      "setting": "single_skill",
      "query": "User request: SEFO 오케스트레이션. What single skill applies?",
      "gold_skills": [
        "sefo-orchestrator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: SEFO 오케스트레이션. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"sefo-orchestrator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0334",
      "setting": "single_skill",
      "query": "Looking for automation that covers: update stock prices from investing.com.",
      "gold_skills": [
        "stock-csv-downloader"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: update stock prices from investing.com.\n\nProposed gold_skills (ordered if composition):\n[\"stock-csv-downloader\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0335",
      "setting": "single_skill",
      "query": "Match this intent to one skill: activate the Ux Researcher agent persona.",
      "gold_skills": [
        "agency-ux-researcher"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: activate the Ux Researcher agent persona.\n\nProposed gold_skills (ordered if composition):\n[\"agency-ux-researcher\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0336",
      "setting": "single_skill",
      "query": "한국어로 'API' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "agency-api-tester"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 'API' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"agency-api-tester\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0337",
      "setting": "single_skill",
      "query": "I need to background research—which skill should handle this?",
      "gold_skills": [
        "kwp-legal-meeting-briefing"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to background research—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-legal-meeting-briefing\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0338",
      "setting": "single_skill",
      "query": "Can you help me x-to-slack) detects content that requires a decision? Pick the most specific matching skill.",
      "gold_skills": [
        "decision-router"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me x-to-slack) detects content that requires a decision? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"decision-router\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0339",
      "setting": "single_skill",
      "query": "I'm trying to any manual journal entry; route me to the right capability.",
      "gold_skills": [
        "kwp-finance-journal-entry-prep"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to any manual journal entry; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-finance-journal-entry-prep\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0340",
      "setting": "single_skill",
      "query": "User request: product area. What single skill applies?",
      "gold_skills": [
        "kwp-customer-support-customer-research"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: product area. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-customer-support-customer-research\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0341",
      "setting": "single_skill",
      "query": "Looking for automation that covers: 논문 레이더.",
      "gold_skills": [
        "paper-auto-classifier"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: 논문 레이더.\n\nProposed gold_skills (ordered if composition):\n[\"paper-auto-classifier\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0342",
      "setting": "single_skill",
      "query": "Match this intent to one skill: manage extensions.",
      "gold_skills": [
        "hf-hub"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: manage extensions.\n\nProposed gold_skills (ordered if composition):\n[\"hf-hub\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0343",
      "setting": "single_skill",
      "query": "Task description: find best model for task. Identify the canonical skill id.",
      "gold_skills": [
        "hf-model-arena"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: find best model for task. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"hf-model-arena\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0344",
      "setting": "single_skill",
      "query": "I need to activate the Mobile App Builder agent persona—which skill should handle this?",
      "gold_skills": [
        "agency-mobile-app-builder"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to activate the Mobile App Builder agent persona—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"agency-mobile-app-builder\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0345",
      "setting": "single_skill",
      "query": "Can you help me Call Prep? Pick the most specific matching skill.",
      "gold_skills": [
        "kwp-sales-call-prep"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me Call Prep? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-sales-call-prep\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0346",
      "setting": "single_skill",
      "query": "I'm trying to 주식 데이터 동기화; route me to the right capability.",
      "gold_skills": [
        "tab-stock-sync"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to 주식 데이터 동기화; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"tab-stock-sync\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0347",
      "setting": "single_skill",
      "query": "User request: tab-technical-analysis. What single skill applies?",
      "gold_skills": [
        "tab-technical-analysis"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: tab-technical-analysis. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"tab-technical-analysis\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0348",
      "setting": "single_skill",
      "query": "Looking for automation that covers: before writing implementation code.",
      "gold_skills": [
        "sp-tdd"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: before writing implementation code.\n\nProposed gold_skills (ordered if composition):\n[\"sp-tdd\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0349",
      "setting": "single_skill",
      "query": "Match this intent to one skill: condensing finance analysis into a structured output.",
      "gold_skills": [
        "alphaear-reporter"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: condensing finance analysis into a structured output.\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-reporter\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0350",
      "setting": "single_skill",
      "query": "Task description: references agency-ui-designer. Identify the canonical skill id.",
      "gold_skills": [
        "agency-ui-designer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: references agency-ui-designer. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"agency-ui-designer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0351",
      "setting": "single_skill",
      "query": "다음 요청과 맞는 워크플로우를 찾아줘: 최적화 관련.",
      "gold_skills": [
        "agency-growth-hacker"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n다음 요청과 맞는 워크플로우를 찾아줘: 최적화 관련.\n\nProposed gold_skills (ordered if composition):\n[\"agency-growth-hacker\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0352",
      "setting": "single_skill",
      "query": "Can you help me references agency-studio-operations? Pick the most specific matching skill.",
      "gold_skills": [
        "agency-studio-operations"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me references agency-studio-operations? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"agency-studio-operations\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0353",
      "setting": "single_skill",
      "query": "I'm trying to Design Critique; route me to the right capability.",
      "gold_skills": [
        "kwp-design-design-critique"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to Design Critique; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-design-design-critique\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0354",
      "setting": "single_skill",
      "query": "User request: Incident Response. What single skill applies?",
      "gold_skills": [
        "kwp-engineering-incident-response"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: Incident Response. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-engineering-incident-response\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0355",
      "setting": "single_skill",
      "query": "Looking for automation that covers: 미팅 준비.",
      "gold_skills": [
        "ai-chief-of-staff"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: 미팅 준비.\n\nProposed gold_skills (ordered if composition):\n[\"ai-chief-of-staff\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0356",
      "setting": "single_skill",
      "query": "한국어로 'DS 분석' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "role-data-scientist"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 'DS 분석' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"role-data-scientist\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0357",
      "setting": "single_skill",
      "query": "Task description: completing tasks. Identify the canonical skill id.",
      "gold_skills": [
        "sp-request-review"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: completing tasks. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"sp-request-review\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0358",
      "setting": "single_skill",
      "query": "I need to run daily AI radar—which skill should handle this?",
      "gold_skills": [
        "hf-trending-intelligence"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to run daily AI radar—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"hf-trending-intelligence\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0359",
      "setting": "single_skill",
      "query": "Can you help me running weekly metrics reviews? Pick the most specific matching skill.",
      "gold_skills": [
        "kwp-product-management-metrics-tracking"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me running weekly metrics reviews? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-product-management-metrics-tracking\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0360",
      "setting": "single_skill",
      "query": "I'm trying to Prospect; route me to the right capability.",
      "gold_skills": [
        "kwp-apollo-prospect"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to Prospect; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-apollo-prospect\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0361",
      "setting": "single_skill",
      "query": "다음 요청과 맞는 워크플로우를 찾아줘: 로드맵 관련.",
      "gold_skills": [
        "kwp-product-management-roadmap-management"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n다음 요청과 맞는 워크플로우를 찾아줘: 로드맵 관련.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-product-management-roadmap-management\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0362",
      "setting": "single_skill",
      "query": "Looking for automation that covers: dependency radar.",
      "gold_skills": [
        "dependency-radar"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: dependency radar.\n\nProposed gold_skills (ordered if composition):\n[\"dependency-radar\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0363",
      "setting": "single_skill",
      "query": "Match this intent to one skill: Interview Prep.",
      "gold_skills": [
        "kwp-human-resources-interview-prep"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: Interview Prep.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-human-resources-interview-prep\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0364",
      "setting": "single_skill",
      "query": "Task description: 옵션 전략. Identify the canonical skill id.",
      "gold_skills": [
        "trading-options-strategy-advisor"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: 옵션 전략. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"trading-options-strategy-advisor\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0365",
      "setting": "single_skill",
      "query": "I need to 올리브영 재고—which skill should handle this?",
      "gold_skills": [
        "daiso-mcp"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to 올리브영 재고—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"daiso-mcp\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0366",
      "setting": "single_skill",
      "query": "한국어로 '리뷰' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "visual-explainer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '리뷰' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"visual-explainer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0367",
      "setting": "single_skill",
      "query": "I'm trying to 대시보드 요약; route me to the right capability.",
      "gold_skills": [
        "tab-dashboard-summary"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to 대시보드 요약; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"tab-dashboard-summary\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0368",
      "setting": "single_skill",
      "query": "User request: 온프레미스 파이프라인. What single skill applies?",
      "gold_skills": [
        "air-gap-orchestrator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: 온프레미스 파이프라인. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"air-gap-orchestrator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0369",
      "setting": "single_skill",
      "query": "Looking for automation that covers: Weekly Prep Brief.",
      "gold_skills": [
        "kwp-common-room-weekly-prep-brief"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: Weekly Prep Brief.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-common-room-weekly-prep-brief\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0370",
      "setting": "single_skill",
      "query": "Match this intent to one skill: mental models for.",
      "gold_skills": [
        "nlm-deep-learn"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: mental models for.\n\nProposed gold_skills (ordered if composition):\n[\"nlm-deep-learn\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0371",
      "setting": "single_skill",
      "query": "다음 요청과 맞는 워크플로우를 찾아줘: HF 스페이스 관련.",
      "gold_skills": [
        "hf-spaces"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n다음 요청과 맞는 워크플로우를 찾아줘: HF 스페이스 관련.\n\nProposed gold_skills (ordered if composition):\n[\"hf-spaces\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0372",
      "setting": "single_skill",
      "query": "I need to Design System Guide — Refined Swiss—which skill should handle this?",
      "gold_skills": [
        "kwp-design-system"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to Design System Guide — Refined Swiss—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-design-system\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0373",
      "setting": "single_skill",
      "query": "Can you help me query notebook? Pick the most specific matching skill.",
      "gold_skills": [
        "notebooklm"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me query notebook? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"notebooklm\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0374",
      "setting": "single_skill",
      "query": "I'm trying to video size reduction; route me to the right capability.",
      "gold_skills": [
        "video-compress"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to video size reduction; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"video-compress\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0375",
      "setting": "single_skill",
      "query": "User request: related work. What single skill applies?",
      "gold_skills": [
        "related-papers-scout"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: related work. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"related-papers-scout\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0376",
      "setting": "single_skill",
      "query": "업무 맥락에서 '리뷰'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?",
      "gold_skills": [
        "pr-review-captain"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n업무 맥락에서 '리뷰'에 가까운 작업을 처리하려고 해. 어떤 스킬이 맞을까?\n\nProposed gold_skills (ordered if composition):\n[\"pr-review-captain\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0377",
      "setting": "single_skill",
      "query": "Match this intent to one skill: Change Management.",
      "gold_skills": [
        "kwp-operations-change-management"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: Change Management.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-operations-change-management\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0378",
      "setting": "single_skill",
      "query": "Task description: run autoskill pipeline. Identify the canonical skill id.",
      "gold_skills": [
        "autoskill-pipeline"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: run autoskill pipeline. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"autoskill-pipeline\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0379",
      "setting": "single_skill",
      "query": "I need to references agency-performance-benchmarker—which skill should handle this?",
      "gold_skills": [
        "agency-performance-benchmarker"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to references agency-performance-benchmarker—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"agency-performance-benchmarker\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0380",
      "setting": "single_skill",
      "query": "Can you help me \"find the bug? Pick the most specific matching skill.",
      "gold_skills": [
        "diagnose"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me \"find the bug? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"diagnose\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0381",
      "setting": "single_skill",
      "query": "다음 요청과 맞는 워크플로우를 찾아줘: 생성 관련.",
      "gold_skills": [
        "i18n-sync"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n다음 요청과 맞는 워크플로우를 찾아줘: 생성 관련.\n\nProposed gold_skills (ordered if composition):\n[\"i18n-sync\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0382",
      "setting": "single_skill",
      "query": "User request: 기능 요청 접수. What single skill applies?",
      "gold_skills": [
        "feature-intake-router"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: 기능 요청 접수. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"feature-intake-router\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0383",
      "setting": "single_skill",
      "query": "Looking for automation that covers: kill criteria for trades.",
      "gold_skills": [
        "trading-trade-hypothesis-ideator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: kill criteria for trades.\n\nProposed gold_skills (ordered if composition):\n[\"trading-trade-hypothesis-ideator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0384",
      "setting": "single_skill",
      "query": "Match this intent to one skill: security compliance scan.",
      "gold_skills": [
        "compliance-gate"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: security compliance scan.\n\nProposed gold_skills (ordered if composition):\n[\"compliance-gate\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0385",
      "setting": "single_skill",
      "query": "Task description: references agency-instagram-curator. Identify the canonical skill id.",
      "gold_skills": [
        "agency-instagram-curator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: references agency-instagram-curator. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"agency-instagram-curator\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0386",
      "setting": "single_skill",
      "query": "I need to needs an enterprise-grade plan spanning 20-50 skills across survey—which skill should handle this?",
      "gold_skills": [
        "full-stack-planner"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to needs an enterprise-grade plan spanning 20-50 skills across survey—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"full-stack-planner\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0387",
      "setting": "single_skill",
      "query": "Can you help me latency problems? Pick the most specific matching skill.",
      "gold_skills": [
        "performance-profiler"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me latency problems? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"performance-profiler\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0388",
      "setting": "single_skill",
      "query": "I'm trying to 재무 데이터 검증; route me to the right capability.",
      "gold_skills": [
        "trading-data-quality-checker"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to 재무 데이터 검증; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"trading-data-quality-checker\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0389",
      "setting": "single_skill",
      "query": "User request: documenting a brand voice. What single skill applies?",
      "gold_skills": [
        "kwp-marketing-brand-voice"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUser request: documenting a brand voice. What single skill applies?\n\nProposed gold_skills (ordered if composition):\n[\"kwp-marketing-brand-voice\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0390",
      "setting": "single_skill",
      "query": "Looking for automation that covers: auto digest meetings.",
      "gold_skills": [
        "planning-meeting-sentinel"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: auto digest meetings.\n\nProposed gold_skills (ordered if composition):\n[\"planning-meeting-sentinel\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0391",
      "setting": "single_skill",
      "query": "한국어로 '검색' 키워드로 시작하는 작업을 맡기고 싶어.",
      "gold_skills": [
        "x-to-slack"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n한국어로 '검색' 키워드로 시작하는 작업을 맡기고 싶어.\n\nProposed gold_skills (ordered if composition):\n[\"x-to-slack\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0392",
      "setting": "single_skill",
      "query": "Task description: stale doc check. Identify the canonical skill id.",
      "gold_skills": [
        "docs-freshness-guardian"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: stale doc check. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"docs-freshness-guardian\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0393",
      "setting": "single_skill",
      "query": "I need to evaluate skill—which skill should handle this?",
      "gold_skills": [
        "skill-optimizer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to evaluate skill—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"skill-optimizer\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0394",
      "setting": "single_skill",
      "query": "Can you help me develop a plugin\" for Cowork? Pick the most specific matching skill.",
      "gold_skills": [
        "kwp-cowork-create-cowork-plugin"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCan you help me develop a plugin\" for Cowork? Pick the most specific matching skill.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-cowork-create-cowork-plugin\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0395",
      "setting": "single_skill",
      "query": "I'm trying to ship it; route me to the right capability.",
      "gold_skills": [
        "ship"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI'm trying to ship it; route me to the right capability.\n\nProposed gold_skills (ordered if composition):\n[\"ship\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0396",
      "setting": "single_skill",
      "query": "다음 요청과 맞는 워크플로우를 찾아줘: 주식 관련.",
      "gold_skills": [
        "weekly-stock-update"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\n다음 요청과 맞는 워크플로우를 찾아줘: 주식 관련.\n\nProposed gold_skills (ordered if composition):\n[\"weekly-stock-update\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0397",
      "setting": "single_skill",
      "query": "Looking for automation that covers: find bottlenecks.",
      "gold_skills": [
        "system-thinker"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nLooking for automation that covers: find bottlenecks.\n\nProposed gold_skills (ordered if composition):\n[\"system-thinker\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0398",
      "setting": "single_skill",
      "query": "Match this intent to one skill: Compensation Benchmarking.",
      "gold_skills": [
        "kwp-human-resources-compensation-benchmarking"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMatch this intent to one skill: Compensation Benchmarking.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-human-resources-compensation-benchmarking\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0399",
      "setting": "single_skill",
      "query": "Task description: user asks to use Claude API. Identify the canonical skill id.",
      "gold_skills": [
        "anthropic-claude-api"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nTask description: user asks to use Claude API. Identify the canonical skill id.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-claude-api\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "SS-0400",
      "setting": "single_skill",
      "query": "I need to refreshing Turtle indicators—which skill should handle this?",
      "gold_skills": [
        "tab-turtle-refresh"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "easy",
      "source": "trigger_paraphrase",
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI need to refreshing Turtle indicators—which skill should handle this?\n\nProposed gold_skills (ordered if composition):\n[\"tab-turtle-refresh\"]\n\nSetting: single_skill\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0001",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Air-Gap Orchestrator, then Google Calendar, Mission Control, then GWS Email Reply. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "air-gap-orchestrator",
        "gws-calendar",
        "mission-control",
        "gws-email-reply"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "gws-calendar",
          "dependent": "air-gap-orchestrator"
        },
        {
          "prerequisite": "mission-control",
          "dependent": "air-gap-orchestrator"
        },
        {
          "prerequisite": "gws-email-reply",
          "dependent": "air-gap-orchestrator"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Air-Gap Orchestrator, then Google Calendar, Mission Control, then GWS Email Reply. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"air-gap-orchestrator\", \"gws-calendar\", \"mission-control\", \"gws-email-reply\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0002",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like tab-strategy-comparison, Daily Stock Check, and tab-stock-sync in order.",
      "gold_skills": [
        "tab-strategy-comparison",
        "daily-stock-check",
        "tab-stock-sync"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "daily-stock-check",
          "dependent": "tab-strategy-comparison"
        },
        {
          "prerequisite": "tab-stock-sync",
          "dependent": "tab-strategy-comparison"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like tab-strategy-comparison, Daily Stock Check, and tab-stock-sync in order.\n\nProposed gold_skills (ordered if composition):\n[\"tab-strategy-comparison\", \"daily-stock-check\", \"tab-stock-sync\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0003",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with tab-screening, then tab-hot-stock-discovery, Daily Stock Check, then tab-technical-analysis. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "tab-screening",
        "tab-hot-stock-discovery",
        "daily-stock-check",
        "tab-technical-analysis"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "tab-hot-stock-discovery",
          "dependent": "tab-screening"
        },
        {
          "prerequisite": "daily-stock-check",
          "dependent": "tab-screening"
        },
        {
          "prerequisite": "tab-technical-analysis",
          "dependent": "tab-screening"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with tab-screening, then tab-hot-stock-discovery, Daily Stock Check, then tab-technical-analysis. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"tab-screening\", \"tab-hot-stock-discovery\", \"daily-stock-check\", \"tab-technical-analysis\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0004",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “NotebookLM Studio: Content Generation”-style work, orchestrated under “NotebookLM Research: Web & Drive Discovery”. Which ordered skill chain fits?",
      "gold_skills": [
        "notebooklm-research",
        "notebooklm-studio"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "notebooklm-studio",
          "dependent": "notebooklm-research"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “NotebookLM Studio: Content Generation”-style work, orchestrated under “NotebookLM Research: Web & Drive Discovery”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"notebooklm-research\", \"notebooklm-studio\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0005",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like md-to-slack-canvas, Slack Messaging Best Practices, and md-to-notion in order.",
      "gold_skills": [
        "md-to-slack-canvas",
        "kwp-slack-slack-messaging",
        "md-to-notion"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "kwp-slack-slack-messaging",
          "dependent": "md-to-slack-canvas"
        },
        {
          "prerequisite": "md-to-notion",
          "dependent": "md-to-slack-canvas"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like md-to-slack-canvas, Slack Messaging Best Practices, and md-to-notion in order.\n\nProposed gold_skills (ordered if composition):\n[\"md-to-slack-canvas\", \"kwp-slack-slack-messaging\", \"md-to-notion\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0006",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with tab-market-breadth, then tab-turtle-refresh, tab-market-environment, then tab-stock-sync. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "tab-market-breadth",
        "tab-turtle-refresh",
        "tab-market-environment",
        "tab-stock-sync"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "tab-turtle-refresh",
          "dependent": "tab-market-breadth"
        },
        {
          "prerequisite": "tab-market-environment",
          "dependent": "tab-market-breadth"
        },
        {
          "prerequisite": "tab-stock-sync",
          "dependent": "tab-market-breadth"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with tab-market-breadth, then tab-turtle-refresh, tab-market-environment, then tab-stock-sync. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"tab-market-breadth\", \"tab-turtle-refresh\", \"tab-market-environment\", \"tab-stock-sync\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0007",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with tab-analysis-run, then tab-screening, tab-technical-analysis, then tab-event-detect. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "tab-analysis-run",
        "tab-screening",
        "tab-technical-analysis",
        "tab-event-detect"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "tab-screening",
          "dependent": "tab-analysis-run"
        },
        {
          "prerequisite": "tab-technical-analysis",
          "dependent": "tab-analysis-run"
        },
        {
          "prerequisite": "tab-event-detect",
          "dependent": "tab-analysis-run"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with tab-analysis-run, then tab-screening, tab-technical-analysis, then tab-event-detect. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"tab-analysis-run\", \"tab-screening\", \"tab-technical-analysis\", \"tab-event-detect\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0008",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with MiroFish Financial Scenario Simulation, then AlphaEar Predictor, Market Environment Analysis, then AlphaEar Sentiment. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "mirofish-financial-sim",
        "alphaear-predictor",
        "trading-market-environment-analysis",
        "alphaear-sentiment"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "alphaear-predictor",
          "dependent": "mirofish-financial-sim"
        },
        {
          "prerequisite": "trading-market-environment-analysis",
          "dependent": "mirofish-financial-sim"
        },
        {
          "prerequisite": "alphaear-sentiment",
          "dependent": "mirofish-financial-sim"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with MiroFish Financial Scenario Simulation, then AlphaEar Predictor, Market Environment Analysis, then AlphaEar Sentiment. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"mirofish-financial-sim\", \"alphaear-predictor\", \"trading-market-environment-analysis\", \"alphaear-sentiment\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0009",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “NotebookLM Studio: Content Generation”-style work, orchestrated under “NotebookLM: Core Notebook Management”. Which ordered skill chain fits?",
      "gold_skills": [
        "notebooklm",
        "notebooklm-studio"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "notebooklm-studio",
          "dependent": "notebooklm"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “NotebookLM Studio: Content Generation”-style work, orchestrated under “NotebookLM: Core Notebook Management”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"notebooklm\", \"notebooklm-studio\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0010",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “github-sprint-digest”-style work, orchestrated under “Weekly Status Report Generator”. Which ordered skill chain fits?",
      "gold_skills": [
        "weekly-status-report",
        "github-sprint-digest"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "github-sprint-digest",
          "dependent": "weekly-status-report"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “github-sprint-digest”-style work, orchestrated under “Weekly Status Report Generator”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"weekly-status-report\", \"github-sprint-digest\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0011",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Performance Profiler, then Backend Expert, SRE / DevOps Expert, then DB Expert. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "performance-profiler",
        "backend-expert",
        "sre-devops-expert",
        "db-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "backend-expert",
          "dependent": "performance-profiler"
        },
        {
          "prerequisite": "sre-devops-expert",
          "dependent": "performance-profiler"
        },
        {
          "prerequisite": "db-expert",
          "dependent": "performance-profiler"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Performance Profiler, then Backend Expert, SRE / DevOps Expert, then DB Expert. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"performance-profiler\", \"backend-expert\", \"sre-devops-expert\", \"db-expert\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0012",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Hugging Face Repos”-style work, orchestrated under “Hugging Face Collections”. Which ordered skill chain fits?",
      "gold_skills": [
        "hf-collections",
        "hf-repos"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "hf-repos",
          "dependent": "hf-collections"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Hugging Face Repos”-style work, orchestrated under “Hugging Face Collections”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"hf-collections\", \"hf-repos\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0013",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Backend Expert, then SRE / DevOps Expert, DB Expert, then Frontend Expert. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "backend-expert",
        "sre-devops-expert",
        "db-expert",
        "frontend-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "sre-devops-expert",
          "dependent": "backend-expert"
        },
        {
          "prerequisite": "db-expert",
          "dependent": "backend-expert"
        },
        {
          "prerequisite": "frontend-expert",
          "dependent": "backend-expert"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Backend Expert, then SRE / DevOps Expert, DB Expert, then Frontend Expert. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"backend-expert\", \"sre-devops-expert\", \"db-expert\", \"frontend-expert\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0014",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like DOCX creation, editing, and analysis, PDF Processing Guide, and PPTX Skill in order.",
      "gold_skills": [
        "anthropic-docx",
        "anthropic-pdf",
        "anthropic-pptx"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "anthropic-pdf",
          "dependent": "anthropic-docx"
        },
        {
          "prerequisite": "anthropic-pptx",
          "dependent": "anthropic-docx"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like DOCX creation, editing, and analysis, PDF Processing Guide, and PPTX Skill in order.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-docx\", \"anthropic-pdf\", \"anthropic-pptx\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0015",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Coding Standards & Best Practices, Frontend Expert, and Backend Expert in order.",
      "gold_skills": [
        "ecc-coding-standards",
        "frontend-expert",
        "backend-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "frontend-expert",
          "dependent": "ecc-coding-standards"
        },
        {
          "prerequisite": "backend-expert",
          "dependent": "ecc-coding-standards"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Coding Standards & Best Practices, Frontend Expert, and Backend Expert in order.\n\nProposed gold_skills (ordered if composition):\n[\"ecc-coding-standards\", \"frontend-expert\", \"backend-expert\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0016",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Gmail, then Google Drive, Google Calendar, then Google Workspace CLI. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "gws-gmail",
        "gws-drive",
        "gws-calendar",
        "gws-workspace"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "gws-drive",
          "dependent": "gws-gmail"
        },
        {
          "prerequisite": "gws-calendar",
          "dependent": "gws-gmail"
        },
        {
          "prerequisite": "gws-workspace",
          "dependent": "gws-gmail"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Gmail, then Google Drive, Google Calendar, then Google Workspace CLI. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"gws-gmail\", \"gws-drive\", \"gws-calendar\", \"gws-workspace\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0017",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “MiroFish Orchestrator”-style work, orchestrated under “MiroFish Graph Explorer”. Which ordered skill chain fits?",
      "gold_skills": [
        "mirofish-graph-explorer",
        "mirofish"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "mirofish",
          "dependent": "mirofish-graph-explorer"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “MiroFish Orchestrator”-style work, orchestrated under “MiroFish Graph Explorer”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"mirofish-graph-explorer\", \"mirofish\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0018",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Breadth Chart Analyst”-style work, orchestrated under “Market Breadth Analyzer Skill”. Which ordered skill chain fits?",
      "gold_skills": [
        "trading-market-breadth-analyzer",
        "trading-breadth-chart-analyst"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "trading-breadth-chart-analyst",
          "dependent": "trading-market-breadth-analyzer"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Breadth Chart Analyst”-style work, orchestrated under “Market Breadth Analyzer Skill”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"trading-market-breadth-analyzer\", \"trading-breadth-chart-analyst\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0019",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Twitter Timeline to Slack Pipeline, Decision Router, and Scrapling in order.",
      "gold_skills": [
        "twitter-timeline-to-slack",
        "decision-router",
        "scrapling"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "decision-router",
          "dependent": "twitter-timeline-to-slack"
        },
        {
          "prerequisite": "scrapling",
          "dependent": "twitter-timeline-to-slack"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Twitter Timeline to Slack Pipeline, Decision Router, and Scrapling in order.\n\nProposed gold_skills (ordered if composition):\n[\"twitter-timeline-to-slack\", \"decision-router\", \"scrapling\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0020",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “NotebookLM: Core Notebook Management”-style work, orchestrated under “NLM arXiv Slides: arXiv Paper to Expert Slide Deck Pipeline”. Which ordered skill chain fits?",
      "gold_skills": [
        "nlm-arxiv-slides",
        "notebooklm"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "notebooklm",
          "dependent": "nlm-arxiv-slides"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “NotebookLM: Core Notebook Management”-style work, orchestrated under “NLM arXiv Slides: arXiv Paper to Expert Slide Deck Pipeline”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"nlm-arxiv-slides\", \"notebooklm\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0021",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like AI Quality Evaluator, Evals Skills, and DOCX creation, editing, and analysis in order.",
      "gold_skills": [
        "ai-quality-evaluator",
        "evals-skills",
        "anthropic-docx"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "evals-skills",
          "dependent": "ai-quality-evaluator"
        },
        {
          "prerequisite": "anthropic-docx",
          "dependent": "ai-quality-evaluator"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like AI Quality Evaluator, Evals Skills, and DOCX creation, editing, and analysis in order.\n\nProposed gold_skills (ordered if composition):\n[\"ai-quality-evaluator\", \"evals-skills\", \"anthropic-docx\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0022",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like KWP Sync, Enrich Lead, and Skill Optimizer in order.",
      "gold_skills": [
        "kwp-sync",
        "kwp-apollo-enrich-lead",
        "skill-optimizer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "kwp-apollo-enrich-lead",
          "dependent": "kwp-sync"
        },
        {
          "prerequisite": "skill-optimizer",
          "dependent": "kwp-sync"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like KWP Sync, Enrich Lead, and Skill Optimizer in order.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-sync\", \"kwp-apollo-enrich-lead\", \"skill-optimizer\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0023",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Docs Freshness Guardian, Deep Review, and Technical Writer in order.",
      "gold_skills": [
        "docs-freshness-guardian",
        "deep-review",
        "technical-writer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "deep-review",
          "dependent": "docs-freshness-guardian"
        },
        {
          "prerequisite": "technical-writer",
          "dependent": "docs-freshness-guardian"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Docs Freshness Guardian, Deep Review, and Technical Writer in order.\n\nProposed gold_skills (ordered if composition):\n[\"docs-freshness-guardian\", \"deep-review\", \"technical-writer\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0024",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Google Docs, Google Drive, and Google Workspace CLI in order.",
      "gold_skills": [
        "gws-docs",
        "gws-drive",
        "gws-workspace"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "gws-drive",
          "dependent": "gws-docs"
        },
        {
          "prerequisite": "gws-workspace",
          "dependent": "gws-docs"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Google Docs, Google Drive, and Google Workspace CLI in order.\n\nProposed gold_skills (ordered if composition):\n[\"gws-docs\", \"gws-drive\", \"gws-workspace\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0025",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Infrastructure Runbook Generator, DOCX creation, editing, and analysis, and Incident to Improvement in order.",
      "gold_skills": [
        "infra-runbook-generator",
        "anthropic-docx",
        "incident-to-improvement"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "anthropic-docx",
          "dependent": "infra-runbook-generator"
        },
        {
          "prerequisite": "incident-to-improvement",
          "dependent": "infra-runbook-generator"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Infrastructure Runbook Generator, DOCX creation, editing, and analysis, and Incident to Improvement in order.\n\nProposed gold_skills (ordered if composition):\n[\"infra-runbook-generator\", \"anthropic-docx\", \"incident-to-improvement\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0026",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Domain-Split Commit”-style work, orchestrated under “Issue-to-Code Draft Generator”. Which ordered skill chain fits?",
      "gold_skills": [
        "issue-to-code",
        "domain-commit"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "domain-commit",
          "dependent": "issue-to-code"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Domain-Split Commit”-style work, orchestrated under “Issue-to-Code Draft Generator”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"issue-to-code\", \"domain-commit\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0027",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Google Workspace Workflows, Gmail, and Google Workspace CLI in order.",
      "gold_skills": [
        "gws-workflows",
        "gws-gmail",
        "gws-workspace"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "gws-gmail",
          "dependent": "gws-workflows"
        },
        {
          "prerequisite": "gws-workspace",
          "dependent": "gws-workflows"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Google Workspace Workflows, Gmail, and Google Workspace CLI in order.\n\nProposed gold_skills (ordered if composition):\n[\"gws-workflows\", \"gws-gmail\", \"gws-workspace\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0028",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Semantic Guard, Security Expert, and Dependency Auditor in order.",
      "gold_skills": [
        "semantic-guard",
        "security-expert",
        "dependency-auditor"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "security-expert",
          "dependent": "semantic-guard"
        },
        {
          "prerequisite": "dependency-auditor",
          "dependent": "semantic-guard"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Semantic Guard, Security Expert, and Dependency Auditor in order.\n\nProposed gold_skills (ordered if composition):\n[\"semantic-guard\", \"security-expert\", \"dependency-auditor\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0029",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Edge Candidate Agent, Daily Stock Check, and AlphaEar Signal Tracker in order.",
      "gold_skills": [
        "trading-edge-candidate-agent",
        "daily-stock-check",
        "alphaear-signal-tracker"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "daily-stock-check",
          "dependent": "trading-edge-candidate-agent"
        },
        {
          "prerequisite": "alphaear-signal-tracker",
          "dependent": "trading-edge-candidate-agent"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Edge Candidate Agent, Daily Stock Check, and AlphaEar Signal Tracker in order.\n\nProposed gold_skills (ordered if composition):\n[\"trading-edge-candidate-agent\", \"daily-stock-check\", \"alphaear-signal-tracker\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0030",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Refactor Simulator, Deep Review, and Simplify in order.",
      "gold_skills": [
        "refactor-simulator",
        "deep-review",
        "simplify"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "deep-review",
          "dependent": "refactor-simulator"
        },
        {
          "prerequisite": "simplify",
          "dependent": "refactor-simulator"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Refactor Simulator, Deep Review, and Simplify in order.\n\nProposed gold_skills (ordered if composition):\n[\"refactor-simulator\", \"deep-review\", \"simplify\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0031",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Technical Writer, PR Review Captain, and Prompt Transformer in order.",
      "gold_skills": [
        "technical-writer",
        "pr-review-captain",
        "prompt-transformer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "pr-review-captain",
          "dependent": "technical-writer"
        },
        {
          "prerequisite": "prompt-transformer",
          "dependent": "technical-writer"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Technical Writer, PR Review Captain, and Prompt Transformer in order.\n\nProposed gold_skills (ordered if composition):\n[\"technical-writer\", \"pr-review-captain\", \"prompt-transformer\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0032",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Hugging Face Spaces, then Hugging Face Repos, Running Workloads on Hugging Face Jobs, then Hugging Face Inference Endpoints. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "hf-spaces",
        "hf-repos",
        "hf-jobs",
        "hf-endpoints"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "hf-repos",
          "dependent": "hf-spaces"
        },
        {
          "prerequisite": "hf-jobs",
          "dependent": "hf-spaces"
        },
        {
          "prerequisite": "hf-endpoints",
          "dependent": "hf-spaces"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Hugging Face Spaces, then Hugging Face Repos, Running Workloads on Hugging Face Jobs, then Hugging Face Inference Endpoints. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"hf-spaces\", \"hf-repos\", \"hf-jobs\", \"hf-endpoints\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0033",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Semantic Guard”-style work, orchestrated under “AutoSkill Pipeline”. Which ordered skill chain fits?",
      "gold_skills": [
        "autoskill-pipeline",
        "semantic-guard"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "semantic-guard",
          "dependent": "autoskill-pipeline"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Semantic Guard”-style work, orchestrated under “AutoSkill Pipeline”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"autoskill-pipeline\", \"semantic-guard\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0034",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with NLM Strategy Hub: Product Strategy Research Hub Pipeline, then PM Product Strategy, NLM Discovery Lab: Product Discovery Lab Pipeline, then NotebookLM: Core Notebook Management. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "nlm-strategy-hub",
        "pm-product-strategy",
        "nlm-discovery-lab",
        "notebooklm"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "pm-product-strategy",
          "dependent": "nlm-strategy-hub"
        },
        {
          "prerequisite": "nlm-discovery-lab",
          "dependent": "nlm-strategy-hub"
        },
        {
          "prerequisite": "notebooklm",
          "dependent": "nlm-strategy-hub"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with NLM Strategy Hub: Product Strategy Research Hub Pipeline, then PM Product Strategy, NLM Discovery Lab: Product Discovery Lab Pipeline, then NotebookLM: Core Notebook Management. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"nlm-strategy-hub\", \"pm-product-strategy\", \"nlm-discovery-lab\", \"notebooklm\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0035",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Google Daily Automation, Google Workspace CLI, and Decision Router in order.",
      "gold_skills": [
        "google-daily",
        "gws-workspace",
        "decision-router"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "gws-workspace",
          "dependent": "google-daily"
        },
        {
          "prerequisite": "decision-router",
          "dependent": "google-daily"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Google Daily Automation, Google Workspace CLI, and Decision Router in order.\n\nProposed gold_skills (ordered if composition):\n[\"google-daily\", \"gws-workspace\", \"decision-router\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0036",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Morning Ship, then Calendar Daily Briefing, Google Daily Automation, then Gmail Daily Triage. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "morning-ship",
        "calendar-daily-briefing",
        "google-daily",
        "gmail-daily-triage"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "calendar-daily-briefing",
          "dependent": "morning-ship"
        },
        {
          "prerequisite": "google-daily",
          "dependent": "morning-ship"
        },
        {
          "prerequisite": "gmail-daily-triage",
          "dependent": "morning-ship"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Morning Ship, then Calendar Daily Briefing, Google Daily Automation, then Gmail Daily Triage. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"morning-ship\", \"calendar-daily-briefing\", \"google-daily\", \"gmail-daily-triage\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0037",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Scrapling, then Defuddle, E2E Testing Skill, then Agent Browser. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "scrapling",
        "defuddle",
        "e2e-testing",
        "agent-browser"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "defuddle",
          "dependent": "scrapling"
        },
        {
          "prerequisite": "e2e-testing",
          "dependent": "scrapling"
        },
        {
          "prerequisite": "agent-browser",
          "dependent": "scrapling"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Scrapling, then Defuddle, E2E Testing Skill, then Agent Browser. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"scrapling\", \"defuddle\", \"e2e-testing\", \"agent-browser\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0038",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with CTO Perspective Analyzer, then Workflow Miner, Backend Expert, then Deep Review. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "role-cto",
        "workflow-miner",
        "backend-expert",
        "deep-review"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "workflow-miner",
          "dependent": "role-cto"
        },
        {
          "prerequisite": "backend-expert",
          "dependent": "role-cto"
        },
        {
          "prerequisite": "deep-review",
          "dependent": "role-cto"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with CTO Perspective Analyzer, then Workflow Miner, Backend Expert, then Deep Review. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"role-cto\", \"workflow-miner\", \"backend-expert\", \"deep-review\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0039",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Figma → TDS Code Generator, Overlay 패턴 가이드, and Screen Description Generator in order.",
      "gold_skills": [
        "figma-to-tds",
        "overlay-layout-patterns",
        "screen-description"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "overlay-layout-patterns",
          "dependent": "figma-to-tds"
        },
        {
          "prerequisite": "screen-description",
          "dependent": "figma-to-tds"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Figma → TDS Code Generator, Overlay 패턴 가이드, and Screen Description Generator in order.\n\nProposed gold_skills (ordered if composition):\n[\"figma-to-tds\", \"overlay-layout-patterns\", \"screen-description\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0040",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like QA / Test Expert, E2E Testing Skill, and CI Quality Gate in order.",
      "gold_skills": [
        "qa-test-expert",
        "e2e-testing",
        "ci-quality-gate"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "e2e-testing",
          "dependent": "qa-test-expert"
        },
        {
          "prerequisite": "ci-quality-gate",
          "dependent": "qa-test-expert"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like QA / Test Expert, E2E Testing Skill, and CI Quality Gate in order.\n\nProposed gold_skills (ordered if composition):\n[\"qa-test-expert\", \"e2e-testing\", \"ci-quality-gate\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0041",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Stock CSV Downloader”-style work, orchestrated under “Daily Stock Check”. Which ordered skill chain fits?",
      "gold_skills": [
        "daily-stock-check",
        "stock-csv-downloader"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "stock-csv-downloader",
          "dependent": "daily-stock-check"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Stock CSV Downloader”-style work, orchestrated under “Daily Stock Check”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"daily-stock-check\", \"stock-csv-downloader\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0042",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Skill Guide Generator, Skill Optimizer, and Technical Writer in order.",
      "gold_skills": [
        "skill-guide-generator",
        "skill-optimizer",
        "technical-writer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "skill-optimizer",
          "dependent": "skill-guide-generator"
        },
        {
          "prerequisite": "technical-writer",
          "dependent": "skill-guide-generator"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Skill Guide Generator, Skill Optimizer, and Technical Writer in order.\n\nProposed gold_skills (ordered if composition):\n[\"skill-guide-generator\", \"skill-optimizer\", \"technical-writer\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0043",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Docs Tutor”-style work, orchestrated under “Onboarding Accelerator”. Which ordered skill chain fits?",
      "gold_skills": [
        "onboarding-accelerator",
        "docs-tutor"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "docs-tutor",
          "dependent": "onboarding-accelerator"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Docs Tutor”-style work, orchestrated under “Onboarding Accelerator”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"onboarding-accelerator\", \"docs-tutor\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0044",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Daily AM Orchestrator, Gmail Daily Triage, and SOD Ship in order.",
      "gold_skills": [
        "daily-am-orchestrator",
        "gmail-daily-triage",
        "sod-ship"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "gmail-daily-triage",
          "dependent": "daily-am-orchestrator"
        },
        {
          "prerequisite": "sod-ship",
          "dependent": "daily-am-orchestrator"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Daily AM Orchestrator, Gmail Daily Triage, and SOD Ship in order.\n\nProposed gold_skills (ordered if composition):\n[\"daily-am-orchestrator\", \"gmail-daily-triage\", \"sod-ship\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0045",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with SOD Ship, then Skill Guide Generator, Release Ship, EOD Ship, then Morning Ship. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "sod-ship",
        "skill-guide-generator",
        "release-ship",
        "eod-ship",
        "morning-ship"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "skill-guide-generator",
          "dependent": "sod-ship"
        },
        {
          "prerequisite": "release-ship",
          "dependent": "sod-ship"
        },
        {
          "prerequisite": "eod-ship",
          "dependent": "sod-ship"
        },
        {
          "prerequisite": "morning-ship",
          "dependent": "sod-ship"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with SOD Ship, then Skill Guide Generator, Release Ship, EOD Ship, then Morning Ship. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"sod-ship\", \"skill-guide-generator\", \"release-ship\", \"eod-ship\", \"morning-ship\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0046",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Code Review All, Deep Review, and Simplify in order.",
      "gold_skills": [
        "code-review-all",
        "deep-review",
        "simplify"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "deep-review",
          "dependent": "code-review-all"
        },
        {
          "prerequisite": "simplify",
          "dependent": "code-review-all"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Code Review All, Deep Review, and Simplify in order.\n\nProposed gold_skills (ordered if composition):\n[\"code-review-all\", \"deep-review\", \"simplify\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0047",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like PR Review Captain, Technical Writer, and Domain-Split Commit in order.",
      "gold_skills": [
        "pr-review-captain",
        "technical-writer",
        "domain-commit"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "technical-writer",
          "dependent": "pr-review-captain"
        },
        {
          "prerequisite": "domain-commit",
          "dependent": "pr-review-captain"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like PR Review Captain, Technical Writer, and Domain-Split Commit in order.\n\nProposed gold_skills (ordered if composition):\n[\"pr-review-captain\", \"technical-writer\", \"domain-commit\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0048",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “AlphaEar Sentiment”-style work, orchestrated under “AlphaEar Signal Tracker”. Which ordered skill chain fits?",
      "gold_skills": [
        "alphaear-signal-tracker",
        "alphaear-sentiment"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "alphaear-sentiment",
          "dependent": "alphaear-signal-tracker"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “AlphaEar Sentiment”-style work, orchestrated under “AlphaEar Signal Tracker”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-signal-tracker\", \"alphaear-sentiment\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0049",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Paper Archive, NLM arXiv Slides: arXiv Paper to Expert Slide Deck Pipeline, and Related Papers Scout in order.",
      "gold_skills": [
        "paper-archive",
        "nlm-arxiv-slides",
        "related-papers-scout"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "nlm-arxiv-slides",
          "dependent": "paper-archive"
        },
        {
          "prerequisite": "related-papers-scout",
          "dependent": "paper-archive"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Paper Archive, NLM arXiv Slides: arXiv Paper to Expert Slide Deck Pipeline, and Related Papers Scout in order.\n\nProposed gold_skills (ordered if composition):\n[\"paper-archive\", \"nlm-arxiv-slides\", \"related-papers-scout\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0050",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with PPT Template Engine, then DOCX Template Engine, PPTX Skill, then Office Template Enforcer. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "ppt-template-engine",
        "docx-template-engine",
        "anthropic-pptx",
        "office-template-enforcer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "docx-template-engine",
          "dependent": "ppt-template-engine"
        },
        {
          "prerequisite": "anthropic-pptx",
          "dependent": "ppt-template-engine"
        },
        {
          "prerequisite": "office-template-enforcer",
          "dependent": "ppt-template-engine"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with PPT Template Engine, then DOCX Template Engine, PPTX Skill, then Office Template Enforcer. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"ppt-template-engine\", \"docx-template-engine\", \"anthropic-pptx\", \"office-template-enforcer\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0051",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Mission Control”-style work, orchestrated under “Workflow Sequential”. Which ordered skill chain fits?",
      "gold_skills": [
        "workflow-sequential",
        "mission-control"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "mission-control",
          "dependent": "workflow-sequential"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Mission Control”-style work, orchestrated under “Workflow Sequential”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"workflow-sequential\", \"mission-control\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0052",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with GWS Email Reply, then Slack Messaging Best Practices, Google Calendar, then Recall: Long-Term Memory Search. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "gws-email-reply",
        "kwp-slack-slack-messaging",
        "gws-calendar",
        "recall"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "kwp-slack-slack-messaging",
          "dependent": "gws-email-reply"
        },
        {
          "prerequisite": "gws-calendar",
          "dependent": "gws-email-reply"
        },
        {
          "prerequisite": "recall",
          "dependent": "gws-email-reply"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with GWS Email Reply, then Slack Messaging Best Practices, Google Calendar, then Recall: Long-Term Memory Search. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"gws-email-reply\", \"kwp-slack-slack-messaging\", \"gws-calendar\", \"recall\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0053",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Eval Harness Skill, AI Quality Evaluator, and Evals Skills in order.",
      "gold_skills": [
        "ecc-eval-harness",
        "ai-quality-evaluator",
        "evals-skills"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "ai-quality-evaluator",
          "dependent": "ecc-eval-harness"
        },
        {
          "prerequisite": "evals-skills",
          "dependent": "ecc-eval-harness"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Eval Harness Skill, AI Quality Evaluator, and Evals Skills in order.\n\nProposed gold_skills (ordered if composition):\n[\"ecc-eval-harness\", \"ai-quality-evaluator\", \"evals-skills\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0054",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Codebase Archaeologist, then Simplify, Domain-Split Commit, then Diagnose. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "codebase-archaeologist",
        "simplify",
        "domain-commit",
        "diagnose"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "simplify",
          "dependent": "codebase-archaeologist"
        },
        {
          "prerequisite": "domain-commit",
          "dependent": "codebase-archaeologist"
        },
        {
          "prerequisite": "diagnose",
          "dependent": "codebase-archaeologist"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Codebase Archaeologist, then Simplify, Domain-Split Commit, then Diagnose. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"codebase-archaeologist\", \"simplify\", \"domain-commit\", \"diagnose\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0055",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Pipeline Builder, Today, and AI Workflow Integrator in order.",
      "gold_skills": [
        "pipeline-builder",
        "today",
        "ai-workflow-integrator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "today",
          "dependent": "pipeline-builder"
        },
        {
          "prerequisite": "ai-workflow-integrator",
          "dependent": "pipeline-builder"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Pipeline Builder, Today, and AI Workflow Integrator in order.\n\nProposed gold_skills (ordered if composition):\n[\"pipeline-builder\", \"today\", \"ai-workflow-integrator\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0056",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Meeting Digest”-style work, orchestrated under “md-to-notion”. Which ordered skill chain fits?",
      "gold_skills": [
        "md-to-notion",
        "meeting-digest"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "meeting-digest",
          "dependent": "md-to-notion"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Meeting Digest”-style work, orchestrated under “md-to-notion”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"md-to-notion\", \"meeting-digest\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0057",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “tab-event-detect”-style work, orchestrated under “tab-patterns”. Which ordered skill chain fits?",
      "gold_skills": [
        "tab-patterns",
        "tab-event-detect"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "tab-event-detect",
          "dependent": "tab-patterns"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “tab-event-detect”-style work, orchestrated under “tab-patterns”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"tab-patterns\", \"tab-event-detect\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0058",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Paperclip Setup, then Paperclip Control, Paperclip Tasks, then Paperclip Agents. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "paperclip-setup",
        "paperclip-control",
        "paperclip-tasks",
        "paperclip-agents"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "paperclip-control",
          "dependent": "paperclip-setup"
        },
        {
          "prerequisite": "paperclip-tasks",
          "dependent": "paperclip-setup"
        },
        {
          "prerequisite": "paperclip-agents",
          "dependent": "paperclip-setup"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Paperclip Setup, then Paperclip Control, Paperclip Tasks, then Paperclip Agents. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"paperclip-setup\", \"paperclip-control\", \"paperclip-tasks\", \"paperclip-agents\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0059",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Data Validation Skill”-style work, orchestrated under “Data Scientist Perspective Analyzer”. Which ordered skill chain fits?",
      "gold_skills": [
        "role-data-scientist",
        "kwp-data-data-validation"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "kwp-data-data-validation",
          "dependent": "role-data-scientist"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Data Validation Skill”-style work, orchestrated under “Data Scientist Perspective Analyzer”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"role-data-scientist\", \"kwp-data-data-validation\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0060",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with DOCX Template Engine, then Office Template Enforcer, DOCX creation, editing, and analysis, then PPT Template Engine. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "docx-template-engine",
        "office-template-enforcer",
        "anthropic-docx",
        "ppt-template-engine"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "office-template-enforcer",
          "dependent": "docx-template-engine"
        },
        {
          "prerequisite": "anthropic-docx",
          "dependent": "docx-template-engine"
        },
        {
          "prerequisite": "ppt-template-engine",
          "dependent": "docx-template-engine"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with DOCX Template Engine, then Office Template Enforcer, DOCX creation, editing, and analysis, then PPT Template Engine. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"docx-template-engine\", \"office-template-enforcer\", \"anthropic-docx\", \"ppt-template-engine\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0061",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Hugging Face Hub CLI”-style work, orchestrated under “Hugging Face Papers”. Which ordered skill chain fits?",
      "gold_skills": [
        "hf-papers",
        "hf-hub"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "hf-hub",
          "dependent": "hf-papers"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Hugging Face Hub CLI”-style work, orchestrated under “Hugging Face Papers”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"hf-papers\", \"hf-hub\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0062",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like tab-bollinger-refresh, tab-stock-sync, and tab-turtle-refresh in order.",
      "gold_skills": [
        "tab-bollinger-refresh",
        "tab-stock-sync",
        "tab-turtle-refresh"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "tab-stock-sync",
          "dependent": "tab-bollinger-refresh"
        },
        {
          "prerequisite": "tab-turtle-refresh",
          "dependent": "tab-bollinger-refresh"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like tab-bollinger-refresh, tab-stock-sync, and tab-turtle-refresh in order.\n\nProposed gold_skills (ordered if composition):\n[\"tab-bollinger-refresh\", \"tab-stock-sync\", \"tab-turtle-refresh\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0063",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like CI Quality Gate, QA / Test Expert, and Dependency Auditor in order.",
      "gold_skills": [
        "ci-quality-gate",
        "qa-test-expert",
        "dependency-auditor"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "qa-test-expert",
          "dependent": "ci-quality-gate"
        },
        {
          "prerequisite": "dependency-auditor",
          "dependent": "ci-quality-gate"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like CI Quality Gate, QA / Test Expert, and Dependency Auditor in order.\n\nProposed gold_skills (ordered if composition):\n[\"ci-quality-gate\", \"qa-test-expert\", \"dependency-auditor\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0064",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with PM Perspective Analyzer, then Workflow Miner, Stakeholder Communications Skill, then Intent Alignment Tracker. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "role-pm",
        "workflow-miner",
        "kwp-product-management-stakeholder-comms",
        "intent-alignment-tracker"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "workflow-miner",
          "dependent": "role-pm"
        },
        {
          "prerequisite": "kwp-product-management-stakeholder-comms",
          "dependent": "role-pm"
        },
        {
          "prerequisite": "intent-alignment-tracker",
          "dependent": "role-pm"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with PM Perspective Analyzer, then Workflow Miner, Stakeholder Communications Skill, then Intent Alignment Tracker. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"role-pm\", \"workflow-miner\", \"kwp-product-management-stakeholder-comms\", \"intent-alignment-tracker\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0065",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with CSO Perspective Analyzer, then Workflow Miner, Skill Composer, then Competitive Analysis Skill. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "role-cso",
        "workflow-miner",
        "skill-composer",
        "kwp-marketing-competitive-analysis"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "workflow-miner",
          "dependent": "role-cso"
        },
        {
          "prerequisite": "skill-composer",
          "dependent": "role-cso"
        },
        {
          "prerequisite": "kwp-marketing-competitive-analysis",
          "dependent": "role-cso"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with CSO Perspective Analyzer, then Workflow Miner, Skill Composer, then Competitive Analysis Skill. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"role-cso\", \"workflow-miner\", \"skill-composer\", \"kwp-marketing-competitive-analysis\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0066",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Daily Stock Check”-style work, orchestrated under “Market Top Detector Skill”. Which ordered skill chain fits?",
      "gold_skills": [
        "trading-market-top-detector",
        "daily-stock-check"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "daily-stock-check",
          "dependent": "trading-market-top-detector"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Daily Stock Check”-style work, orchestrated under “Market Top Detector Skill”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"trading-market-top-detector\", \"daily-stock-check\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0067",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Meeting Digest, Notion Meeting Sync, and PPTX Skill in order.",
      "gold_skills": [
        "meeting-digest",
        "notion-meeting-sync",
        "anthropic-pptx"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "notion-meeting-sync",
          "dependent": "meeting-digest"
        },
        {
          "prerequisite": "anthropic-pptx",
          "dependent": "meeting-digest"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Meeting Digest, Notion Meeting Sync, and PPTX Skill in order.\n\nProposed gold_skills (ordered if composition):\n[\"meeting-digest\", \"notion-meeting-sync\", \"anthropic-pptx\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0068",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like SEFO Federation, SEFO Governance, and SEFO Orchestrator in order.",
      "gold_skills": [
        "sefo-federation",
        "sefo-governance",
        "sefo-orchestrator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "sefo-governance",
          "dependent": "sefo-federation"
        },
        {
          "prerequisite": "sefo-orchestrator",
          "dependent": "sefo-federation"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like SEFO Federation, SEFO Governance, and SEFO Orchestrator in order.\n\nProposed gold_skills (ordered if composition):\n[\"sefo-federation\", \"sefo-governance\", \"sefo-orchestrator\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0069",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Problem Definition, then PM Execution, Simplify, then Frontend Expert. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "problem-definition",
        "pm-execution",
        "simplify",
        "frontend-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "pm-execution",
          "dependent": "problem-definition"
        },
        {
          "prerequisite": "simplify",
          "dependent": "problem-definition"
        },
        {
          "prerequisite": "frontend-expert",
          "dependent": "problem-definition"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Problem Definition, then PM Execution, Simplify, then Frontend Expert. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"problem-definition\", \"pm-execution\", \"simplify\", \"frontend-expert\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0070",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Standup Digest, Google Calendar, and Calendar Daily Briefing in order.",
      "gold_skills": [
        "standup-digest",
        "gws-calendar",
        "calendar-daily-briefing"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "gws-calendar",
          "dependent": "standup-digest"
        },
        {
          "prerequisite": "calendar-daily-briefing",
          "dependent": "standup-digest"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Standup Digest, Google Calendar, and Calendar Daily Briefing in order.\n\nProposed gold_skills (ordered if composition):\n[\"standup-digest\", \"gws-calendar\", \"calendar-daily-briefing\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0071",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “PM Execution”-style work, orchestrated under “PRD Research Factory”. Which ordered skill chain fits?",
      "gold_skills": [
        "prd-research-factory",
        "pm-execution"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "pm-execution",
          "dependent": "prd-research-factory"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “PM Execution”-style work, orchestrated under “PRD Research Factory”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"prd-research-factory\", \"pm-execution\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0072",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with tab-stock-sync, then Stock CSV Downloader, tab-fundamental-sync, then Daily Stock Check. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "tab-stock-sync",
        "stock-csv-downloader",
        "tab-fundamental-sync",
        "daily-stock-check"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "stock-csv-downloader",
          "dependent": "tab-stock-sync"
        },
        {
          "prerequisite": "tab-fundamental-sync",
          "dependent": "tab-stock-sync"
        },
        {
          "prerequisite": "daily-stock-check",
          "dependent": "tab-stock-sync"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with tab-stock-sync, then Stock CSV Downloader, tab-fundamental-sync, then Daily Stock Check. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"tab-stock-sync\", \"stock-csv-downloader\", \"tab-fundamental-sync\", \"daily-stock-check\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0073",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Design Review, then Figma → TDS Code Generator, AI Platform FE - FSD 개발 및 마이그레이션, then Screen Implementation Orchestrator. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "design-review",
        "figma-to-tds",
        "fsd-development",
        "implement-screen"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "figma-to-tds",
          "dependent": "design-review"
        },
        {
          "prerequisite": "fsd-development",
          "dependent": "design-review"
        },
        {
          "prerequisite": "implement-screen",
          "dependent": "design-review"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Design Review, then Figma → TDS Code Generator, AI Platform FE - FSD 개발 및 마이그레이션, then Screen Implementation Orchestrator. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"design-review\", \"figma-to-tds\", \"fsd-development\", \"implement-screen\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0074",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Workflow Eval-Opt, Mission Control, and Ship in order.",
      "gold_skills": [
        "workflow-eval-opt",
        "mission-control",
        "ship"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "mission-control",
          "dependent": "workflow-eval-opt"
        },
        {
          "prerequisite": "ship",
          "dependent": "workflow-eval-opt"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Workflow Eval-Opt, Mission Control, and Ship in order.\n\nProposed gold_skills (ordered if composition):\n[\"workflow-eval-opt\", \"mission-control\", \"ship\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0075",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Dependency Radar, Deep Review, and PM Execution in order.",
      "gold_skills": [
        "dependency-radar",
        "deep-review",
        "pm-execution"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "deep-review",
          "dependent": "dependency-radar"
        },
        {
          "prerequisite": "pm-execution",
          "dependent": "dependency-radar"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Dependency Radar, Deep Review, and PM Execution in order.\n\nProposed gold_skills (ordered if composition):\n[\"dependency-radar\", \"deep-review\", \"pm-execution\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0076",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Screen Description Generator, Figma → TDS Code Generator, and Screen Implementation Orchestrator in order.",
      "gold_skills": [
        "screen-description",
        "figma-to-tds",
        "implement-screen"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "figma-to-tds",
          "dependent": "screen-description"
        },
        {
          "prerequisite": "implement-screen",
          "dependent": "screen-description"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Screen Description Generator, Figma → TDS Code Generator, and Screen Implementation Orchestrator in order.\n\nProposed gold_skills (ordered if composition):\n[\"screen-description\", \"figma-to-tds\", \"implement-screen\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0077",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Simplify”-style work, orchestrated under “Workflow Parallel”. Which ordered skill chain fits?",
      "gold_skills": [
        "workflow-parallel",
        "simplify"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "simplify",
          "dependent": "workflow-parallel"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Simplify”-style work, orchestrated under “Workflow Parallel”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"workflow-parallel\", \"simplify\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0078",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Overlay 패턴 가이드, Figma → TDS Code Generator, and Screen Implementation Orchestrator in order.",
      "gold_skills": [
        "overlay-layout-patterns",
        "figma-to-tds",
        "implement-screen"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "figma-to-tds",
          "dependent": "overlay-layout-patterns"
        },
        {
          "prerequisite": "implement-screen",
          "dependent": "overlay-layout-patterns"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Overlay 패턴 가이드, Figma → TDS Code Generator, and Screen Implementation Orchestrator in order.\n\nProposed gold_skills (ordered if composition):\n[\"overlay-layout-patterns\", \"figma-to-tds\", \"implement-screen\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0079",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with tab-market-environment, then tab-news-fetch, tab-stock-sync, then tab-market-breadth. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "tab-market-environment",
        "tab-news-fetch",
        "tab-stock-sync",
        "tab-market-breadth"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "tab-news-fetch",
          "dependent": "tab-market-environment"
        },
        {
          "prerequisite": "tab-stock-sync",
          "dependent": "tab-market-environment"
        },
        {
          "prerequisite": "tab-market-breadth",
          "dependent": "tab-market-environment"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with tab-market-environment, then tab-news-fetch, tab-stock-sync, then tab-market-breadth. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"tab-market-environment\", \"tab-news-fetch\", \"tab-stock-sync\", \"tab-market-breadth\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0080",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Frontend Expert”-style work, orchestrated under “Web Artifacts Builder”. Which ordered skill chain fits?",
      "gold_skills": [
        "anthropic-web-artifacts-builder",
        "frontend-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "frontend-expert",
          "dependent": "anthropic-web-artifacts-builder"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Frontend Expert”-style work, orchestrated under “Web Artifacts Builder”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-web-artifacts-builder\", \"frontend-expert\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0081",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Paperclip Agents, then Paperclip Setup, Paperclip Control, then Paperclip Tasks. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "paperclip-agents",
        "paperclip-setup",
        "paperclip-control",
        "paperclip-tasks"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "paperclip-setup",
          "dependent": "paperclip-agents"
        },
        {
          "prerequisite": "paperclip-control",
          "dependent": "paperclip-agents"
        },
        {
          "prerequisite": "paperclip-tasks",
          "dependent": "paperclip-agents"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Paperclip Agents, then Paperclip Setup, Paperclip Control, then Paperclip Tasks. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"paperclip-agents\", \"paperclip-setup\", \"paperclip-control\", \"paperclip-tasks\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0082",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Market Environment Analysis, AlphaEar News, and Daily Stock Check in order.",
      "gold_skills": [
        "trading-market-environment-analysis",
        "alphaear-news",
        "daily-stock-check"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "alphaear-news",
          "dependent": "trading-market-environment-analysis"
        },
        {
          "prerequisite": "daily-stock-check",
          "dependent": "trading-market-environment-analysis"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Market Environment Analysis, AlphaEar News, and Daily Stock Check in order.\n\nProposed gold_skills (ordered if composition):\n[\"trading-market-environment-analysis\", \"alphaear-news\", \"daily-stock-check\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0083",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Create distinctive, production-grade frontend interfaces with high design quality, Frontend Expert, and UX Expert in order.",
      "gold_skills": [
        "anthropic-frontend-design",
        "frontend-expert",
        "ux-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "frontend-expert",
          "dependent": "anthropic-frontend-design"
        },
        {
          "prerequisite": "ux-expert",
          "dependent": "anthropic-frontend-design"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Create distinctive, production-grade frontend interfaces with high design quality, Frontend Expert, and UX Expert in order.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-frontend-design\", \"frontend-expert\", \"ux-expert\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0084",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “User Research”-style work, orchestrated under “UX Designer Perspective Analyzer”. Which ordered skill chain fits?",
      "gold_skills": [
        "role-ux-designer",
        "kwp-design-user-research"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "kwp-design-user-research",
          "dependent": "role-ux-designer"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “User Research”-style work, orchestrated under “UX Designer Perspective Analyzer”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"role-ux-designer\", \"kwp-design-user-research\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0085",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Google Chat, Gmail, and Google Workspace CLI in order.",
      "gold_skills": [
        "gws-chat",
        "gws-gmail",
        "gws-workspace"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "gws-gmail",
          "dependent": "gws-chat"
        },
        {
          "prerequisite": "gws-workspace",
          "dependent": "gws-chat"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Google Chat, Gmail, and Google Workspace CLI in order.\n\nProposed gold_skills (ordered if composition):\n[\"gws-chat\", \"gws-gmail\", \"gws-workspace\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0086",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Recipe: Save Gmail Message to Google Docs, Google Docs, and Google Drive in order.",
      "gold_skills": [
        "gws-recipe-save-email-to-doc",
        "gws-docs",
        "gws-drive"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "gws-docs",
          "dependent": "gws-recipe-save-email-to-doc"
        },
        {
          "prerequisite": "gws-drive",
          "dependent": "gws-recipe-save-email-to-doc"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Recipe: Save Gmail Message to Google Docs, Google Docs, and Google Drive in order.\n\nProposed gold_skills (ordered if composition):\n[\"gws-recipe-save-email-to-doc\", \"gws-docs\", \"gws-drive\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0087",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Running Workloads on Hugging Face Jobs, TRL Training on Hugging Face Jobs, and Daily Stock Check in order.",
      "gold_skills": [
        "hf-jobs",
        "hf-model-trainer",
        "daily-stock-check"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "hf-model-trainer",
          "dependent": "hf-jobs"
        },
        {
          "prerequisite": "daily-stock-check",
          "dependent": "hf-jobs"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Running Workloads on Hugging Face Jobs, TRL Training on Hugging Face Jobs, and Daily Stock Check in order.\n\nProposed gold_skills (ordered if composition):\n[\"hf-jobs\", \"hf-model-trainer\", \"daily-stock-check\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0088",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “MiroFish Financial Scenario Simulation”-style work, orchestrated under “MiroFish Public Opinion Simulation”. Which ordered skill chain fits?",
      "gold_skills": [
        "mirofish-opinion-sim",
        "mirofish-financial-sim"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "mirofish-financial-sim",
          "dependent": "mirofish-opinion-sim"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “MiroFish Financial Scenario Simulation”-style work, orchestrated under “MiroFish Public Opinion Simulation”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"mirofish-opinion-sim\", \"mirofish-financial-sim\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0089",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like email-auto-reply, smart-meeting-scheduler, and Gmail Daily Triage in order.",
      "gold_skills": [
        "email-auto-reply",
        "smart-meeting-scheduler",
        "gmail-daily-triage"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "smart-meeting-scheduler",
          "dependent": "email-auto-reply"
        },
        {
          "prerequisite": "gmail-daily-triage",
          "dependent": "email-auto-reply"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like email-auto-reply, smart-meeting-scheduler, and Gmail Daily Triage in order.\n\nProposed gold_skills (ordered if composition):\n[\"email-auto-reply\", \"smart-meeting-scheduler\", \"gmail-daily-triage\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0090",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like AlphaEar DeepEar Lite, Weekly Stock Update, and Daily Stock Check in order.",
      "gold_skills": [
        "alphaear-deepear-lite",
        "weekly-stock-update",
        "daily-stock-check"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "weekly-stock-update",
          "dependent": "alphaear-deepear-lite"
        },
        {
          "prerequisite": "daily-stock-check",
          "dependent": "alphaear-deepear-lite"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like AlphaEar DeepEar Lite, Weekly Stock Update, and Daily Stock Check in order.\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-deepear-lite\", \"weekly-stock-update\", \"daily-stock-check\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0091",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Helm Validator, then K8s Manifest Validator, IaC Review Agent, Terraform Reviewer, then SRE / DevOps Expert. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "helm-validator",
        "k8s-manifest-validator",
        "iac-review-agent",
        "terraform-reviewer",
        "sre-devops-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "k8s-manifest-validator",
          "dependent": "helm-validator"
        },
        {
          "prerequisite": "iac-review-agent",
          "dependent": "helm-validator"
        },
        {
          "prerequisite": "terraform-reviewer",
          "dependent": "helm-validator"
        },
        {
          "prerequisite": "sre-devops-expert",
          "dependent": "helm-validator"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Helm Validator, then K8s Manifest Validator, IaC Review Agent, Terraform Reviewer, then SRE / DevOps Expert. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"helm-validator\", \"k8s-manifest-validator\", \"iac-review-agent\", \"terraform-reviewer\", \"sre-devops-expert\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0092",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “md-to-notion”-style work, orchestrated under “Deep Research Pipeline”. Which ordered skill chain fits?",
      "gold_skills": [
        "deep-research-pipeline",
        "md-to-notion"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "md-to-notion",
          "dependent": "deep-research-pipeline"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “md-to-notion”-style work, orchestrated under “Deep Research Pipeline”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"deep-research-pipeline\", \"md-to-notion\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0093",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with MiroFish Orchestrator, then Technical Analyst, Today, Daily Stock Check, then AlphaEar News. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "mirofish",
        "trading-technical-analyst",
        "today",
        "daily-stock-check",
        "alphaear-news"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "trading-technical-analyst",
          "dependent": "mirofish"
        },
        {
          "prerequisite": "today",
          "dependent": "mirofish"
        },
        {
          "prerequisite": "daily-stock-check",
          "dependent": "mirofish"
        },
        {
          "prerequisite": "alphaear-news",
          "dependent": "mirofish"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with MiroFish Orchestrator, then Technical Analyst, Today, Daily Stock Check, then AlphaEar News. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"mirofish\", \"trading-technical-analyst\", \"today\", \"daily-stock-check\", \"alphaear-news\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0094",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Today”-style work, orchestrated under “Cognee”. Which ordered skill chain fits?",
      "gold_skills": [
        "cognee",
        "today"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "today",
          "dependent": "cognee"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Today”-style work, orchestrated under “Cognee”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"cognee\", \"today\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0095",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like tab-hot-stock-discovery, tab-stock-sync, and tab-screening in order.",
      "gold_skills": [
        "tab-hot-stock-discovery",
        "tab-stock-sync",
        "tab-screening"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "tab-stock-sync",
          "dependent": "tab-hot-stock-discovery"
        },
        {
          "prerequisite": "tab-screening",
          "dependent": "tab-hot-stock-discovery"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like tab-hot-stock-discovery, tab-stock-sync, and tab-screening in order.\n\nProposed gold_skills (ordered if composition):\n[\"tab-hot-stock-discovery\", \"tab-stock-sync\", \"tab-screening\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0096",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like AutoSkill Merger, AutoSkill Extractor, and AutoSkill Evolve in order.",
      "gold_skills": [
        "autoskill-merger",
        "autoskill-extractor",
        "autoskill-evolve"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "autoskill-extractor",
          "dependent": "autoskill-merger"
        },
        {
          "prerequisite": "autoskill-evolve",
          "dependent": "autoskill-merger"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like AutoSkill Merger, AutoSkill Extractor, and AutoSkill Evolve in order.\n\nProposed gold_skills (ordered if composition):\n[\"autoskill-merger\", \"autoskill-extractor\", \"autoskill-evolve\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0097",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with AI Platform FE - FSD 개발 및 마이그레이션, then Figma → TDS Code Generator, Overlay 패턴 가이드, Screen Description Generator, then Screen Implementation Orchestrator. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "fsd-development",
        "figma-to-tds",
        "overlay-layout-patterns",
        "screen-description",
        "implement-screen"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "figma-to-tds",
          "dependent": "fsd-development"
        },
        {
          "prerequisite": "overlay-layout-patterns",
          "dependent": "fsd-development"
        },
        {
          "prerequisite": "screen-description",
          "dependent": "fsd-development"
        },
        {
          "prerequisite": "implement-screen",
          "dependent": "fsd-development"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with AI Platform FE - FSD 개발 및 마이그레이션, then Figma → TDS Code Generator, Overlay 패턴 가이드, Screen Description Generator, then Screen Implementation Orchestrator. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"fsd-development\", \"figma-to-tds\", \"overlay-layout-patterns\", \"screen-description\", \"implement-screen\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0098",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Skill Composer, then Workflow Sequential, Mission Control, Workflow Parallel, then Skill Optimizer. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "skill-composer",
        "workflow-sequential",
        "mission-control",
        "workflow-parallel",
        "skill-optimizer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "workflow-sequential",
          "dependent": "skill-composer"
        },
        {
          "prerequisite": "mission-control",
          "dependent": "skill-composer"
        },
        {
          "prerequisite": "workflow-parallel",
          "dependent": "skill-composer"
        },
        {
          "prerequisite": "skill-optimizer",
          "dependent": "skill-composer"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Skill Composer, then Workflow Sequential, Mission Control, Workflow Parallel, then Skill Optimizer. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"skill-composer\", \"workflow-sequential\", \"mission-control\", \"workflow-parallel\", \"skill-optimizer\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0099",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Market Environment Analysis”-style work, orchestrated under “Trading Expert Perspective Analyzer”. Which ordered skill chain fits?",
      "gold_skills": [
        "role-trading-expert",
        "trading-market-environment-analysis"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "trading-market-environment-analysis",
          "dependent": "role-trading-expert"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Market Environment Analysis”-style work, orchestrated under “Trading Expert Perspective Analyzer”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"role-trading-expert\", \"trading-market-environment-analysis\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0100",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Executive Summary Generator Agent Personality”-style work, orchestrated under “Executive Briefing Generator”. Which ordered skill chain fits?",
      "gold_skills": [
        "executive-briefing",
        "agency-executive-summary-generator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "agency-executive-summary-generator",
          "dependent": "executive-briefing"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Executive Summary Generator Agent Personality”-style work, orchestrated under “Executive Briefing Generator”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"executive-briefing\", \"agency-executive-summary-generator\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0101",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like tab-technical-analysis, tab-turtle-refresh, and tab-bollinger-refresh in order.",
      "gold_skills": [
        "tab-technical-analysis",
        "tab-turtle-refresh",
        "tab-bollinger-refresh"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "tab-turtle-refresh",
          "dependent": "tab-technical-analysis"
        },
        {
          "prerequisite": "tab-bollinger-refresh",
          "dependent": "tab-technical-analysis"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like tab-technical-analysis, tab-turtle-refresh, and tab-bollinger-refresh in order.\n\nProposed gold_skills (ordered if composition):\n[\"tab-technical-analysis\", \"tab-turtle-refresh\", \"tab-bollinger-refresh\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0102",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Slack Messaging Best Practices”-style work, orchestrated under “Slack Agent Development”. Which ordered skill chain fits?",
      "gold_skills": [
        "slack-agent",
        "kwp-slack-slack-messaging"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "kwp-slack-slack-messaging",
          "dependent": "slack-agent"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Slack Messaging Best Practices”-style work, orchestrated under “Slack Agent Development”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"slack-agent\", \"kwp-slack-slack-messaging\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0103",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Decision Router, then X-to-Slack: Tweet Intelligence Pipeline, Slack Messaging Best Practices, Google Daily Automation, then Bespin News Digest. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "decision-router",
        "x-to-slack",
        "kwp-slack-slack-messaging",
        "google-daily",
        "bespin-news-digest"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "x-to-slack",
          "dependent": "decision-router"
        },
        {
          "prerequisite": "kwp-slack-slack-messaging",
          "dependent": "decision-router"
        },
        {
          "prerequisite": "google-daily",
          "dependent": "decision-router"
        },
        {
          "prerequisite": "bespin-news-digest",
          "dependent": "decision-router"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Decision Router, then X-to-Slack: Tweet Intelligence Pipeline, Slack Messaging Best Practices, Google Daily Automation, then Bespin News Digest. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"decision-router\", \"x-to-slack\", \"kwp-slack-slack-messaging\", \"google-daily\", \"bespin-news-digest\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0104",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Plans, Prompt Architect, and Mission Control in order.",
      "gold_skills": [
        "plans",
        "prompt-architect",
        "mission-control"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "prompt-architect",
          "dependent": "plans"
        },
        {
          "prerequisite": "mission-control",
          "dependent": "plans"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Plans, Prompt Architect, and Mission Control in order.\n\nProposed gold_skills (ordered if composition):\n[\"plans\", \"prompt-architect\", \"mission-control\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0105",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Feedback Meeting Scheduler, then smart-meeting-scheduler, Google Calendar, Sprint Orchestrator, then github-sprint-digest. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "feedback-meeting-scheduler",
        "smart-meeting-scheduler",
        "gws-calendar",
        "sprint-orchestrator",
        "github-sprint-digest"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "smart-meeting-scheduler",
          "dependent": "feedback-meeting-scheduler"
        },
        {
          "prerequisite": "gws-calendar",
          "dependent": "feedback-meeting-scheduler"
        },
        {
          "prerequisite": "sprint-orchestrator",
          "dependent": "feedback-meeting-scheduler"
        },
        {
          "prerequisite": "github-sprint-digest",
          "dependent": "feedback-meeting-scheduler"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Feedback Meeting Scheduler, then smart-meeting-scheduler, Google Calendar, Sprint Orchestrator, then github-sprint-digest. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"feedback-meeting-scheduler\", \"smart-meeting-scheduler\", \"gws-calendar\", \"sprint-orchestrator\", \"github-sprint-digest\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0106",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with NLM Video: Local Document to Expert Video Pipeline, then NotebookLM: Core Notebook Management, Transcribee, NotebookLM Studio: Content Generation, then NLM Slides: Local Document to Expert Slide Deck Pipeline. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "nlm-video",
        "notebooklm",
        "transcribee",
        "notebooklm-studio",
        "nlm-slides"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "notebooklm",
          "dependent": "nlm-video"
        },
        {
          "prerequisite": "transcribee",
          "dependent": "nlm-video"
        },
        {
          "prerequisite": "notebooklm-studio",
          "dependent": "nlm-video"
        },
        {
          "prerequisite": "nlm-slides",
          "dependent": "nlm-video"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with NLM Video: Local Document to Expert Video Pipeline, then NotebookLM: Core Notebook Management, Transcribee, NotebookLM Studio: Content Generation, then NLM Slides: Local Document to Expert Slide Deck Pipeline. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"nlm-video\", \"notebooklm\", \"transcribee\", \"notebooklm-studio\", \"nlm-slides\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0107",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “PDF Processing Guide”-style work, orchestrated under “Paper Review Pipeline”. Which ordered skill chain fits?",
      "gold_skills": [
        "paper-review",
        "anthropic-pdf"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "anthropic-pdf",
          "dependent": "paper-review"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “PDF Processing Guide”-style work, orchestrated under “Paper Review Pipeline”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"paper-review\", \"anthropic-pdf\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0108",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Intent Alignment Tracker, then AI Quality Evaluator, Deep Review, then Evals Skills. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "intent-alignment-tracker",
        "ai-quality-evaluator",
        "deep-review",
        "evals-skills"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "ai-quality-evaluator",
          "dependent": "intent-alignment-tracker"
        },
        {
          "prerequisite": "deep-review",
          "dependent": "intent-alignment-tracker"
        },
        {
          "prerequisite": "evals-skills",
          "dependent": "intent-alignment-tracker"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Intent Alignment Tracker, then AI Quality Evaluator, Deep Review, then Evals Skills. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"intent-alignment-tracker\", \"ai-quality-evaluator\", \"deep-review\", \"evals-skills\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0109",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Office Template Enforcer, PPT Template Engine, and DOCX creation, editing, and analysis in order.",
      "gold_skills": [
        "office-template-enforcer",
        "ppt-template-engine",
        "anthropic-docx"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "ppt-template-engine",
          "dependent": "office-template-enforcer"
        },
        {
          "prerequisite": "anthropic-docx",
          "dependent": "office-template-enforcer"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Office Template Enforcer, PPT Template Engine, and DOCX creation, editing, and analysis in order.\n\nProposed gold_skills (ordered if composition):\n[\"office-template-enforcer\", \"ppt-template-engine\", \"anthropic-docx\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0110",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Frontend Pipeline, Design Review, and Figma → TDS Code Generator in order.",
      "gold_skills": [
        "fe-pipeline",
        "design-review",
        "figma-to-tds"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "design-review",
          "dependent": "fe-pipeline"
        },
        {
          "prerequisite": "figma-to-tds",
          "dependent": "fe-pipeline"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Frontend Pipeline, Design Review, and Figma → TDS Code Generator in order.\n\nProposed gold_skills (ordered if composition):\n[\"fe-pipeline\", \"design-review\", \"figma-to-tds\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0111",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Model Arena, then Hugging Face Hub CLI, TRL Training on Hugging Face Jobs, then Today. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "hf-model-arena",
        "hf-hub",
        "hf-model-trainer",
        "today"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "hf-hub",
          "dependent": "hf-model-arena"
        },
        {
          "prerequisite": "hf-model-trainer",
          "dependent": "hf-model-arena"
        },
        {
          "prerequisite": "today",
          "dependent": "hf-model-arena"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Model Arena, then Hugging Face Hub CLI, TRL Training on Hugging Face Jobs, then Today. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"hf-model-arena\", \"hf-hub\", \"hf-model-trainer\", \"today\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0112",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “PR Review Captain”-style work, orchestrated under “Cursor Automations”. Which ordered skill chain fits?",
      "gold_skills": [
        "cursor-automations",
        "pr-review-captain"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "pr-review-captain",
          "dependent": "cursor-automations"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “PR Review Captain”-style work, orchestrated under “Cursor Automations”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"cursor-automations\", \"pr-review-captain\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0113",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “IaC Review Agent”-style work, orchestrated under “Infrastructure Drift Detector”. Which ordered skill chain fits?",
      "gold_skills": [
        "infra-drift-detector",
        "iac-review-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "iac-review-agent",
          "dependent": "infra-drift-detector"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “IaC Review Agent”-style work, orchestrated under “Infrastructure Drift Detector”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"infra-drift-detector\", \"iac-review-agent\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0114",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with AlphaEar Stock, then Stock CSV Downloader, Weekly Stock Update, then Daily Stock Check. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "alphaear-stock",
        "stock-csv-downloader",
        "weekly-stock-update",
        "daily-stock-check"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "stock-csv-downloader",
          "dependent": "alphaear-stock"
        },
        {
          "prerequisite": "weekly-stock-update",
          "dependent": "alphaear-stock"
        },
        {
          "prerequisite": "daily-stock-check",
          "dependent": "alphaear-stock"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with AlphaEar Stock, then Stock CSV Downloader, Weekly Stock Update, then Daily Stock Check. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-stock\", \"stock-csv-downloader\", \"weekly-stock-update\", \"daily-stock-check\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0115",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “PM Product Strategy”-style work, orchestrated under “PM Execution”. Which ordered skill chain fits?",
      "gold_skills": [
        "pm-execution",
        "pm-product-strategy"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "pm-product-strategy",
          "dependent": "pm-execution"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “PM Product Strategy”-style work, orchestrated under “PM Execution”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"pm-execution\", \"pm-product-strategy\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0116",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Today, then tab-stock-sync, tab-hot-stock-discovery, tab-dualma-refresh, then Daily Stock Check. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "today",
        "tab-stock-sync",
        "tab-hot-stock-discovery",
        "tab-dualma-refresh",
        "daily-stock-check"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "tab-stock-sync",
          "dependent": "today"
        },
        {
          "prerequisite": "tab-hot-stock-discovery",
          "dependent": "today"
        },
        {
          "prerequisite": "tab-dualma-refresh",
          "dependent": "today"
        },
        {
          "prerequisite": "daily-stock-check",
          "dependent": "today"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Today, then tab-stock-sync, tab-hot-stock-discovery, tab-dualma-refresh, then Daily Stock Check. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"today\", \"tab-stock-sync\", \"tab-hot-stock-discovery\", \"tab-dualma-refresh\", \"daily-stock-check\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0117",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Weekly Stock Update”-style work, orchestrated under “Stock CSV Downloader”. Which ordered skill chain fits?",
      "gold_skills": [
        "stock-csv-downloader",
        "weekly-stock-update"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "weekly-stock-update",
          "dependent": "stock-csv-downloader"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Weekly Stock Update”-style work, orchestrated under “Stock CSV Downloader”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"stock-csv-downloader\", \"weekly-stock-update\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0118",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Demo Forge”-style work, orchestrated under “Sales Perspective Analyzer”. Which ordered skill chain fits?",
      "gold_skills": [
        "role-sales",
        "demo-forge"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "demo-forge",
          "dependent": "role-sales"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Demo Forge”-style work, orchestrated under “Sales Perspective Analyzer”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"role-sales\", \"demo-forge\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0119",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Finance Perspective Analyzer, then Finance Tracker Agent Personality, Workflow Miner, then Audit Support. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "role-finance",
        "agency-finance-tracker",
        "workflow-miner",
        "kwp-finance-audit-support"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "agency-finance-tracker",
          "dependent": "role-finance"
        },
        {
          "prerequisite": "workflow-miner",
          "dependent": "role-finance"
        },
        {
          "prerequisite": "kwp-finance-audit-support",
          "dependent": "role-finance"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Finance Perspective Analyzer, then Finance Tracker Agent Personality, Workflow Miner, then Audit Support. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"role-finance\", \"agency-finance-tracker\", \"workflow-miner\", \"kwp-finance-audit-support\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0120",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Release Commander, then Ship, i18n Sync Manager, then Deep Review. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "release-commander",
        "ship",
        "i18n-sync",
        "deep-review"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "ship",
          "dependent": "release-commander"
        },
        {
          "prerequisite": "i18n-sync",
          "dependent": "release-commander"
        },
        {
          "prerequisite": "deep-review",
          "dependent": "release-commander"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Release Commander, then Ship, i18n Sync Manager, then Deep Review. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"release-commander\", \"ship\", \"i18n-sync\", \"deep-review\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0121",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with PM Product Discovery, then Campaign Planning Skill, User Research Synthesis Skill, then PM Product Strategy. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "pm-product-discovery",
        "kwp-marketing-campaign-planning",
        "kwp-product-management-user-research-synthesis",
        "pm-product-strategy"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "kwp-marketing-campaign-planning",
          "dependent": "pm-product-discovery"
        },
        {
          "prerequisite": "kwp-product-management-user-research-synthesis",
          "dependent": "pm-product-discovery"
        },
        {
          "prerequisite": "pm-product-strategy",
          "dependent": "pm-product-discovery"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with PM Product Discovery, then Campaign Planning Skill, User Research Synthesis Skill, then PM Product Strategy. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"pm-product-discovery\", \"kwp-marketing-campaign-planning\", \"kwp-product-management-user-research-synthesis\", \"pm-product-strategy\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0122",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Calendar Daily Briefing, then Google Workspace CLI, AI Chief of Staff -- Personal Assistant, then Google Calendar. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "calendar-daily-briefing",
        "gws-workspace",
        "ai-chief-of-staff",
        "gws-calendar"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "gws-workspace",
          "dependent": "calendar-daily-briefing"
        },
        {
          "prerequisite": "ai-chief-of-staff",
          "dependent": "calendar-daily-briefing"
        },
        {
          "prerequisite": "gws-calendar",
          "dependent": "calendar-daily-briefing"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Calendar Daily Briefing, then Google Workspace CLI, AI Chief of Staff -- Personal Assistant, then Google Calendar. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"calendar-daily-briefing\", \"gws-workspace\", \"ai-chief-of-staff\", \"gws-calendar\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0123",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Backtest Expert, then AlphaEar DeepEar Lite, Daily Stock Check, then Weekly Stock Update. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "trading-backtest-expert",
        "alphaear-deepear-lite",
        "daily-stock-check",
        "weekly-stock-update"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "alphaear-deepear-lite",
          "dependent": "trading-backtest-expert"
        },
        {
          "prerequisite": "daily-stock-check",
          "dependent": "trading-backtest-expert"
        },
        {
          "prerequisite": "weekly-stock-update",
          "dependent": "trading-backtest-expert"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Backtest Expert, then AlphaEar DeepEar Lite, Daily Stock Check, then Weekly Stock Update. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"trading-backtest-expert\", \"alphaear-deepear-lite\", \"daily-stock-check\", \"weekly-stock-update\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0124",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “AutoSkill Evolve”-style work, orchestrated under “AutoSkill Extractor”. Which ordered skill chain fits?",
      "gold_skills": [
        "autoskill-extractor",
        "autoskill-evolve"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "autoskill-evolve",
          "dependent": "autoskill-extractor"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “AutoSkill Evolve”-style work, orchestrated under “AutoSkill Extractor”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"autoskill-extractor\", \"autoskill-evolve\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0125",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Bespin News Digest, then X-to-Slack: Tweet Intelligence Pipeline, Decision Router, Gmail Daily Triage, then Twitter Timeline to Slack Pipeline. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "bespin-news-digest",
        "x-to-slack",
        "decision-router",
        "gmail-daily-triage",
        "twitter-timeline-to-slack"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "x-to-slack",
          "dependent": "bespin-news-digest"
        },
        {
          "prerequisite": "decision-router",
          "dependent": "bespin-news-digest"
        },
        {
          "prerequisite": "gmail-daily-triage",
          "dependent": "bespin-news-digest"
        },
        {
          "prerequisite": "twitter-timeline-to-slack",
          "dependent": "bespin-news-digest"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Bespin News Digest, then X-to-Slack: Tweet Intelligence Pipeline, Decision Router, Gmail Daily Triage, then Twitter Timeline to Slack Pipeline. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"bespin-news-digest\", \"x-to-slack\", \"decision-router\", \"gmail-daily-triage\", \"twitter-timeline-to-slack\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0126",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with E2E Overhaul, then QA / Test Expert, Test Suite, then E2E Testing Skill. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "e2e-overhaul",
        "qa-test-expert",
        "test-suite",
        "e2e-testing"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "qa-test-expert",
          "dependent": "e2e-overhaul"
        },
        {
          "prerequisite": "test-suite",
          "dependent": "e2e-overhaul"
        },
        {
          "prerequisite": "e2e-testing",
          "dependent": "e2e-overhaul"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with E2E Overhaul, then QA / Test Expert, Test Suite, then E2E Testing Skill. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"e2e-overhaul\", \"qa-test-expert\", \"test-suite\", \"e2e-testing\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0127",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like tab-glossary, tab-stock-sync, and Technical Writer in order.",
      "gold_skills": [
        "tab-glossary",
        "tab-stock-sync",
        "technical-writer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "tab-stock-sync",
          "dependent": "tab-glossary"
        },
        {
          "prerequisite": "technical-writer",
          "dependent": "tab-glossary"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like tab-glossary, tab-stock-sync, and Technical Writer in order.\n\nProposed gold_skills (ordered if composition):\n[\"tab-glossary\", \"tab-stock-sync\", \"technical-writer\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0128",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Daily Stock Check”-style work, orchestrated under “FinViz Screener”. Which ordered skill chain fits?",
      "gold_skills": [
        "trading-finviz-screener",
        "daily-stock-check"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "daily-stock-check",
          "dependent": "trading-finviz-screener"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Daily Stock Check”-style work, orchestrated under “FinViz Screener”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"trading-finviz-screener\", \"daily-stock-check\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0129",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with AlphaXiv Paper Lookup, then Defuddle, Paper Review Pipeline, then Related Papers Scout. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "alphaxiv-paper-lookup",
        "defuddle",
        "paper-review",
        "related-papers-scout"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "defuddle",
          "dependent": "alphaxiv-paper-lookup"
        },
        {
          "prerequisite": "paper-review",
          "dependent": "alphaxiv-paper-lookup"
        },
        {
          "prerequisite": "related-papers-scout",
          "dependent": "alphaxiv-paper-lookup"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with AlphaXiv Paper Lookup, then Defuddle, Paper Review Pipeline, then Related Papers Scout. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"alphaxiv-paper-lookup\", \"defuddle\", \"paper-review\", \"related-papers-scout\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0130",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Terraform Reviewer, SRE / DevOps Expert, and IaC Review Agent in order.",
      "gold_skills": [
        "terraform-reviewer",
        "sre-devops-expert",
        "iac-review-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "sre-devops-expert",
          "dependent": "terraform-reviewer"
        },
        {
          "prerequisite": "iac-review-agent",
          "dependent": "terraform-reviewer"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Terraform Reviewer, SRE / DevOps Expert, and IaC Review Agent in order.\n\nProposed gold_skills (ordered if composition):\n[\"terraform-reviewer\", \"sre-devops-expert\", \"iac-review-agent\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0131",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Morning, Today, and Stock CSV Downloader in order.",
      "gold_skills": [
        "morning",
        "today",
        "stock-csv-downloader"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "today",
          "dependent": "morning"
        },
        {
          "prerequisite": "stock-csv-downloader",
          "dependent": "morning"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Morning, Today, and Stock CSV Downloader in order.\n\nProposed gold_skills (ordered if composition):\n[\"morning\", \"today\", \"stock-csv-downloader\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0132",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Skill Composer”-style work, orchestrated under “Mission Control”. Which ordered skill chain fits?",
      "gold_skills": [
        "mission-control",
        "skill-composer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "skill-composer",
          "dependent": "mission-control"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Skill Composer”-style work, orchestrated under “Mission Control”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"mission-control\", \"skill-composer\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0133",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Hugging Face Dataset Viewer”-style work, orchestrated under “Overview”. Which ordered skill chain fits?",
      "gold_skills": [
        "hf-datasets",
        "hf-dataset-viewer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "hf-dataset-viewer",
          "dependent": "hf-datasets"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Hugging Face Dataset Viewer”-style work, orchestrated under “Overview”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"hf-datasets\", \"hf-dataset-viewer\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0134",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Daily Stock Check”-style work, orchestrated under “Execute HuggingFace Hub operations via the hf CLI”. Which ordered skill chain fits?",
      "gold_skills": [
        "hf-cli",
        "daily-stock-check"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "daily-stock-check",
          "dependent": "hf-cli"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Daily Stock Check”-style work, orchestrated under “Execute HuggingFace Hub operations via the hf CLI”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"hf-cli\", \"daily-stock-check\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0135",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like daily-strategy-post, Presentation Strategist, and Executive Briefing Generator in order.",
      "gold_skills": [
        "daily-strategy-post",
        "presentation-strategist",
        "executive-briefing"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "presentation-strategist",
          "dependent": "daily-strategy-post"
        },
        {
          "prerequisite": "executive-briefing",
          "dependent": "daily-strategy-post"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like daily-strategy-post, Presentation Strategist, and Executive Briefing Generator in order.\n\nProposed gold_skills (ordered if composition):\n[\"daily-strategy-post\", \"presentation-strategist\", \"executive-briefing\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0136",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Hugging Face Models”-style work, orchestrated under “Hugging Face Inference Endpoints”. Which ordered skill chain fits?",
      "gold_skills": [
        "hf-endpoints",
        "hf-models"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "hf-models",
          "dependent": "hf-endpoints"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Hugging Face Models”-style work, orchestrated under “Hugging Face Inference Endpoints”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"hf-endpoints\", \"hf-models\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0137",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Developer Perspective Analyzer, then Workflow Miner, QA / Test Expert, Semantic Guard, then Test Suite. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "role-developer",
        "workflow-miner",
        "qa-test-expert",
        "semantic-guard",
        "test-suite"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "workflow-miner",
          "dependent": "role-developer"
        },
        {
          "prerequisite": "qa-test-expert",
          "dependent": "role-developer"
        },
        {
          "prerequisite": "semantic-guard",
          "dependent": "role-developer"
        },
        {
          "prerequisite": "test-suite",
          "dependent": "role-developer"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Developer Perspective Analyzer, then Workflow Miner, QA / Test Expert, Semantic Guard, then Test Suite. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"role-developer\", \"workflow-miner\", \"qa-test-expert\", \"semantic-guard\", \"test-suite\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0138",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Weekly Stock Update”-style work, orchestrated under “PM Sentinel”. Which ordered skill chain fits?",
      "gold_skills": [
        "pm-sentinel",
        "weekly-stock-update"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "weekly-stock-update",
          "dependent": "pm-sentinel"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Weekly Stock Update”-style work, orchestrated under “PM Sentinel”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"pm-sentinel\", \"weekly-stock-update\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0139",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Meeting Digest”-style work, orchestrated under “Planning Meeting Sentinel”. Which ordered skill chain fits?",
      "gold_skills": [
        "planning-meeting-sentinel",
        "meeting-digest"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "meeting-digest",
          "dependent": "planning-meeting-sentinel"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Meeting Digest”-style work, orchestrated under “Planning Meeting Sentinel”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"planning-meeting-sentinel\", \"meeting-digest\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0140",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Jira/Linear Adapter, Sprint Orchestrator, and Commit-to-Issue in order.",
      "gold_skills": [
        "jira-linear-adapter",
        "sprint-orchestrator",
        "commit-to-issue"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "sprint-orchestrator",
          "dependent": "jira-linear-adapter"
        },
        {
          "prerequisite": "commit-to-issue",
          "dependent": "jira-linear-adapter"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Jira/Linear Adapter, Sprint Orchestrator, and Commit-to-Issue in order.\n\nProposed gold_skills (ordered if composition):\n[\"jira-linear-adapter\", \"sprint-orchestrator\", \"commit-to-issue\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0141",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Hugging Face Hub CLI, then Hugging Face Discussions, Hugging Face Inference Endpoints, then Hugging Face Models. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "hf-hub",
        "hf-discussions",
        "hf-endpoints",
        "hf-models"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "hf-discussions",
          "dependent": "hf-hub"
        },
        {
          "prerequisite": "hf-endpoints",
          "dependent": "hf-hub"
        },
        {
          "prerequisite": "hf-models",
          "dependent": "hf-hub"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Hugging Face Hub CLI, then Hugging Face Discussions, Hugging Face Inference Endpoints, then Hugging Face Models. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"hf-hub\", \"hf-discussions\", \"hf-endpoints\", \"hf-models\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0142",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Google Calendar, Google Workspace Workflows, and Gmail in order.",
      "gold_skills": [
        "gws-calendar",
        "gws-workflows",
        "gws-gmail"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "gws-workflows",
          "dependent": "gws-calendar"
        },
        {
          "prerequisite": "gws-gmail",
          "dependent": "gws-calendar"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Google Calendar, Google Workspace Workflows, and Gmail in order.\n\nProposed gold_skills (ordered if composition):\n[\"gws-calendar\", \"gws-workflows\", \"gws-gmail\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0143",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like AutoSkill Judge, AutoSkill Evolve, and Skill Optimizer in order.",
      "gold_skills": [
        "autoskill-judge",
        "autoskill-evolve",
        "skill-optimizer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "autoskill-evolve",
          "dependent": "autoskill-judge"
        },
        {
          "prerequisite": "skill-optimizer",
          "dependent": "autoskill-judge"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like AutoSkill Judge, AutoSkill Evolve, and Skill Optimizer in order.\n\nProposed gold_skills (ordered if composition):\n[\"autoskill-judge\", \"autoskill-evolve\", \"skill-optimizer\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0144",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Trade Hypothesis Ideator, Daily Stock Check, and Backtest Expert in order.",
      "gold_skills": [
        "trading-trade-hypothesis-ideator",
        "daily-stock-check",
        "trading-backtest-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "daily-stock-check",
          "dependent": "trading-trade-hypothesis-ideator"
        },
        {
          "prerequisite": "trading-backtest-expert",
          "dependent": "trading-trade-hypothesis-ideator"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Trade Hypothesis Ideator, Daily Stock Check, and Backtest Expert in order.\n\nProposed gold_skills (ordered if composition):\n[\"trading-trade-hypothesis-ideator\", \"daily-stock-check\", \"trading-backtest-expert\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0145",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Theme Factory Skill, Design System Management, and Anthropic Brand Styling in order.",
      "gold_skills": [
        "anthropic-theme-factory",
        "kwp-design-design-system-management",
        "anthropic-brand-guidelines"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "kwp-design-design-system-management",
          "dependent": "anthropic-theme-factory"
        },
        {
          "prerequisite": "anthropic-brand-guidelines",
          "dependent": "anthropic-theme-factory"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Theme Factory Skill, Design System Management, and Anthropic Brand Styling in order.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-theme-factory\", \"kwp-design-design-system-management\", \"anthropic-brand-guidelines\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0146",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Full-Stack Planner, then Mission Control, Plans, then Writing Plans. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "full-stack-planner",
        "mission-control",
        "plans",
        "sp-writing-plans"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "mission-control",
          "dependent": "full-stack-planner"
        },
        {
          "prerequisite": "plans",
          "dependent": "full-stack-planner"
        },
        {
          "prerequisite": "sp-writing-plans",
          "dependent": "full-stack-planner"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Full-Stack Planner, then Mission Control, Plans, then Writing Plans. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"full-stack-planner\", \"mission-control\", \"plans\", \"sp-writing-plans\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0147",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Risk Assessment”-style work, orchestrated under “Security Engineer Perspective Analyzer”. Which ordered skill chain fits?",
      "gold_skills": [
        "role-security-engineer",
        "kwp-operations-risk-assessment"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "kwp-operations-risk-assessment",
          "dependent": "role-security-engineer"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Risk Assessment”-style work, orchestrated under “Security Engineer Perspective Analyzer”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"role-security-engineer\", \"kwp-operations-risk-assessment\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0148",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Today”-style work, orchestrated under “Setup Doctor”. Which ordered skill chain fits?",
      "gold_skills": [
        "setup-doctor",
        "today"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "today",
          "dependent": "setup-doctor"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Today”-style work, orchestrated under “Setup Doctor”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"setup-doctor\", \"today\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0149",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like AlphaEar Reporter, AlphaEar Logic Visualizer, and Daily Stock Check in order.",
      "gold_skills": [
        "alphaear-reporter",
        "alphaear-logic-visualizer",
        "daily-stock-check"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "alphaear-logic-visualizer",
          "dependent": "alphaear-reporter"
        },
        {
          "prerequisite": "daily-stock-check",
          "dependent": "alphaear-reporter"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like AlphaEar Reporter, AlphaEar Logic Visualizer, and Daily Stock Check in order.\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-reporter\", \"alphaear-logic-visualizer\", \"daily-stock-check\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0150",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Content Creation Skill”-style work, orchestrated under “PM Go-to-Market”. Which ordered skill chain fits?",
      "gold_skills": [
        "pm-go-to-market",
        "kwp-marketing-content-creation"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "kwp-marketing-content-creation",
          "dependent": "pm-go-to-market"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Content Creation Skill”-style work, orchestrated under “PM Go-to-Market”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"pm-go-to-market\", \"kwp-marketing-content-creation\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0151",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Defuddle, Agent Browser, and Transcribee in order.",
      "gold_skills": [
        "defuddle",
        "agent-browser",
        "transcribee"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "agent-browser",
          "dependent": "defuddle"
        },
        {
          "prerequisite": "transcribee",
          "dependent": "defuddle"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Defuddle, Agent Browser, and Transcribee in order.\n\nProposed gold_skills (ordered if composition):\n[\"defuddle\", \"agent-browser\", \"transcribee\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0152",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Gmail”-style work, orchestrated under “smart-meeting-scheduler”. Which ordered skill chain fits?",
      "gold_skills": [
        "smart-meeting-scheduler",
        "gws-gmail"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "gws-gmail",
          "dependent": "smart-meeting-scheduler"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Gmail”-style work, orchestrated under “smart-meeting-scheduler”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"smart-meeting-scheduler\", \"gws-gmail\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0153",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “SEFO Federation”-style work, orchestrated under “SEFO Governance”. Which ordered skill chain fits?",
      "gold_skills": [
        "sefo-governance",
        "sefo-federation"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "sefo-federation",
          "dependent": "sefo-governance"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “SEFO Federation”-style work, orchestrated under “SEFO Governance”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"sefo-governance\", \"sefo-federation\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0154",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “AlphaEar Sentiment”-style work, orchestrated under “Market News Analyst”. Which ordered skill chain fits?",
      "gold_skills": [
        "trading-market-news-analyst",
        "alphaear-sentiment"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "alphaear-sentiment",
          "dependent": "trading-market-news-analyst"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “AlphaEar Sentiment”-style work, orchestrated under “Market News Analyst”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"trading-market-news-analyst\", \"alphaear-sentiment\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0155",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Trackio - Experiment Tracking for ML Training, then TRL Training on Hugging Face Jobs, AI Quality Evaluator, then Weekly Stock Update. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "hf-trackio",
        "hf-model-trainer",
        "ai-quality-evaluator",
        "weekly-stock-update"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "hf-model-trainer",
          "dependent": "hf-trackio"
        },
        {
          "prerequisite": "ai-quality-evaluator",
          "dependent": "hf-trackio"
        },
        {
          "prerequisite": "weekly-stock-update",
          "dependent": "hf-trackio"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Trackio - Experiment Tracking for ML Training, then TRL Training on Hugging Face Jobs, AI Quality Evaluator, then Weekly Stock Update. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"hf-trackio\", \"hf-model-trainer\", \"ai-quality-evaluator\", \"weekly-stock-update\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0156",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with tab-llm-agents, then tab-stock-sync, Daily Stock Check, then tab-genai-features. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "tab-llm-agents",
        "tab-stock-sync",
        "daily-stock-check",
        "tab-genai-features"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "tab-stock-sync",
          "dependent": "tab-llm-agents"
        },
        {
          "prerequisite": "daily-stock-check",
          "dependent": "tab-llm-agents"
        },
        {
          "prerequisite": "tab-genai-features",
          "dependent": "tab-llm-agents"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with tab-llm-agents, then tab-stock-sync, Daily Stock Check, then tab-genai-features. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"tab-llm-agents\", \"tab-stock-sync\", \"daily-stock-check\", \"tab-genai-features\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0157",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Auto Fine-Tune Planner, Today, and Overview in order.",
      "gold_skills": [
        "hf-auto-finetune-planner",
        "today",
        "hf-evaluation"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "today",
          "dependent": "hf-auto-finetune-planner"
        },
        {
          "prerequisite": "hf-evaluation",
          "dependent": "hf-auto-finetune-planner"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Auto Fine-Tune Planner, Today, and Overview in order.\n\nProposed gold_skills (ordered if composition):\n[\"hf-auto-finetune-planner\", \"today\", \"hf-evaluation\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0158",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “smart-meeting-scheduler”-style work, orchestrated under “Proactive Meeting Scheduler”. Which ordered skill chain fits?",
      "gold_skills": [
        "proactive-meeting-scheduler",
        "smart-meeting-scheduler"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "smart-meeting-scheduler",
          "dependent": "proactive-meeting-scheduler"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “smart-meeting-scheduler”-style work, orchestrated under “Proactive Meeting Scheduler”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"proactive-meeting-scheduler\", \"smart-meeting-scheduler\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0159",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Market Top Detector Skill”-style work, orchestrated under “Uptrend Analyzer Skill”. Which ordered skill chain fits?",
      "gold_skills": [
        "trading-uptrend-analyzer",
        "trading-market-top-detector"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "trading-market-top-detector",
          "dependent": "trading-uptrend-analyzer"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Market Top Detector Skill”-style work, orchestrated under “Uptrend Analyzer Skill”. Which ordered skill chain fits?\n\nProposed gold_skills (ordered if composition):\n[\"trading-uptrend-analyzer\", \"trading-market-top-detector\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0160",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Local Development Runner, Service Health Doctor, and SRE / DevOps Expert in order.",
      "gold_skills": [
        "local-dev-runner",
        "service-health-doctor",
        "sre-devops-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "service-health-doctor",
          "dependent": "local-dev-runner"
        },
        {
          "prerequisite": "sre-devops-expert",
          "dependent": "local-dev-runner"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Local Development Runner, Service Health Doctor, and SRE / DevOps Expert in order.\n\nProposed gold_skills (ordered if composition):\n[\"local-dev-runner\", \"service-health-doctor\", \"sre-devops-expert\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0161",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Cross-Role Topic Analysis Dispatcher, then Today, CSO Perspective Analyzer, CEO Perspective Analyzer, then UX Designer Perspective Analyzer. Return the ordered skill ids that must compose together.",
      "gold_skills": [
        "role-dispatcher",
        "today",
        "role-cso",
        "role-ceo",
        "role-ux-designer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "today",
          "dependent": "role-dispatcher"
        },
        {
          "prerequisite": "role-cso",
          "dependent": "role-dispatcher"
        },
        {
          "prerequisite": "role-ceo",
          "dependent": "role-dispatcher"
        },
        {
          "prerequisite": "role-ux-designer",
          "dependent": "role-dispatcher"
        }
      ],
      "judge_verdict": "accept",
      "judge_confidence": 0.82,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Cross-Role Topic Analysis Dispatcher, then Today, CSO Perspective Analyzer, CEO Perspective Analyzer, then UX Designer Perspective Analyzer. Return the ordered skill ids that must compose together.\n\nProposed gold_skills (ordered if composition):\n[\"role-dispatcher\", \"today\", \"role-cso\", \"role-ceo\", \"role-ux-designer\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0162",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “NotebookLM Research: Web & Drive Discovery”-style work, orchestrated under “NotebookLM Studio: Content Generation”. Which ordered skill chain fits? [variant 161]",
      "gold_skills": [
        "notebooklm-studio",
        "notebooklm-research"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "notebooklm-research",
          "dependent": "notebooklm-studio"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “NotebookLM Research: Web & Drive Discovery”-style work, orchestrated under “NotebookLM Studio: Content Generation”. Which ordered skill chain fits? [variant 161]\n\nProposed gold_skills (ordered if composition):\n[\"notebooklm-studio\", \"notebooklm-research\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0163",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Overview, then TRL Training on Hugging Face Jobs, Daily Stock Check, then AI Quality Evaluator. Return the ordered skill ids that must compose together. [variant 162]",
      "gold_skills": [
        "hf-evaluation",
        "hf-model-trainer",
        "daily-stock-check",
        "ai-quality-evaluator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "hf-model-trainer",
          "dependent": "hf-evaluation"
        },
        {
          "prerequisite": "daily-stock-check",
          "dependent": "hf-evaluation"
        },
        {
          "prerequisite": "ai-quality-evaluator",
          "dependent": "hf-evaluation"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Overview, then TRL Training on Hugging Face Jobs, Daily Stock Check, then AI Quality Evaluator. Return the ordered skill ids that must compose together. [variant 162]\n\nProposed gold_skills (ordered if composition):\n[\"hf-evaluation\", \"hf-model-trainer\", \"daily-stock-check\", \"ai-quality-evaluator\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0164",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with tab-kiwoom, then tab-strategy-comparison, Daily Stock Check, tab-stock-sync, then US Stock Analysis. Return the ordered skill ids that must compose together. [variant 163]",
      "gold_skills": [
        "tab-kiwoom",
        "tab-strategy-comparison",
        "daily-stock-check",
        "tab-stock-sync",
        "trading-us-stock-analysis"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "tab-strategy-comparison",
          "dependent": "tab-kiwoom"
        },
        {
          "prerequisite": "daily-stock-check",
          "dependent": "tab-kiwoom"
        },
        {
          "prerequisite": "tab-stock-sync",
          "dependent": "tab-kiwoom"
        },
        {
          "prerequisite": "trading-us-stock-analysis",
          "dependent": "tab-kiwoom"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nMulti-step automation: start with tab-kiwoom, then tab-strategy-comparison, Daily Stock Check, tab-stock-sync, then US Stock Analysis. Return the ordered skill ids that must compose together. [variant 163]\n\nProposed gold_skills (ordered if composition):\n[\"tab-kiwoom\", \"tab-strategy-comparison\", \"daily-stock-check\", \"tab-stock-sync\", \"trading-us-stock-analysis\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0165",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with AutoResearchClaw, then NLM arXiv Slides: arXiv Paper to Expert Slide Deck Pipeline, Paper Review Pipeline, AlphaXiv Paper Lookup, then Related Papers Scout. Return the ordered skill ids that must compose together. [variant 164]",
      "gold_skills": [
        "auto-research",
        "nlm-arxiv-slides",
        "paper-review",
        "alphaxiv-paper-lookup",
        "related-papers-scout"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "nlm-arxiv-slides",
          "dependent": "auto-research"
        },
        {
          "prerequisite": "paper-review",
          "dependent": "auto-research"
        },
        {
          "prerequisite": "alphaxiv-paper-lookup",
          "dependent": "auto-research"
        },
        {
          "prerequisite": "related-papers-scout",
          "dependent": "auto-research"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nMulti-step automation: start with AutoResearchClaw, then NLM arXiv Slides: arXiv Paper to Expert Slide Deck Pipeline, Paper Review Pipeline, AlphaXiv Paper Lookup, then Related Papers Scout. Return the ordered skill ids that must compose together. [variant 164]\n\nProposed gold_skills (ordered if composition):\n[\"auto-research\", \"nlm-arxiv-slides\", \"paper-review\", \"alphaxiv-paper-lookup\", \"related-papers-scout\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0166",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Autonomous Loops Skill, Today, and Ralph Loop in order. [variant 165]",
      "gold_skills": [
        "ecc-autonomous-loops",
        "today",
        "ralph-loop"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "today",
          "dependent": "ecc-autonomous-loops"
        },
        {
          "prerequisite": "ralph-loop",
          "dependent": "ecc-autonomous-loops"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Autonomous Loops Skill, Today, and Ralph Loop in order. [variant 165]\n\nProposed gold_skills (ordered if composition):\n[\"ecc-autonomous-loops\", \"today\", \"ralph-loop\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0167",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Gmail Daily Triage, then Google Calendar, Google Workspace CLI, then DOCX creation, editing, and analysis. Return the ordered skill ids that must compose together. [variant 166]",
      "gold_skills": [
        "gmail-daily-triage",
        "gws-calendar",
        "gws-workspace",
        "anthropic-docx"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "gws-calendar",
          "dependent": "gmail-daily-triage"
        },
        {
          "prerequisite": "gws-workspace",
          "dependent": "gmail-daily-triage"
        },
        {
          "prerequisite": "anthropic-docx",
          "dependent": "gmail-daily-triage"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Gmail Daily Triage, then Google Calendar, Google Workspace CLI, then DOCX creation, editing, and analysis. Return the ordered skill ids that must compose together. [variant 166]\n\nProposed gold_skills (ordered if composition):\n[\"gmail-daily-triage\", \"gws-calendar\", \"gws-workspace\", \"anthropic-docx\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0168",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Daily Stock Check”-style work, orchestrated under “US Stock Analysis”. Which ordered skill chain fits? [variant 167]",
      "gold_skills": [
        "trading-us-stock-analysis",
        "daily-stock-check"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "daily-stock-check",
          "dependent": "trading-us-stock-analysis"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Daily Stock Check”-style work, orchestrated under “US Stock Analysis”. Which ordered skill chain fits? [variant 167]\n\nProposed gold_skills (ordered if composition):\n[\"trading-us-stock-analysis\", \"daily-stock-check\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0169",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with US Market Bubble Detection Skill (Revised v2.1), then AlphaEar DeepEar Lite, Daily Stock Check, then Weekly Stock Update. Return the ordered skill ids that must compose together. [variant 168]",
      "gold_skills": [
        "trading-us-market-bubble-detector",
        "alphaear-deepear-lite",
        "daily-stock-check",
        "weekly-stock-update"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "alphaear-deepear-lite",
          "dependent": "trading-us-market-bubble-detector"
        },
        {
          "prerequisite": "daily-stock-check",
          "dependent": "trading-us-market-bubble-detector"
        },
        {
          "prerequisite": "weekly-stock-update",
          "dependent": "trading-us-market-bubble-detector"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nMulti-step automation: start with US Market Bubble Detection Skill (Revised v2.1), then AlphaEar DeepEar Lite, Daily Stock Check, then Weekly Stock Update. Return the ordered skill ids that must compose together. [variant 168]\n\nProposed gold_skills (ordered if composition):\n[\"trading-us-market-bubble-detector\", \"alphaear-deepear-lite\", \"daily-stock-check\", \"weekly-stock-update\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0170",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “tab-bollinger-refresh”-style work, orchestrated under “tab-turtle-refresh”. Which ordered skill chain fits? [variant 169]",
      "gold_skills": [
        "tab-turtle-refresh",
        "tab-bollinger-refresh"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "tab-bollinger-refresh",
          "dependent": "tab-turtle-refresh"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “tab-bollinger-refresh”-style work, orchestrated under “tab-turtle-refresh”. Which ordered skill chain fits? [variant 169]\n\nProposed gold_skills (ordered if composition):\n[\"tab-turtle-refresh\", \"tab-bollinger-refresh\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0171",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like NLM Deep Learn: Accelerated Learning Pipeline, NotebookLM Studio: Content Generation, and NLM Discovery Lab: Product Discovery Lab Pipeline in order. [variant 170]",
      "gold_skills": [
        "nlm-deep-learn",
        "notebooklm-studio",
        "nlm-discovery-lab"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "notebooklm-studio",
          "dependent": "nlm-deep-learn"
        },
        {
          "prerequisite": "nlm-discovery-lab",
          "dependent": "nlm-deep-learn"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like NLM Deep Learn: Accelerated Learning Pipeline, NotebookLM Studio: Content Generation, and NLM Discovery Lab: Product Discovery Lab Pipeline in order. [variant 170]\n\nProposed gold_skills (ordered if composition):\n[\"nlm-deep-learn\", \"notebooklm-studio\", \"nlm-discovery-lab\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0172",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like NLM Slides: Local Document to Expert Slide Deck Pipeline, NotebookLM Studio: Content Generation, and NotebookLM: Core Notebook Management in order. [variant 171]",
      "gold_skills": [
        "nlm-slides",
        "notebooklm-studio",
        "notebooklm"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "notebooklm-studio",
          "dependent": "nlm-slides"
        },
        {
          "prerequisite": "notebooklm",
          "dependent": "nlm-slides"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like NLM Slides: Local Document to Expert Slide Deck Pipeline, NotebookLM Studio: Content Generation, and NotebookLM: Core Notebook Management in order. [variant 171]\n\nProposed gold_skills (ordered if composition):\n[\"nlm-slides\", \"notebooklm-studio\", \"notebooklm\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0173",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with TRL Training on Hugging Face Jobs, then Execute HuggingFace Hub operations via the hf CLI, AlphaEar Predictor, then Daily Stock Check. Return the ordered skill ids that must compose together. [variant 172]",
      "gold_skills": [
        "hf-model-trainer",
        "hf-cli",
        "alphaear-predictor",
        "daily-stock-check"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "hf-cli",
          "dependent": "hf-model-trainer"
        },
        {
          "prerequisite": "alphaear-predictor",
          "dependent": "hf-model-trainer"
        },
        {
          "prerequisite": "daily-stock-check",
          "dependent": "hf-model-trainer"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nMulti-step automation: start with TRL Training on Hugging Face Jobs, then Execute HuggingFace Hub operations via the hf CLI, AlphaEar Predictor, then Daily Stock Check. Return the ordered skill ids that must compose together. [variant 172]\n\nProposed gold_skills (ordered if composition):\n[\"hf-model-trainer\", \"hf-cli\", \"alphaear-predictor\", \"daily-stock-check\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0174",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Performance Profiler”-style work, orchestrated under “DB Expert”. Which ordered skill chain fits? [variant 173]",
      "gold_skills": [
        "db-expert",
        "performance-profiler"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "performance-profiler",
          "dependent": "db-expert"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Performance Profiler”-style work, orchestrated under “DB Expert”. Which ordered skill chain fits? [variant 173]\n\nProposed gold_skills (ordered if composition):\n[\"db-expert\", \"performance-profiler\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0175",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Demo Forge, then Transcribee, PR Review Captain, then Visual Explainer. Return the ordered skill ids that must compose together. [variant 174]",
      "gold_skills": [
        "demo-forge",
        "transcribee",
        "pr-review-captain",
        "visual-explainer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "transcribee",
          "dependent": "demo-forge"
        },
        {
          "prerequisite": "pr-review-captain",
          "dependent": "demo-forge"
        },
        {
          "prerequisite": "visual-explainer",
          "dependent": "demo-forge"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Demo Forge, then Transcribee, PR Review Captain, then Visual Explainer. Return the ordered skill ids that must compose together. [variant 174]\n\nProposed gold_skills (ordered if composition):\n[\"demo-forge\", \"transcribee\", \"pr-review-captain\", \"visual-explainer\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0176",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Recipe: Audit External Drive Sharing, Google Drive, and Compliance & Governance in order. [variant 175]",
      "gold_skills": [
        "gws-recipe-audit-sharing",
        "gws-drive",
        "compliance-governance"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "gws-drive",
          "dependent": "gws-recipe-audit-sharing"
        },
        {
          "prerequisite": "compliance-governance",
          "dependent": "gws-recipe-audit-sharing"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Recipe: Audit External Drive Sharing, Google Drive, and Compliance & Governance in order. [variant 175]\n\nProposed gold_skills (ordered if composition):\n[\"gws-recipe-audit-sharing\", \"gws-drive\", \"compliance-governance\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0177",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Sprint Orchestrator, then github-sprint-digest, SOD Ship, then Commit-to-Issue. Return the ordered skill ids that must compose together. [variant 176]",
      "gold_skills": [
        "sprint-orchestrator",
        "github-sprint-digest",
        "sod-ship",
        "commit-to-issue"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "github-sprint-digest",
          "dependent": "sprint-orchestrator"
        },
        {
          "prerequisite": "sod-ship",
          "dependent": "sprint-orchestrator"
        },
        {
          "prerequisite": "commit-to-issue",
          "dependent": "sprint-orchestrator"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Sprint Orchestrator, then github-sprint-digest, SOD Ship, then Commit-to-Issue. Return the ordered skill ids that must compose together. [variant 176]\n\nProposed gold_skills (ordered if composition):\n[\"sprint-orchestrator\", \"github-sprint-digest\", \"sod-ship\", \"commit-to-issue\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0178",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “DOCX creation, editing, and analysis”-style work, orchestrated under “Doc Co-Authoring Workflow”. Which ordered skill chain fits? [variant 177]",
      "gold_skills": [
        "anthropic-doc-coauthoring",
        "anthropic-docx"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "anthropic-docx",
          "dependent": "anthropic-doc-coauthoring"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “DOCX creation, editing, and analysis”-style work, orchestrated under “Doc Co-Authoring Workflow”. Which ordered skill chain fits? [variant 177]\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-doc-coauthoring\", \"anthropic-docx\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0179",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Compliance Gate, then Dependency Auditor, Release Commander, then Deep Review. Return the ordered skill ids that must compose together. [variant 178]",
      "gold_skills": [
        "compliance-gate",
        "dependency-auditor",
        "release-commander",
        "deep-review"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "dependency-auditor",
          "dependent": "compliance-gate"
        },
        {
          "prerequisite": "release-commander",
          "dependent": "compliance-gate"
        },
        {
          "prerequisite": "deep-review",
          "dependent": "compliance-gate"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Compliance Gate, then Dependency Auditor, Release Commander, then Deep Review. Return the ordered skill ids that must compose together. [variant 178]\n\nProposed gold_skills (ordered if composition):\n[\"compliance-gate\", \"dependency-auditor\", \"release-commander\", \"deep-review\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0180",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Frontend Expert”-style work, orchestrated under “i18n Sync Manager”. Which ordered skill chain fits? [variant 179]",
      "gold_skills": [
        "i18n-sync",
        "frontend-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "frontend-expert",
          "dependent": "i18n-sync"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Frontend Expert”-style work, orchestrated under “i18n Sync Manager”. Which ordered skill chain fits? [variant 179]\n\nProposed gold_skills (ordered if composition):\n[\"i18n-sync\", \"frontend-expert\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0181",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Paperclip Control, then Paperclip Setup, Paperclip Tasks, then Paperclip Agents. Return the ordered skill ids that must compose together. [variant 180]",
      "gold_skills": [
        "paperclip-control",
        "paperclip-setup",
        "paperclip-tasks",
        "paperclip-agents"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "paperclip-setup",
          "dependent": "paperclip-control"
        },
        {
          "prerequisite": "paperclip-tasks",
          "dependent": "paperclip-control"
        },
        {
          "prerequisite": "paperclip-agents",
          "dependent": "paperclip-control"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Paperclip Control, then Paperclip Setup, Paperclip Tasks, then Paperclip Agents. Return the ordered skill ids that must compose together. [variant 180]\n\nProposed gold_skills (ordered if composition):\n[\"paperclip-control\", \"paperclip-setup\", \"paperclip-tasks\", \"paperclip-agents\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0182",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like tab-dashboard-summary, Daily Stock Check, and tab-market-environment in order. [variant 181]",
      "gold_skills": [
        "tab-dashboard-summary",
        "daily-stock-check",
        "tab-market-environment"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "daily-stock-check",
          "dependent": "tab-dashboard-summary"
        },
        {
          "prerequisite": "tab-market-environment",
          "dependent": "tab-dashboard-summary"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like tab-dashboard-summary, Daily Stock Check, and tab-market-environment in order. [variant 181]\n\nProposed gold_skills (ordered if composition):\n[\"tab-dashboard-summary\", \"daily-stock-check\", \"tab-market-environment\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0183",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like PM Marketing and Growth, Brand Voice Skill, and Campaign Planning Skill in order. [variant 182]",
      "gold_skills": [
        "pm-marketing-growth",
        "kwp-marketing-brand-voice",
        "kwp-marketing-campaign-planning"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "kwp-marketing-brand-voice",
          "dependent": "pm-marketing-growth"
        },
        {
          "prerequisite": "kwp-marketing-campaign-planning",
          "dependent": "pm-marketing-growth"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like PM Marketing and Growth, Brand Voice Skill, and Campaign Planning Skill in order. [variant 182]\n\nProposed gold_skills (ordered if composition):\n[\"pm-marketing-growth\", \"kwp-marketing-brand-voice\", \"kwp-marketing-campaign-planning\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0184",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Trending Intelligence, md-to-notion, and Slack Messaging Best Practices in order. [variant 183]",
      "gold_skills": [
        "hf-trending-intelligence",
        "md-to-notion",
        "kwp-slack-slack-messaging"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "md-to-notion",
          "dependent": "hf-trending-intelligence"
        },
        {
          "prerequisite": "kwp-slack-slack-messaging",
          "dependent": "hf-trending-intelligence"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Trending Intelligence, md-to-notion, and Slack Messaging Best Practices in order. [variant 183]\n\nProposed gold_skills (ordered if composition):\n[\"hf-trending-intelligence\", \"md-to-notion\", \"kwp-slack-slack-messaging\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0185",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like AlphaEar Search, AlphaEar News, and AlphaEar Stock in order. [variant 184]",
      "gold_skills": [
        "alphaear-search",
        "alphaear-news",
        "alphaear-stock"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "alphaear-news",
          "dependent": "alphaear-search"
        },
        {
          "prerequisite": "alphaear-stock",
          "dependent": "alphaear-search"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like AlphaEar Search, AlphaEar News, and AlphaEar Stock in order. [variant 184]\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-search\", \"alphaear-news\", \"alphaear-stock\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0186",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Prompt Architect, Technical Writer, and Prompt Transformer in order. [variant 185]",
      "gold_skills": [
        "prompt-architect",
        "technical-writer",
        "prompt-transformer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "technical-writer",
          "dependent": "prompt-architect"
        },
        {
          "prerequisite": "prompt-transformer",
          "dependent": "prompt-architect"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Prompt Architect, Technical Writer, and Prompt Transformer in order. [variant 185]\n\nProposed gold_skills (ordered if composition):\n[\"prompt-architect\", \"technical-writer\", \"prompt-transformer\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0187",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with knowledge-to-paper, then PPTX Skill, Cognee, then AutoResearchClaw. Return the ordered skill ids that must compose together. [variant 186]",
      "gold_skills": [
        "knowledge-to-paper",
        "anthropic-pptx",
        "cognee",
        "auto-research"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "anthropic-pptx",
          "dependent": "knowledge-to-paper"
        },
        {
          "prerequisite": "cognee",
          "dependent": "knowledge-to-paper"
        },
        {
          "prerequisite": "auto-research",
          "dependent": "knowledge-to-paper"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nMulti-step automation: start with knowledge-to-paper, then PPTX Skill, Cognee, then AutoResearchClaw. Return the ordered skill ids that must compose together. [variant 186]\n\nProposed gold_skills (ordered if composition):\n[\"knowledge-to-paper\", \"anthropic-pptx\", \"cognee\", \"auto-research\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0188",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Paperclip Tasks, Paperclip Agents, and Paperclip Setup in order. [variant 187]",
      "gold_skills": [
        "paperclip-tasks",
        "paperclip-agents",
        "paperclip-setup"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "paperclip-agents",
          "dependent": "paperclip-tasks"
        },
        {
          "prerequisite": "paperclip-setup",
          "dependent": "paperclip-tasks"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Paperclip Tasks, Paperclip Agents, and Paperclip Setup in order. [variant 187]\n\nProposed gold_skills (ordered if composition):\n[\"paperclip-tasks\", \"paperclip-agents\", \"paperclip-setup\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0189",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Paper Auto-Classifier, then Paper Archive, Paper Review Pipeline, then AlphaXiv Paper Lookup. Return the ordered skill ids that must compose together. [variant 188]",
      "gold_skills": [
        "paper-auto-classifier",
        "paper-archive",
        "paper-review",
        "alphaxiv-paper-lookup"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "paper-archive",
          "dependent": "paper-auto-classifier"
        },
        {
          "prerequisite": "paper-review",
          "dependent": "paper-auto-classifier"
        },
        {
          "prerequisite": "alphaxiv-paper-lookup",
          "dependent": "paper-auto-classifier"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Paper Auto-Classifier, then Paper Archive, Paper Review Pipeline, then AlphaXiv Paper Lookup. Return the ordered skill ids that must compose together. [variant 188]\n\nProposed gold_skills (ordered if composition):\n[\"paper-auto-classifier\", \"paper-archive\", \"paper-review\", \"alphaxiv-paper-lookup\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0190",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Google Calendar”-style work, orchestrated under “Recipe: Create Calendar Events from a Sheet”. Which ordered skill chain fits? [variant 189]",
      "gold_skills": [
        "gws-recipe-events-from-sheet",
        "gws-calendar"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "gws-calendar",
          "dependent": "gws-recipe-events-from-sheet"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Google Calendar”-style work, orchestrated under “Recipe: Create Calendar Events from a Sheet”. Which ordered skill chain fits? [variant 189]\n\nProposed gold_skills (ordered if composition):\n[\"gws-recipe-events-from-sheet\", \"gws-calendar\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0191",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “PDF Processing Guide”-style work, orchestrated under “Requirements for Outputs”. Which ordered skill chain fits? [variant 190]",
      "gold_skills": [
        "anthropic-xlsx",
        "anthropic-pdf"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "anthropic-pdf",
          "dependent": "anthropic-xlsx"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “PDF Processing Guide”-style work, orchestrated under “Requirements for Outputs”. Which ordered skill chain fits? [variant 190]\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-xlsx\", \"anthropic-pdf\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0192",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with NLM Launch Kit: Go-to-Market Launch Kit Generator Pipeline, then NotebookLM: Core Notebook Management, PM Marketing and Growth, then NLM Strategy Hub: Product Strategy Research Hub Pipeline. Return the ordered skill ids that must compose together. [variant 191]",
      "gold_skills": [
        "nlm-launch-kit",
        "notebooklm",
        "pm-marketing-growth",
        "nlm-strategy-hub"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "notebooklm",
          "dependent": "nlm-launch-kit"
        },
        {
          "prerequisite": "pm-marketing-growth",
          "dependent": "nlm-launch-kit"
        },
        {
          "prerequisite": "nlm-strategy-hub",
          "dependent": "nlm-launch-kit"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nMulti-step automation: start with NLM Launch Kit: Go-to-Market Launch Kit Generator Pipeline, then NotebookLM: Core Notebook Management, PM Marketing and Growth, then NLM Strategy Hub: Product Strategy Research Hub Pipeline. Return the ordered skill ids that must compose together. [variant 191]\n\nProposed gold_skills (ordered if composition):\n[\"nlm-launch-kit\", \"notebooklm\", \"pm-marketing-growth\", \"nlm-strategy-hub\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0193",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “Slack Messaging Best Practices”-style work, orchestrated under “Slack GIF Creator”. Which ordered skill chain fits? [variant 192]",
      "gold_skills": [
        "anthropic-slack-gif-creator",
        "kwp-slack-slack-messaging"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "kwp-slack-slack-messaging",
          "dependent": "anthropic-slack-gif-creator"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “Slack Messaging Best Practices”-style work, orchestrated under “Slack GIF Creator”. Which ordered skill chain fits? [variant 192]\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-slack-gif-creator\", \"kwp-slack-slack-messaging\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0194",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with tab-genai-features, then Daily Stock Check, tab-llm-agents, then tab-event-detect. Return the ordered skill ids that must compose together. [variant 193]",
      "gold_skills": [
        "tab-genai-features",
        "daily-stock-check",
        "tab-llm-agents",
        "tab-event-detect"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "daily-stock-check",
          "dependent": "tab-genai-features"
        },
        {
          "prerequisite": "tab-llm-agents",
          "dependent": "tab-genai-features"
        },
        {
          "prerequisite": "tab-event-detect",
          "dependent": "tab-genai-features"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nMulti-step automation: start with tab-genai-features, then Daily Stock Check, tab-llm-agents, then tab-event-detect. Return the ordered skill ids that must compose together. [variant 193]\n\nProposed gold_skills (ordered if composition):\n[\"tab-genai-features\", \"daily-stock-check\", \"tab-llm-agents\", \"tab-event-detect\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0195",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Sentence Polisher, then Prompt Transformer, Guideline Generation, then GWS Email Reply. Return the ordered skill ids that must compose together. [variant 194]",
      "gold_skills": [
        "sentence-polisher",
        "prompt-transformer",
        "kwp-brand-voice-guideline-generation",
        "gws-email-reply"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "prompt-transformer",
          "dependent": "sentence-polisher"
        },
        {
          "prerequisite": "kwp-brand-voice-guideline-generation",
          "dependent": "sentence-polisher"
        },
        {
          "prerequisite": "gws-email-reply",
          "dependent": "sentence-polisher"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Sentence Polisher, then Prompt Transformer, Guideline Generation, then GWS Email Reply. Return the ordered skill ids that must compose together. [variant 194]\n\nProposed gold_skills (ordered if composition):\n[\"sentence-polisher\", \"prompt-transformer\", \"kwp-brand-voice-guideline-generation\", \"gws-email-reply\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0196",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like SRE / DevOps Expert, Local Development Runner, and CI Quality Gate in order. [variant 195]",
      "gold_skills": [
        "sre-devops-expert",
        "local-dev-runner",
        "ci-quality-gate"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "local-dev-runner",
          "dependent": "sre-devops-expert"
        },
        {
          "prerequisite": "ci-quality-gate",
          "dependent": "sre-devops-expert"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like SRE / DevOps Expert, Local Development Runner, and CI Quality Gate in order. [variant 195]\n\nProposed gold_skills (ordered if composition):\n[\"sre-devops-expert\", \"local-dev-runner\", \"ci-quality-gate\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0197",
      "setting": "multi_skill_composition",
      "query": "Combine calendar/email prep, equity screening, and reporting in one morning flow using skills like Critical Review, Today, and Deep Review in order. [variant 196]",
      "gold_skills": [
        "critical-review",
        "today",
        "deep-review"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "today",
          "dependent": "critical-review"
        },
        {
          "prerequisite": "deep-review",
          "dependent": "critical-review"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nCombine calendar/email prep, equity screening, and reporting in one morning flow using skills like Critical Review, Today, and Deep Review in order. [variant 196]\n\nProposed gold_skills (ordered if composition):\n[\"critical-review\", \"today\", \"deep-review\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0198",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Recipe: Send Personalized Emails from a Sheet, then Gmail, Google Sheets, then Google Chat. Return the ordered skill ids that must compose together. [variant 197]",
      "gold_skills": [
        "gws-recipe-personalized-emails",
        "gws-gmail",
        "gws-sheets",
        "gws-chat"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "gws-gmail",
          "dependent": "gws-recipe-personalized-emails"
        },
        {
          "prerequisite": "gws-sheets",
          "dependent": "gws-recipe-personalized-emails"
        },
        {
          "prerequisite": "gws-chat",
          "dependent": "gws-recipe-personalized-emails"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Recipe: Send Personalized Emails from a Sheet, then Gmail, Google Sheets, then Google Chat. Return the ordered skill ids that must compose together. [variant 197]\n\nProposed gold_skills (ordered if composition):\n[\"gws-recipe-personalized-emails\", \"gws-gmail\", \"gws-sheets\", \"gws-chat\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0199",
      "setting": "multi_skill_composition",
      "query": "I want a single workflow that first runs “SEFO Governance”-style work, orchestrated under “SEFO Orchestrator”. Which ordered skill chain fits? [variant 198]",
      "gold_skills": [
        "sefo-orchestrator",
        "sefo-governance"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "sefo-governance",
          "dependent": "sefo-orchestrator"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nI want a single workflow that first runs “SEFO Governance”-style work, orchestrated under “SEFO Orchestrator”. Which ordered skill chain fits? [variant 198]\n\nProposed gold_skills (ordered if composition):\n[\"sefo-orchestrator\", \"sefo-governance\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "MC-0200",
      "setting": "multi_skill_composition",
      "query": "Multi-step automation: start with Automation Strategist, then AI Quality Evaluator, Pipeline Builder, Compliance & Governance, then SRE / DevOps Expert. Return the ordered skill ids that must compose together. [variant 199]",
      "gold_skills": [
        "automation-strategist",
        "ai-quality-evaluator",
        "pipeline-builder",
        "compliance-governance",
        "sre-devops-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "composition_extraction",
      "composition_dependencies": [
        {
          "prerequisite": "ai-quality-evaluator",
          "dependent": "automation-strategist"
        },
        {
          "prerequisite": "pipeline-builder",
          "dependent": "automation-strategist"
        },
        {
          "prerequisite": "compliance-governance",
          "dependent": "automation-strategist"
        },
        {
          "prerequisite": "sre-devops-expert",
          "dependent": "automation-strategist"
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nMulti-step automation: start with Automation Strategist, then AI Quality Evaluator, Pipeline Builder, Compliance & Governance, then SRE / DevOps Expert. Return the ordered skill ids that must compose together. [variant 199]\n\nProposed gold_skills (ordered if composition):\n[\"automation-strategist\", \"ai-quality-evaluator\", \"pipeline-builder\", \"compliance-governance\", \"sre-devops-expert\"]\n\nSetting: multi_skill_composition\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0001",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: usability analysis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "ux-expert"
      ],
      "distractor_skills": [
        "agency-ux-architect",
        "kwp-design-accessibility-review",
        "frontend-expert",
        "design-architect",
        "compliance-governance",
        "kwp-design-design-system-management",
        "kwp-design-user-research",
        "agency-brand-guardian",
        "ci-quality-gate",
        "ui-suite",
        "agency-xr-interface-architect",
        "agency-security-engineer",
        "agency-accessibility-auditor",
        "kwp-design-design-critique",
        "backend-expert",
        "agency-ux-researcher",
        "agency-xr-immersive-developer",
        "agency-frontend-developer",
        "anthropic-frontend-design",
        "role-ux-designer",
        "kwp-engineering-code-review",
        "agency-visionos-spatial-engineer",
        "agency-app-store-optimizer",
        "security-expert",
        "performance-profiler",
        "agency-tiktok-strategist",
        "gws-recipe-audit-sharing",
        "agency-reddit-community-builder",
        "agency-instagram-curator",
        "anthropic-web-artifacts-builder",
        "i18n-sync",
        "agency-infrastructure-maintainer",
        "e2e-testing"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: usability analysis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"ux-expert\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-ux-architect\", \"kwp-design-accessibility-review\", \"frontend-expert\", \"design-architect\", \"compliance-governance\", \"kwp-design-design-system-management\", \"kwp-design-user-research\", \"agency-brand-guardian\", \"ci-quality-gate\", \"ui-suite\", \"agency-xr-interface-architect\", \"agency-security-engineer\", \"agency-accessibility-auditor\", \"kwp-design-design-critique\", \"backend-expert\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0002",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: accessibility audit. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "ux-expert"
      ],
      "distractor_skills": [
        "agency-ux-researcher",
        "kwp-design-accessibility-review",
        "frontend-expert",
        "design-architect",
        "compliance-governance",
        "kwp-design-design-system-management",
        "kwp-design-user-research",
        "agency-brand-guardian",
        "ci-quality-gate",
        "ui-suite",
        "agency-xr-interface-architect",
        "agency-security-engineer",
        "agency-accessibility-auditor",
        "kwp-design-design-critique",
        "backend-expert",
        "agency-xr-immersive-developer",
        "agency-frontend-developer",
        "anthropic-frontend-design",
        "role-ux-designer",
        "kwp-engineering-code-review",
        "agency-visionos-spatial-engineer",
        "agency-app-store-optimizer",
        "security-expert"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: accessibility audit. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"ux-expert\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-ux-researcher\", \"kwp-design-accessibility-review\", \"frontend-expert\", \"design-architect\", \"compliance-governance\", \"kwp-design-design-system-management\", \"kwp-design-user-research\", \"agency-brand-guardian\", \"ci-quality-gate\", \"ui-suite\", \"agency-xr-interface-architect\", \"agency-security-engineer\", \"agency-accessibility-auditor\", \"kwp-design-design-critique\", \"backend-expert\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0003",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: a UX review. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "ux-expert"
      ],
      "distractor_skills": [
        "agency-ui-designer",
        "kwp-design-accessibility-review",
        "frontend-expert",
        "design-architect",
        "compliance-governance",
        "kwp-design-design-system-management",
        "kwp-design-user-research",
        "agency-brand-guardian",
        "ci-quality-gate",
        "ui-suite",
        "agency-xr-interface-architect",
        "agency-security-engineer",
        "agency-accessibility-auditor",
        "kwp-design-design-critique",
        "backend-expert",
        "agency-ux-researcher",
        "agency-xr-immersive-developer",
        "agency-frontend-developer",
        "anthropic-frontend-design",
        "role-ux-designer",
        "kwp-engineering-code-review",
        "agency-visionos-spatial-engineer",
        "agency-app-store-optimizer",
        "security-expert",
        "performance-profiler",
        "agency-tiktok-strategist",
        "gws-recipe-audit-sharing",
        "agency-reddit-community-builder",
        "agency-instagram-curator",
        "anthropic-web-artifacts-builder",
        "i18n-sync",
        "agency-infrastructure-maintainer",
        "e2e-testing",
        "agency-xr-cockpit-interaction-specialist",
        "kwp-engineering-tech-debt",
        "agency-tool-evaluator",
        "agency-macos-spatial-metal-engineer",
        "agency-lsp-index-engineer",
        "agency-legal-compliance-checker"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: a UX review. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"ux-expert\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-ui-designer\", \"kwp-design-accessibility-review\", \"frontend-expert\", \"design-architect\", \"compliance-governance\", \"kwp-design-design-system-management\", \"kwp-design-user-research\", \"agency-brand-guardian\", \"ci-quality-gate\", \"ui-suite\", \"agency-xr-interface-architect\", \"agency-security-engineer\", \"agency-accessibility-auditor\", \"kwp-design-design-critique\", \"backend-expert\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0004",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: component refactoring. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "frontend-expert"
      ],
      "distractor_skills": [
        "agency-frontend-developer",
        "qa-test-expert",
        "ux-expert",
        "e2e-testing",
        "performance-profiler",
        "kwp-design-design-system-management",
        "backend-expert",
        "kwp-engineering-testing-strategy",
        "agency-security-engineer",
        "db-expert",
        "anthropic-webapp-testing",
        "ci-quality-gate",
        "agency-test-results-analyzer",
        "ui-suite",
        "skill-optimizer",
        "kwp-design-accessibility-review",
        "design-architect",
        "compliance-governance",
        "agency-accessibility-auditor",
        "anthropic-web-artifacts-builder",
        "i18n-sync",
        "kwp-engineering-tech-debt",
        "agency-tool-evaluator",
        "anthropic-frontend-design",
        "agency-api-tester",
        "kwp-engineering-code-review",
        "agency-performance-benchmarker",
        "agency-app-store-optimizer",
        "agency-tiktok-strategist",
        "agency-instagram-curator"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: component refactoring. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"frontend-expert\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-frontend-developer\", \"qa-test-expert\", \"ux-expert\", \"e2e-testing\", \"performance-profiler\", \"kwp-design-design-system-management\", \"backend-expert\", \"kwp-engineering-testing-strategy\", \"agency-security-engineer\", \"db-expert\", \"anthropic-webapp-testing\", \"ci-quality-gate\", \"agency-test-results-analyzer\", \"ui-suite\", \"skill-optimizer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0005",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: references agency-backend-architect. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-backend-architect"
      ],
      "distractor_skills": [
        "backend-expert",
        "agency-security-engineer",
        "agency-devops-automator",
        "agency-brand-guardian",
        "agency-performance-benchmarker",
        "agency-infrastructure-maintainer",
        "agency-xr-immersive-developer",
        "agency-frontend-developer",
        "agency-xr-interface-architect",
        "agency-ai-engineer",
        "agency-ui-designer",
        "agency-terminal-integration-specialist",
        "agency-ux-researcher",
        "agency-ux-architect",
        "agency-rapid-prototyper",
        "agency-visionos-spatial-engineer",
        "agency-data-engineer",
        "agency-app-store-optimizer",
        "agency-api-tester",
        "agency-growth-hacker",
        "agency-xr-cockpit-interaction-specialist",
        "agency-wechat-official-account-manager",
        "agency-experiment-tracker",
        "agency-technical-writer",
        "agency-mobile-app-builder",
        "agency-instagram-curator",
        "agency-accessibility-auditor",
        "agency-tool-evaluator",
        "agency-senior-project-manager",
        "agency-lsp-index-engineer",
        "agency-senior-developer",
        "agency-finance-tracker",
        "agency-autonomous-optimization-architect",
        "agency-test-results-analyzer",
        "agency-inclusive-visuals-specialist",
        "agency-social-media-strategist",
        "agency-workflow-optimizer",
        "agency-behavioral-nudge-engine",
        "agency-agents-orchestrator",
        "agency-trend-researcher",
        "agency-report-distribution-agent",
        "agency-tiktok-strategist",
        "agency-data-consolidation-agent"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: references agency-backend-architect. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-backend-architect\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"backend-expert\", \"agency-security-engineer\", \"agency-devops-automator\", \"agency-brand-guardian\", \"agency-performance-benchmarker\", \"agency-infrastructure-maintainer\", \"agency-xr-immersive-developer\", \"agency-frontend-developer\", \"agency-xr-interface-architect\", \"agency-ai-engineer\", \"agency-ui-designer\", \"agency-terminal-integration-specialist\", \"agency-ux-researcher\", \"agency-ux-architect\", \"agency-rapid-prototyper\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0006",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Security Engineer agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-security-engineer"
      ],
      "distractor_skills": [
        "security-expert",
        "agency-devops-automator",
        "agency-backend-architect",
        "agency-frontend-developer",
        "agency-visionos-spatial-engineer",
        "agency-xr-immersive-developer",
        "agency-terminal-integration-specialist",
        "agency-brand-guardian",
        "agency-mobile-app-builder",
        "agency-xr-interface-architect",
        "agency-app-store-optimizer",
        "agency-tool-evaluator",
        "agency-macos-spatial-metal-engineer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-ai-engineer",
        "agency-ux-researcher",
        "agency-ux-architect",
        "agency-trend-researcher",
        "agency-experiment-tracker",
        "agency-performance-benchmarker",
        "agency-lsp-index-engineer",
        "agency-data-engineer",
        "agency-test-results-analyzer",
        "agency-inclusive-visuals-specialist",
        "agency-api-tester",
        "agency-infrastructure-maintainer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Security Engineer agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-security-engineer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"security-expert\", \"agency-devops-automator\", \"agency-backend-architect\", \"agency-frontend-developer\", \"agency-visionos-spatial-engineer\", \"agency-xr-immersive-developer\", \"agency-terminal-integration-specialist\", \"agency-brand-guardian\", \"agency-mobile-app-builder\", \"agency-xr-interface-architect\", \"agency-app-store-optimizer\", \"agency-tool-evaluator\", \"agency-macos-spatial-metal-engineer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-ai-engineer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0007",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: comment on a discussion. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "hf-discussions"
      ],
      "distractor_skills": [
        "hf-models",
        "hf-collections",
        "hf-repos",
        "hf-buckets",
        "hf-hub",
        "hf-endpoints",
        "hf-spaces",
        "hf-cli",
        "hf-papers",
        "hf-datasets",
        "hf-dataset-viewer",
        "gws-drive",
        "gws-docs",
        "gws-chat",
        "gws-sheets",
        "hf-auto-finetune-planner",
        "hf-model-trainer",
        "domain-commit",
        "hf-model-arena",
        "hf-trending-intelligence",
        "hf-trackio",
        "gws-calendar",
        "hf-evaluation",
        "gws-gmail",
        "hf-jobs",
        "ship",
        "gws-workspace",
        "kwp-operations-vendor-management",
        "gws-workflows",
        "agency-devops-automator",
        "pr-review-captain",
        "paperclip-tasks",
        "gws-recipe-events-from-sheet",
        "gws-recipe-doc-from-template"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: comment on a discussion. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"hf-discussions\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"hf-models\", \"hf-collections\", \"hf-repos\", \"hf-buckets\", \"hf-hub\", \"hf-endpoints\", \"hf-spaces\", \"hf-cli\", \"hf-papers\", \"hf-datasets\", \"hf-dataset-viewer\", \"gws-drive\", \"gws-docs\", \"gws-chat\", \"gws-sheets\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0008",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: decision documentation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-product-management-stakeholder-comms"
      ],
      "distractor_skills": [
        "kwp-customer-support-escalation",
        "kwp-design-design-handoff",
        "kwp-marketing-campaign-planning",
        "kwp-customer-support-response-drafting",
        "kwp-product-management-metrics-tracking",
        "kwp-product-management-feature-spec",
        "kwp-finance-close-management",
        "kwp-slack-slack-search",
        "kwp-design-system",
        "kwp-product-management-competitive-analysis",
        "kwp-sales-create-an-asset",
        "kwp-human-resources-people-analytics",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-data-data-validation"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: decision documentation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-product-management-stakeholder-comms\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-customer-support-escalation\", \"kwp-design-design-handoff\", \"kwp-marketing-campaign-planning\", \"kwp-customer-support-response-drafting\", \"kwp-product-management-metrics-tracking\", \"kwp-product-management-feature-spec\", \"kwp-finance-close-management\", \"kwp-slack-slack-search\", \"kwp-design-system\", \"kwp-product-management-competitive-analysis\", \"kwp-sales-create-an-asset\", \"kwp-human-resources-people-analytics\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-data-data-validation\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0009",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: bespin-news-digest. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "decision-router"
      ],
      "distractor_skills": [
        "daily-am-orchestrator",
        "daily-strategy-post",
        "setup-doctor",
        "calendar-daily-briefing",
        "cursor-automations",
        "auto-research-distribute",
        "eod-ship",
        "related-papers-scout",
        "feedback-meeting-scheduler",
        "full-stack-planner",
        "automation-strategist",
        "daily-stock-check",
        "sprint-orchestrator",
        "prd-research-factory",
        "executive-briefing",
        "twitter-timeline-to-slack",
        "feature-intake-router",
        "auto-research",
        "critical-review",
        "bespin-news-digest",
        "ci-quality-gate",
        "simplify",
        "context-engineer",
        "commit-to-issue",
        "ppt-template-engine",
        "onboarding-accelerator",
        "unified-intel-intake",
        "cursor-sync",
        "alphaxiv-paper-lookup",
        "plans",
        "technical-writer",
        "helm-validator",
        "pipeline-builder",
        "domain-commit"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: bespin-news-digest. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"decision-router\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"daily-am-orchestrator\", \"daily-strategy-post\", \"setup-doctor\", \"calendar-daily-briefing\", \"cursor-automations\", \"auto-research-distribute\", \"eod-ship\", \"related-papers-scout\", \"feedback-meeting-scheduler\", \"full-stack-planner\", \"automation-strategist\", \"daily-stock-check\", \"sprint-orchestrator\", \"prd-research-factory\", \"executive-briefing\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0010",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Knowledge Synthesis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-enterprise-search-knowledge-synthesis"
      ],
      "distractor_skills": [
        "kwp-enterprise-search-source-management",
        "kwp-enterprise-search-search-strategy",
        "kwp-slack-slack-search",
        "kwp-human-resources-interview-prep",
        "kwp-design-design-handoff",
        "kwp-human-resources-people-analytics",
        "kwp-data-data-context-extractor",
        "kwp-finance-variance-analysis",
        "kwp-operations-vendor-management",
        "kwp-finance-close-management",
        "kwp-apollo-enrich-lead",
        "kwp-customer-support-customer-research",
        "kwp-sales-create-an-asset",
        "kwp-sales-account-research",
        "kwp-product-management-competitive-analysis",
        "kwp-marketing-performance-analytics",
        "kwp-design-system",
        "kwp-apollo-prospect",
        "kwp-common-room-account-research",
        "kwp-sales-draft-outreach",
        "kwp-sales-competitive-intelligence",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-org-planning",
        "kwp-data-statistical-analysis",
        "kwp-data-data-validation",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-finance-financial-statements",
        "kwp-product-management-feature-spec",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-finance-audit-support",
        "kwp-data-interactive-dashboard-builder",
        "kwp-customer-support-escalation",
        "kwp-legal-meeting-briefing"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Knowledge Synthesis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-enterprise-search-knowledge-synthesis\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-enterprise-search-source-management\", \"kwp-enterprise-search-search-strategy\", \"kwp-slack-slack-search\", \"kwp-human-resources-interview-prep\", \"kwp-design-design-handoff\", \"kwp-human-resources-people-analytics\", \"kwp-data-data-context-extractor\", \"kwp-finance-variance-analysis\", \"kwp-operations-vendor-management\", \"kwp-finance-close-management\", \"kwp-apollo-enrich-lead\", \"kwp-customer-support-customer-research\", \"kwp-sales-create-an-asset\", \"kwp-sales-account-research\", \"kwp-product-management-competitive-analysis\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0011",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: schedule a meeting. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "smart-meeting-scheduler"
      ],
      "distractor_skills": [
        "gws-calendar",
        "calendar-daily-briefing",
        "proactive-meeting-scheduler",
        "gws-recipe-events-from-sheet",
        "daily-am-orchestrator",
        "gws-email-reply",
        "feedback-meeting-scheduler",
        "email-auto-reply",
        "gws-workflows",
        "gws-recipe-doc-from-template",
        "gws-workspace",
        "morning-ship",
        "ai-chief-of-staff",
        "gws-gmail",
        "cursor-automations",
        "gws-recipe-personalized-emails",
        "bespin-news-digest",
        "gws-recipe-save-email-to-doc",
        "gws-chat",
        "morning",
        "gws-docs",
        "hf-papers",
        "gws-sheets",
        "standup-digest",
        "kwp-sales-daily-briefing",
        "gws-recipe-audit-sharing",
        "google-daily",
        "related-papers-scout",
        "md-enhance-publish",
        "knowledge-daily-aggregator",
        "autoskill-extractor",
        "kwp-sales-call-prep"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: schedule a meeting. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"smart-meeting-scheduler\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"gws-calendar\", \"calendar-daily-briefing\", \"proactive-meeting-scheduler\", \"gws-recipe-events-from-sheet\", \"daily-am-orchestrator\", \"gws-email-reply\", \"feedback-meeting-scheduler\", \"email-auto-reply\", \"gws-workflows\", \"gws-recipe-doc-from-template\", \"gws-workspace\", \"morning-ship\", \"ai-chief-of-staff\", \"gws-gmail\", \"cursor-automations\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0012",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: skill-aware routing. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "sefo-orchestrator"
      ],
      "distractor_skills": [
        "sefo-federation",
        "sefo-governance",
        "mission-control copy",
        "paperclip-setup",
        "mission-control",
        "ai-workflow-integrator",
        "plans",
        "autoskill-merger",
        "sp-core",
        "ecc-agentic-engineering",
        "paperclip-agents",
        "full-stack-planner",
        "agency-roster",
        "workflow-sequential",
        "agency-studio-producer",
        "public-apis",
        "agency-technical-writer",
        "issue-to-code",
        "prompt-transformer",
        "ecc-continuous-learning",
        "ai-chief-of-staff",
        "paperclip-control",
        "agency-infrastructure-maintainer",
        "scrapling",
        "pm-sentinel",
        "agency-data-consolidation-agent",
        "paperclip-tasks",
        "iac-review-agent",
        "autoskill-judge",
        "paper-review",
        "skill-optimizer",
        "air-gap-orchestrator",
        "ecc-autonomous-loops",
        "kwp-operations-change-management",
        "gws-workspace",
        "gws-workflows",
        "system-thinker",
        "kwp-productivity-task-management",
        "agency-visual-storyteller",
        "ralph-loop"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: skill-aware routing. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"sefo-orchestrator\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"sefo-federation\", \"sefo-governance\", \"mission-control copy\", \"paperclip-setup\", \"mission-control\", \"ai-workflow-integrator\", \"plans\", \"autoskill-merger\", \"sp-core\", \"ecc-agentic-engineering\", \"paperclip-agents\", \"full-stack-planner\", \"agency-roster\", \"workflow-sequential\", \"agency-studio-producer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0013",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Ai Engineer agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-ai-engineer"
      ],
      "distractor_skills": [
        "agency-data-engineer",
        "agency-xr-immersive-developer",
        "agency-brand-guardian",
        "agency-terminal-integration-specialist",
        "agency-data-consolidation-agent",
        "agency-app-store-optimizer",
        "agency-tiktok-strategist",
        "agency-macos-spatial-metal-engineer",
        "agency-lsp-index-engineer",
        "agency-instagram-curator",
        "agency-xr-cockpit-interaction-specialist",
        "agency-api-tester",
        "agency-infrastructure-maintainer",
        "agency-xr-interface-architect",
        "agency-ux-architect",
        "agency-mobile-app-builder",
        "agency-growth-hacker",
        "agency-visionos-spatial-engineer",
        "agency-tool-evaluator",
        "agency-experiment-tracker",
        "agency-reddit-community-builder",
        "agency-test-results-analyzer",
        "agency-studio-operations",
        "agency-developer-advocate",
        "agency-workflow-optimizer",
        "agency-legal-compliance-checker",
        "agency-rapid-prototyper",
        "agency-devops-automator",
        "agency-senior-developer",
        "agency-trend-researcher",
        "agency-senior-project-manager",
        "agency-feedback-synthesizer",
        "agency-security-engineer",
        "agency-sales-data-extraction-agent",
        "agency-inclusive-visuals-specialist",
        "agency-finance-tracker",
        "agency-report-distribution-agent",
        "agency-twitter-engager",
        "agency-behavioral-nudge-engine",
        "agency-agents-orchestrator",
        "agency-visual-storyteller",
        "agency-xiaohongshu-specialist"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Ai Engineer agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-ai-engineer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-data-engineer\", \"agency-xr-immersive-developer\", \"agency-brand-guardian\", \"agency-terminal-integration-specialist\", \"agency-data-consolidation-agent\", \"agency-app-store-optimizer\", \"agency-tiktok-strategist\", \"agency-macos-spatial-metal-engineer\", \"agency-lsp-index-engineer\", \"agency-instagram-curator\", \"agency-xr-cockpit-interaction-specialist\", \"agency-api-tester\", \"agency-infrastructure-maintainer\", \"agency-xr-interface-architect\", \"agency-ux-architect\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0014",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: filter_cells from qc_plotting import plot_qc_distributions # Only if visualization needed adata = ad.read_h5ad('input.h5ad') calculate_qc_metrics(adata. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-bio-research-single-cell-rna-qc"
      ],
      "distractor_skills": [
        "kwp-bio-research-scvi-tools",
        "kwp-data-data-exploration",
        "kwp-product-management-competitive-analysis",
        "kwp-human-resources-people-analytics",
        "kwp-data-data-visualization",
        "kwp-data-data-validation",
        "kwp-human-resources-interview-prep",
        "kwp-data-interactive-dashboard-builder",
        "kwp-slack-slack-search",
        "kwp-legal-legal-risk-assessment",
        "kwp-common-room-account-research",
        "kwp-data-data-context-extractor",
        "kwp-data-statistical-analysis",
        "kwp-finance-variance-analysis",
        "kwp-finance-financial-statements",
        "kwp-operations-vendor-management",
        "kwp-finance-audit-support",
        "kwp-customer-support-escalation",
        "kwp-finance-close-management",
        "kwp-design-design-handoff"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: filter_cells from qc_plotting import plot_qc_distributions # Only if visualization needed adata = ad.read_h5ad('input.h5ad') calculate_qc_metrics(adata. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-bio-research-single-cell-rna-qc\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-bio-research-scvi-tools\", \"kwp-data-data-exploration\", \"kwp-product-management-competitive-analysis\", \"kwp-human-resources-people-analytics\", \"kwp-data-data-visualization\", \"kwp-data-data-validation\", \"kwp-human-resources-interview-prep\", \"kwp-data-interactive-dashboard-builder\", \"kwp-slack-slack-search\", \"kwp-legal-legal-risk-assessment\", \"kwp-common-room-account-research\", \"kwp-data-data-context-extractor\", \"kwp-data-statistical-analysis\", \"kwp-finance-variance-analysis\", \"kwp-finance-financial-statements\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0015",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: agency-roster. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-roster"
      ],
      "distractor_skills": [
        "agency-app-store-optimizer",
        "agency-workflow-optimizer",
        "agency-developer-advocate",
        "agency-ux-architect",
        "agency-mobile-app-builder",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-terminal-integration-specialist",
        "agency-xr-interface-architect",
        "agency-behavioral-nudge-engine",
        "agency-rapid-prototyper",
        "agency-data-consolidation-agent",
        "agency-infrastructure-maintainer",
        "agency-tool-evaluator",
        "agency-senior-project-manager",
        "agency-macos-spatial-metal-engineer",
        "agency-lsp-index-engineer",
        "agency-legal-compliance-checker",
        "agency-brand-guardian",
        "agency-test-results-analyzer",
        "agency-sales-data-extraction-agent",
        "agency-inclusive-visuals-specialist",
        "agency-api-tester",
        "agency-visual-storyteller",
        "agency-visionos-spatial-engineer",
        "agency-senior-developer",
        "agency-performance-benchmarker",
        "agency-cultural-intelligence-strategist",
        "agency-evidence-collector",
        "agency-tiktok-strategist",
        "agency-studio-operations",
        "agency-devops-automator",
        "agency-agents-orchestrator",
        "agency-reddit-community-builder",
        "agency-instagram-curator",
        "agency-accessibility-auditor",
        "agency-xiaohongshu-specialist",
        "agency-report-distribution-agent",
        "agency-frontend-developer",
        "agency-whimsy-injector",
        "agency-support-responder",
        "agency-agentic-identity-trust-architect"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: agency-roster. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-roster\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-app-store-optimizer\", \"agency-workflow-optimizer\", \"agency-developer-advocate\", \"agency-ux-architect\", \"agency-mobile-app-builder\", \"agency-xr-immersive-developer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-terminal-integration-specialist\", \"agency-xr-interface-architect\", \"agency-behavioral-nudge-engine\", \"agency-rapid-prototyper\", \"agency-data-consolidation-agent\", \"agency-infrastructure-maintainer\", \"agency-tool-evaluator\", \"agency-senior-project-manager\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0016",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: adapt a learning framework. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "autoskill-pipeline"
      ],
      "distractor_skills": [
        "autoskill-evolve",
        "autoskill-extractor",
        "autoskill-judge",
        "autoskill-merger",
        "release-ship",
        "deep-research-pipeline",
        "daily-strategy-post",
        "hf-trending-intelligence",
        "auto-research",
        "auto-research-distribute",
        "agency-agents-orchestrator",
        "test-suite",
        "release-commander",
        "iac-review-agent",
        "email-research-dispatcher",
        "deep-review",
        "agency-visionos-spatial-engineer",
        "agency-senior-developer",
        "role-cso"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: adapt a learning framework. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"autoskill-pipeline\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"autoskill-evolve\", \"autoskill-extractor\", \"autoskill-judge\", \"autoskill-merger\", \"release-ship\", \"deep-research-pipeline\", \"daily-strategy-post\", \"hf-trending-intelligence\", \"auto-research\", \"auto-research-distribute\", \"agency-agents-orchestrator\", \"test-suite\", \"release-commander\", \"iac-review-agent\", \"email-research-dispatcher\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0017",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: AlphaEar Sentiment. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "alphaear-sentiment"
      ],
      "distractor_skills": [
        "alphaear-deepear-lite",
        "alphaear-news",
        "alphaear-search",
        "alphaear-predictor",
        "trading-market-environment-analysis",
        "alphaear-signal-tracker",
        "alphaear-reporter",
        "alphaear-stock",
        "trading-breadth-chart-analyst",
        "trading-market-news-analyst",
        "hf-model-trainer",
        "alphaear-logic-visualizer",
        "daily-stock-check",
        "stock-csv-downloader",
        "role-trading-expert",
        "trading-market-top-detector",
        "trading-market-breadth-analyzer",
        "mirofish-financial-sim",
        "trading-uptrend-analyzer",
        "trading-theme-detector",
        "trading-us-stock-analysis",
        "hf-evaluation",
        "hf-cli",
        "trading-trade-hypothesis-ideator",
        "tab-news-fetch",
        "mirofish",
        "trading-sector-analyst",
        "weekly-stock-update",
        "tab-report-generate",
        "trading-stanley-druckenmiller-investment",
        "trading-technical-analyst",
        "tab-kiwoom",
        "qa-test-expert"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: AlphaEar Sentiment. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-sentiment\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"alphaear-deepear-lite\", \"alphaear-news\", \"alphaear-search\", \"alphaear-predictor\", \"trading-market-environment-analysis\", \"alphaear-signal-tracker\", \"alphaear-reporter\", \"alphaear-stock\", \"trading-breadth-chart-analyst\", \"trading-market-news-analyst\", \"hf-model-trainer\", \"alphaear-logic-visualizer\", \"daily-stock-check\", \"stock-csv-downloader\", \"role-trading-expert\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0018",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Notion 문서 동기화. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "notion-docs-sync"
      ],
      "distractor_skills": [
        "github-workflow-automation",
        "swagger-api-doc-generator",
        "implement-screen",
        "screen-description",
        "design-review",
        "figma-to-tds",
        "fsd-development",
        "notion-meeting-sync",
        "overlay-layout-patterns",
        "fe-pipeline",
        "local-dev-setup",
        "sp-parallel-agents",
        "tab-turtle-refresh",
        "tab-fundamental-sync",
        "tab-dualma-refresh",
        "tab-bollinger-refresh",
        "tab-llm-agents",
        "tab-hot-stock-discovery",
        "google-daily",
        "sp-writing-plans",
        "tab-screening",
        "tab-market-breadth",
        "kwp-sync",
        "sp-subagent-dev"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Notion 문서 동기화. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"notion-docs-sync\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"github-workflow-automation\", \"swagger-api-doc-generator\", \"implement-screen\", \"screen-description\", \"design-review\", \"figma-to-tds\", \"fsd-development\", \"notion-meeting-sync\", \"overlay-layout-patterns\", \"fe-pipeline\", \"local-dev-setup\", \"sp-parallel-agents\", \"tab-turtle-refresh\", \"tab-fundamental-sync\", \"tab-dualma-refresh\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0019",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: evaluate skill. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "skill-optimizer"
      ],
      "distractor_skills": [
        "autoskill-judge",
        "skill-guide-generator",
        "skill-seekers",
        "kwp-sync",
        "skill-composer",
        "workflow-miner",
        "anthropic-skill-creator",
        "autoskill-extractor",
        "autoskill-merger",
        "agency-test-results-analyzer",
        "ui-suite",
        "intent-alignment-tracker",
        "agency-terminal-integration-specialist",
        "agency-api-tester",
        "agency-app-store-optimizer",
        "prompt-transformer",
        "agency-xr-interface-architect",
        "test-suite",
        "frontend-expert",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "i18n-sync",
        "agency-tool-evaluator",
        "agency-brand-guardian",
        "sp-writing-skills",
        "skill-autoimprove",
        "evals-skills",
        "agency-sales-data-extraction-agent",
        "agency-roster",
        "agency-inclusive-visuals-specialist",
        "kwp-engineering-code-review"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: evaluate skill. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"skill-optimizer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"autoskill-judge\", \"skill-guide-generator\", \"skill-seekers\", \"kwp-sync\", \"skill-composer\", \"workflow-miner\", \"anthropic-skill-creator\", \"autoskill-extractor\", \"autoskill-merger\", \"agency-test-results-analyzer\", \"ui-suite\", \"intent-alignment-tracker\", \"agency-terminal-integration-specialist\", \"agency-api-tester\", \"agency-app-store-optimizer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0020",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: preparing documentation for reproducibility. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-data-data-validation"
      ],
      "distractor_skills": [
        "kwp-finance-audit-support",
        "kwp-product-management-competitive-analysis",
        "kwp-human-resources-people-analytics",
        "kwp-data-statistical-analysis",
        "kwp-human-resources-interview-prep",
        "kwp-finance-variance-analysis",
        "kwp-finance-financial-statements",
        "kwp-data-data-context-extractor",
        "kwp-data-data-exploration",
        "kwp-marketing-performance-analytics",
        "kwp-marketing-brand-voice",
        "kwp-operations-vendor-management",
        "kwp-slack-slack-search",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-data-interactive-dashboard-builder",
        "kwp-finance-close-management",
        "kwp-design-design-handoff",
        "kwp-design-system",
        "kwp-legal-meeting-briefing",
        "kwp-legal-compliance",
        "kwp-finance-reconciliation",
        "kwp-common-room-account-research",
        "kwp-finance-journal-entry-prep",
        "kwp-marketing-competitive-analysis",
        "kwp-legal-contract-review",
        "kwp-data-data-visualization",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-product-management-feature-spec",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-legal-canned-responses",
        "kwp-customer-support-escalation",
        "kwp-apollo-enrich-lead",
        "kwp-customer-support-customer-research",
        "kwp-cowork-plugin-customizer",
        "kwp-apollo-prospect",
        "kwp-sales-daily-briefing"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: preparing documentation for reproducibility. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-data-data-validation\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-finance-audit-support\", \"kwp-product-management-competitive-analysis\", \"kwp-human-resources-people-analytics\", \"kwp-data-statistical-analysis\", \"kwp-human-resources-interview-prep\", \"kwp-finance-variance-analysis\", \"kwp-finance-financial-statements\", \"kwp-data-data-context-extractor\", \"kwp-data-data-exploration\", \"kwp-marketing-performance-analytics\", \"kwp-marketing-brand-voice\", \"kwp-operations-vendor-management\", \"kwp-slack-slack-search\", \"kwp-bio-research-single-cell-rna-qc\", \"kwp-data-interactive-dashboard-builder\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0021",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 스킬 병합. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "autoskill-merger"
      ],
      "distractor_skills": [
        "autoskill-judge",
        "autoskill-evolve",
        "autoskill-extractor",
        "sp-writing-skills",
        "anthropic-skill-creator",
        "agency-roster",
        "kwp-sync",
        "skill-optimizer",
        "skill-guide-generator",
        "autoskill-pipeline",
        "prompt-transformer",
        "skill-composer",
        "ship",
        "commit-to-issue",
        "intent-alignment-tracker",
        "agency-data-consolidation-agent",
        "skill-seekers",
        "hf-repos",
        "kwp-engineering-tech-debt",
        "kwp-human-resources-recruiting-pipeline",
        "gws-recipe-doc-from-template",
        "agency-xr-immersive-developer",
        "ecc-configure",
        "technical-writer",
        "kwp-engineering-code-review",
        "agency-app-store-optimizer",
        "kwp-human-resources-interview-prep",
        "workflow-miner",
        "kwp-design-accessibility-review",
        "gws-sheets",
        "gws-recipe-save-email-to-doc",
        "sefo-orchestrator",
        "mission-control copy",
        "kwp-operations-risk-assessment",
        "anthropic-claude-api",
        "paperclip-agents",
        "hf-collections",
        "docs-tutor-setup",
        "agency-xr-interface-architect",
        "stock-csv-downloader",
        "kwp-design-design-system-management",
        "kwp-design-design-critique",
        "kwp-data-data-visualization",
        "agency-ux-architect",
        "agency-mobile-app-builder",
        "kwp-design-system",
        "agency-visual-storyteller"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 스킬 병합. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"autoskill-merger\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"autoskill-judge\", \"autoskill-evolve\", \"autoskill-extractor\", \"sp-writing-skills\", \"anthropic-skill-creator\", \"agency-roster\", \"kwp-sync\", \"skill-optimizer\", \"skill-guide-generator\", \"autoskill-pipeline\", \"prompt-transformer\", \"skill-composer\", \"ship\", \"commit-to-issue\", \"intent-alignment-tracker\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0022",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: create launch kit. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "nlm-launch-kit"
      ],
      "distractor_skills": [
        "nlm-strategy-hub",
        "nlm-discovery-lab",
        "nlm-deep-learn",
        "nlm-slides",
        "nlm-arxiv-slides",
        "nlm-video",
        "notebooklm-studio",
        "pm-go-to-market",
        "prd-research-factory",
        "role-cso",
        "paper-review",
        "pm-marketing-growth",
        "notion-meeting-sync",
        "daily-strategy-post",
        "notebooklm",
        "pm-market-research",
        "role-ceo",
        "mirofish",
        "critical-review",
        "pm-product-strategy",
        "kwp-brand-voice-guideline-generation",
        "kwp-product-management-competitive-analysis",
        "notebooklm-research",
        "deep-research-pipeline",
        "pm-execution",
        "full-stack-planner",
        "docs-tutor-setup",
        "alphaxiv-paper-lookup",
        "knowledge-to-paper",
        "hf-buckets",
        "role-sales",
        "kwp-marketing-competitive-analysis",
        "onboarding-accelerator",
        "trading-breadth-chart-analyst",
        "gws-recipe-doc-from-template",
        "e2e-testing"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: create launch kit. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"nlm-launch-kit\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"nlm-strategy-hub\", \"nlm-discovery-lab\", \"nlm-deep-learn\", \"nlm-slides\", \"nlm-arxiv-slides\", \"nlm-video\", \"notebooklm-studio\", \"pm-go-to-market\", \"prd-research-factory\", \"role-cso\", \"paper-review\", \"pm-marketing-growth\", \"notion-meeting-sync\", \"daily-strategy-post\", \"notebooklm\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0023",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: run a daily stock check. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "daily-stock-check"
      ],
      "distractor_skills": [
        "stock-csv-downloader",
        "weekly-stock-update",
        "alphaear-stock",
        "today",
        "trading-sector-analyst",
        "tab-stock-sync",
        "tab-turtle-refresh",
        "daily-strategy-post",
        "tab-strategy-comparison",
        "trading-technical-analyst",
        "commit-to-issue",
        "alphaear-deepear-lite"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: run a daily stock check. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"daily-stock-check\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"stock-csv-downloader\", \"weekly-stock-update\", \"alphaear-stock\", \"today\", \"trading-sector-analyst\", \"tab-stock-sync\", \"tab-turtle-refresh\", \"daily-strategy-post\", \"tab-strategy-comparison\", \"trading-technical-analyst\", \"commit-to-issue\", \"alphaear-deepear-lite\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0024",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Resource Planning. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-operations-resource-planning"
      ],
      "distractor_skills": [
        "kwp-human-resources-org-planning",
        "kwp-operations-change-management",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-marketing-campaign-planning",
        "kwp-engineering-testing-strategy",
        "kwp-operations-process-optimization",
        "kwp-human-resources-people-analytics",
        "kwp-engineering-system-design",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-slack-slack-search",
        "kwp-finance-close-management",
        "kwp-design-design-handoff",
        "kwp-design-ux-writing",
        "kwp-bio-research-scientific-problem-selection",
        "kwp-sales-daily-briefing",
        "kwp-sales-account-research",
        "kwp-marketing-performance-analytics",
        "kwp-design-user-research",
        "kwp-sales-competitive-intelligence",
        "kwp-finance-variance-analysis",
        "kwp-sales-call-prep",
        "kwp-product-management-feature-spec",
        "kwp-operations-compliance-tracking",
        "kwp-data-interactive-dashboard-builder",
        "kwp-customer-support-escalation",
        "kwp-design-system",
        "kwp-product-management-competitive-analysis",
        "kwp-human-resources-employee-handbook",
        "kwp-legal-legal-risk-assessment",
        "kwp-common-room-account-research",
        "kwp-marketing-brand-voice",
        "kwp-sales-draft-outreach",
        "kwp-product-management-metrics-tracking",
        "kwp-data-data-validation",
        "kwp-engineering-tech-debt",
        "kwp-engineering-documentation",
        "kwp-finance-financial-statements",
        "kwp-common-room-call-prep",
        "kwp-finance-audit-support",
        "kwp-productivity-task-management",
        "kwp-legal-meeting-briefing",
        "kwp-customer-support-knowledge-management",
        "kwp-customer-support-customer-research",
        "kwp-apollo-enrich-lead",
        "kwp-product-management-user-research-synthesis"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Resource Planning. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-operations-resource-planning\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-human-resources-org-planning\", \"kwp-operations-change-management\", \"kwp-operations-vendor-management\", \"kwp-human-resources-interview-prep\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-marketing-campaign-planning\", \"kwp-engineering-testing-strategy\", \"kwp-operations-process-optimization\", \"kwp-human-resources-people-analytics\", \"kwp-engineering-system-design\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-slack-slack-search\", \"kwp-finance-close-management\", \"kwp-design-design-handoff\", \"kwp-design-ux-writing\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0025",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Zhihu Strategist agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-zhihu-strategist"
      ],
      "distractor_skills": [
        "agency-twitter-engager",
        "agency-xiaohongshu-specialist",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "agency-instagram-curator",
        "agency-app-store-optimizer",
        "agency-wechat-official-account-manager",
        "agency-brand-guardian",
        "agency-social-media-strategist",
        "agency-xr-interface-architect",
        "agency-tool-evaluator",
        "agency-whimsy-injector",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-test-results-analyzer",
        "agency-api-tester",
        "agency-terminal-integration-specialist",
        "agency-workflow-optimizer",
        "agency-growth-hacker",
        "agency-content-creator",
        "agency-visual-storyteller",
        "agency-senior-developer",
        "agency-lsp-index-engineer",
        "agency-experiment-tracker",
        "agency-inclusive-visuals-specialist",
        "agency-infrastructure-maintainer",
        "agency-behavioral-nudge-engine",
        "agency-developer-advocate",
        "agency-cultural-intelligence-strategist",
        "agency-ux-architect",
        "agency-mobile-app-builder",
        "agency-visionos-spatial-engineer",
        "agency-trend-researcher",
        "agency-support-responder",
        "agency-senior-project-manager",
        "agency-macos-spatial-metal-engineer",
        "agency-evidence-collector",
        "agency-report-distribution-agent",
        "agency-legal-compliance-checker",
        "agency-rapid-prototyper",
        "agency-data-consolidation-agent",
        "agency-feedback-synthesizer",
        "agency-sales-data-extraction-agent",
        "agency-finance-tracker",
        "agency-studio-operations",
        "agency-ai-engineer",
        "agency-agents-orchestrator",
        "agency-project-shepherd",
        "agency-image-prompt-engineer",
        "agency-reality-checker"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Zhihu Strategist agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-zhihu-strategist\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-twitter-engager\", \"agency-xiaohongshu-specialist\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"agency-instagram-curator\", \"agency-app-store-optimizer\", \"agency-wechat-official-account-manager\", \"agency-brand-guardian\", \"agency-social-media-strategist\", \"agency-xr-interface-architect\", \"agency-tool-evaluator\", \"agency-whimsy-injector\", \"agency-xr-immersive-developer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-test-results-analyzer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0026",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: screening stocks. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "tab-screening"
      ],
      "distractor_skills": [
        "tab-hot-stock-discovery",
        "tab-report-generate",
        "tab-genai-features",
        "tab-dualma-refresh",
        "tab-analysis-run",
        "tab-strategy-comparison",
        "tab-llm-agents",
        "tab-turtle-refresh",
        "tab-bollinger-refresh",
        "tab-market-environment",
        "tab-dashboard-summary",
        "tab-fundamental-sync",
        "tab-event-detect",
        "tab-stock-sync",
        "tab-technical-analysis",
        "tab-glossary",
        "tab-market-breadth",
        "tab-patterns",
        "tab-kiwoom",
        "today",
        "tab-news-fetch",
        "daily-stock-check",
        "trading-theme-detector",
        "trading-technical-analyst",
        "trading-breadth-chart-analyst",
        "trading-sector-analyst",
        "alphaear-predictor",
        "morning-ship",
        "trading-us-stock-analysis",
        "trading-stanley-druckenmiller-investment",
        "alphaear-reporter",
        "hf-cli",
        "alphaear-deepear-lite",
        "mirofish-financial-sim",
        "trading-market-environment-analysis",
        "stock-csv-downloader",
        "hf-jobs",
        "trading-finviz-screener",
        "trading-market-breadth-analyzer",
        "morning",
        "alphaear-stock",
        "trading-backtest-expert",
        "autoskill-pipeline",
        "weekly-stock-update",
        "trading-scenario-analyzer",
        "daily-strategy-post"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: screening stocks. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"tab-screening\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"tab-hot-stock-discovery\", \"tab-report-generate\", \"tab-genai-features\", \"tab-dualma-refresh\", \"tab-analysis-run\", \"tab-strategy-comparison\", \"tab-llm-agents\", \"tab-turtle-refresh\", \"tab-bollinger-refresh\", \"tab-market-environment\", \"tab-dashboard-summary\", \"tab-fundamental-sync\", \"tab-event-detect\", \"tab-stock-sync\", \"tab-technical-analysis\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0027",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: references agency-instagram-curator. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-instagram-curator"
      ],
      "distractor_skills": [
        "agency-xiaohongshu-specialist",
        "agency-tiktok-strategist",
        "agency-reddit-community-builder",
        "agency-app-store-optimizer",
        "agency-wechat-official-account-manager",
        "agency-zhihu-strategist",
        "agency-twitter-engager",
        "agency-tool-evaluator",
        "agency-terminal-integration-specialist",
        "agency-brand-guardian",
        "agency-content-creator",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-visual-storyteller",
        "agency-test-results-analyzer",
        "agency-api-tester",
        "agency-xr-interface-architect",
        "agency-infrastructure-maintainer",
        "agency-workflow-optimizer",
        "agency-senior-developer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: references agency-instagram-curator. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-instagram-curator\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-xiaohongshu-specialist\", \"agency-tiktok-strategist\", \"agency-reddit-community-builder\", \"agency-app-store-optimizer\", \"agency-wechat-official-account-manager\", \"agency-zhihu-strategist\", \"agency-twitter-engager\", \"agency-tool-evaluator\", \"agency-terminal-integration-specialist\", \"agency-brand-guardian\", \"agency-content-creator\", \"agency-xr-immersive-developer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-visual-storyteller\", \"agency-test-results-analyzer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0028",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: market conditions. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "trading-market-environment-analysis"
      ],
      "distractor_skills": [
        "trading-breadth-chart-analyst",
        "trading-stanley-druckenmiller-investment",
        "trading-uptrend-analyzer",
        "trading-market-news-analyst",
        "trading-us-stock-analysis",
        "alphaear-sentiment",
        "trading-sector-analyst",
        "trading-market-top-detector",
        "trading-theme-detector",
        "alphaear-deepear-lite",
        "trading-market-breadth-analyzer",
        "trading-technical-analyst",
        "trading-trade-hypothesis-ideator",
        "trading-position-sizer",
        "stock-csv-downloader",
        "trading-finviz-screener",
        "trading-scenario-analyzer",
        "trading-edge-candidate-agent",
        "trading-options-strategy-advisor",
        "trading-backtest-expert",
        "trading-options-theta"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: market conditions. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"trading-market-environment-analysis\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"trading-breadth-chart-analyst\", \"trading-stanley-druckenmiller-investment\", \"trading-uptrend-analyzer\", \"trading-market-news-analyst\", \"trading-us-stock-analysis\", \"alphaear-sentiment\", \"trading-sector-analyst\", \"trading-market-top-detector\", \"trading-theme-detector\", \"alphaear-deepear-lite\", \"trading-market-breadth-analyzer\", \"trading-technical-analyst\", \"trading-trade-hypothesis-ideator\", \"trading-position-sizer\", \"stock-csv-downloader\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0029",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 인텔 파이프라인. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "unified-intel-intake"
      ],
      "distractor_skills": [
        "twitter-timeline-to-slack",
        "paper-auto-classifier",
        "bespin-news-digest",
        "iac-review-agent",
        "x-to-slack",
        "auto-research-distribute",
        "daily-am-orchestrator",
        "daily-strategy-post",
        "auto-research",
        "knowledge-to-paper",
        "hf-trending-intelligence",
        "autoskill-pipeline",
        "decision-router",
        "release-commander",
        "release-ship",
        "eod-ship",
        "sprint-orchestrator",
        "morning-ship",
        "role-dispatcher",
        "weekly-status-report",
        "agency-agents-orchestrator",
        "portfolio-report-generator",
        "agency-lsp-index-engineer",
        "deep-research-pipeline",
        "github-sprint-digest",
        "ship",
        "agency-workflow-optimizer",
        "kwp-engineering-code-review",
        "alphaxiv-paper-lookup",
        "standup-digest",
        "kwp-design-accessibility-review",
        "related-papers-scout",
        "agency-content-creator",
        "knowledge-daily-aggregator",
        "paper-review"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 인텔 파이프라인. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"unified-intel-intake\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"twitter-timeline-to-slack\", \"paper-auto-classifier\", \"bespin-news-digest\", \"iac-review-agent\", \"x-to-slack\", \"auto-research-distribute\", \"daily-am-orchestrator\", \"daily-strategy-post\", \"auto-research\", \"knowledge-to-paper\", \"hf-trending-intelligence\", \"autoskill-pipeline\", \"decision-router\", \"release-commander\", \"release-ship\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0030",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: refine this. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "workflow-eval-opt"
      ],
      "distractor_skills": [
        "workflow-sequential",
        "workflow-parallel",
        "workflow-miner",
        "skill-optimizer",
        "kwp-operations-process-optimization",
        "ralph-loop",
        "kwp-design-accessibility-review",
        "kwp-design-user-research",
        "sentence-polisher",
        "agency-test-results-analyzer",
        "role-data-scientist",
        "sefo-orchestrator",
        "issue-to-code",
        "prompt-transformer",
        "paperclip-control",
        "qa-test-expert",
        "gws-calendar",
        "kwp-operations-vendor-management",
        "intent-alignment-tracker",
        "ecc-eval-harness",
        "email-auto-reply",
        "paperclip-tasks",
        "skill-autoimprove",
        "autoskill-judge",
        "hf-evaluation",
        "role-developer",
        "e2e-testing",
        "evals-skills",
        "sefo-federation",
        "transcribee",
        "paperclip-setup",
        "compliance-gate",
        "paperclip-agents",
        "kwp-engineering-code-review",
        "gws-workspace",
        "autoskill-merger",
        "trading-trade-hypothesis-ideator",
        "kwp-productivity-task-management",
        "kwp-design-ux-writing",
        "trading-finviz-screener",
        "ai-quality-evaluator",
        "kwp-operations-resource-planning",
        "kwp-design-design-critique",
        "office-template-enforcer",
        "mission-control",
        "kwp-brand-voice-brand-voice-enforcement"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: refine this. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"workflow-eval-opt\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"workflow-sequential\", \"workflow-parallel\", \"workflow-miner\", \"skill-optimizer\", \"kwp-operations-process-optimization\", \"ralph-loop\", \"kwp-design-accessibility-review\", \"kwp-design-user-research\", \"sentence-polisher\", \"agency-test-results-analyzer\", \"role-data-scientist\", \"sefo-orchestrator\", \"issue-to-code\", \"prompt-transformer\", \"paperclip-control\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0031",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: interacting with Kiwoom. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "tab-kiwoom"
      ],
      "distractor_skills": [
        "tab-report-generate",
        "tab-stock-sync",
        "tab-llm-agents",
        "tab-dashboard-summary",
        "tab-fundamental-sync",
        "tab-strategy-comparison",
        "tab-glossary",
        "tab-market-breadth",
        "tab-turtle-refresh",
        "tab-patterns",
        "tab-dualma-refresh",
        "tab-analysis-run",
        "tab-screening",
        "tab-news-fetch",
        "tab-genai-features",
        "tab-market-environment",
        "stock-csv-downloader",
        "tab-bollinger-refresh",
        "tab-event-detect",
        "tab-hot-stock-discovery",
        "weekly-stock-update",
        "alphaear-predictor",
        "tab-technical-analysis",
        "alphaear-reporter",
        "trading-market-environment-analysis",
        "hf-jobs",
        "alphaear-news",
        "daily-stock-check"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: interacting with Kiwoom. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"tab-kiwoom\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"tab-report-generate\", \"tab-stock-sync\", \"tab-llm-agents\", \"tab-dashboard-summary\", \"tab-fundamental-sync\", \"tab-strategy-comparison\", \"tab-glossary\", \"tab-market-breadth\", \"tab-turtle-refresh\", \"tab-patterns\", \"tab-dualma-refresh\", \"tab-analysis-run\", \"tab-screening\", \"tab-news-fetch\", \"tab-genai-features\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0032",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: wants a detailed analysis of the latest Bespin Global news clipping. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "bespin-news-digest"
      ],
      "distractor_skills": [
        "email-research-dispatcher",
        "daily-am-orchestrator",
        "twitter-timeline-to-slack",
        "morning-ship",
        "github-sprint-digest",
        "morning",
        "portfolio-report-generator",
        "calendar-daily-briefing",
        "md-enhance-publish",
        "daily-strategy-post",
        "unified-intel-intake",
        "knowledge-to-paper",
        "gws-recipe-save-email-to-doc",
        "gws-workflows",
        "ai-chief-of-staff",
        "gws-drive",
        "tab-news-fetch",
        "eod-ship",
        "tab-report-generate",
        "gmail-daily-triage",
        "gws-recipe-personalized-emails",
        "hf-trending-intelligence",
        "gws-chat",
        "autoskill-pipeline",
        "weekly-status-report",
        "standup-digest",
        "gws-email-reply",
        "smart-meeting-scheduler",
        "critical-review",
        "proactive-meeting-scheduler",
        "paper-auto-classifier",
        "notion-meeting-sync",
        "decision-router",
        "gws-recipe-doc-from-template",
        "deep-research-pipeline",
        "x-to-slack",
        "gws-gmail",
        "gws-workspace"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: wants a detailed analysis of the latest Bespin Global news clipping. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"bespin-news-digest\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"email-research-dispatcher\", \"daily-am-orchestrator\", \"twitter-timeline-to-slack\", \"morning-ship\", \"github-sprint-digest\", \"morning\", \"portfolio-report-generator\", \"calendar-daily-briefing\", \"md-enhance-publish\", \"daily-strategy-post\", \"unified-intel-intake\", \"knowledge-to-paper\", \"gws-recipe-save-email-to-doc\", \"gws-workflows\", \"ai-chief-of-staff\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0033",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: detecting events. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "tab-event-detect"
      ],
      "distractor_skills": [
        "tab-analysis-run",
        "tab-patterns",
        "tab-fundamental-sync",
        "tab-genai-features",
        "tab-screening",
        "tab-turtle-refresh",
        "tab-strategy-comparison",
        "tab-dualma-refresh",
        "tab-technical-analysis",
        "tab-llm-agents",
        "tab-hot-stock-discovery",
        "tab-market-breadth",
        "tab-dashboard-summary",
        "tab-news-fetch",
        "tab-report-generate",
        "tab-stock-sync",
        "tab-bollinger-refresh",
        "tab-glossary",
        "tab-market-environment",
        "tab-kiwoom",
        "gws-recipe-events-from-sheet",
        "kwp-human-resources-people-analytics",
        "kwp-data-statistical-analysis",
        "skill-seekers",
        "hf-dataset-viewer",
        "autoskill-pipeline",
        "daily-stock-check",
        "anthropic-skill-creator",
        "agency-rapid-prototyper",
        "email-research-dispatcher",
        "kwp-data-data-context-extractor",
        "fe-pipeline",
        "kwp-design-design-handoff",
        "morning-ship",
        "swagger-api-doc-generator",
        "kwp-bio-research-nextflow-development",
        "sefo-orchestrator",
        "x-to-slack",
        "kwp-data-data-validation",
        "gws-recipe-doc-from-template"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: detecting events. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"tab-event-detect\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"tab-analysis-run\", \"tab-patterns\", \"tab-fundamental-sync\", \"tab-genai-features\", \"tab-screening\", \"tab-turtle-refresh\", \"tab-strategy-comparison\", \"tab-dualma-refresh\", \"tab-technical-analysis\", \"tab-llm-agents\", \"tab-hot-stock-discovery\", \"tab-market-breadth\", \"tab-dashboard-summary\", \"tab-news-fetch\", \"tab-report-generate\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0034",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: validate all checks pass. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "ci-quality-gate"
      ],
      "distractor_skills": [
        "k8s-manifest-validator",
        "helm-validator",
        "setup-doctor",
        "security-expert",
        "code-review-all",
        "dependency-auditor",
        "compliance-gate",
        "sre-devops-expert",
        "performance-profiler",
        "qa-test-expert",
        "domain-commit",
        "e2e-testing",
        "ecc-verification-loop",
        "ux-expert",
        "technical-writer",
        "commit-to-issue",
        "daily-am-orchestrator",
        "simplify",
        "daily-stock-check",
        "frontend-expert",
        "calendar-daily-briefing",
        "semantic-guard",
        "terraform-reviewer",
        "alphaxiv-paper-lookup",
        "auto-research-distribute",
        "deep-review",
        "daily-strategy-post",
        "e2e-overhaul",
        "eod-ship",
        "refactor-simulator",
        "backend-expert",
        "ui-suite",
        "problem-definition",
        "local-dev-runner",
        "automation-strategist",
        "pipeline-builder"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: validate all checks pass. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"ci-quality-gate\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"k8s-manifest-validator\", \"helm-validator\", \"setup-doctor\", \"security-expert\", \"code-review-all\", \"dependency-auditor\", \"compliance-gate\", \"sre-devops-expert\", \"performance-profiler\", \"qa-test-expert\", \"domain-commit\", \"e2e-testing\", \"ecc-verification-loop\", \"ux-expert\", \"technical-writer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0035",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 에어갭 모드. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "air-gap-orchestrator"
      ],
      "distractor_skills": [
        "paperclip-setup",
        "compliance-governance",
        "hf-endpoints",
        "test-suite",
        "release-commander",
        "workflow-sequential",
        "plans",
        "auto-research",
        "ai-workflow-integrator",
        "sefo-orchestrator",
        "paperclip-agents",
        "dependency-auditor",
        "sefo-federation",
        "jira-linear-adapter",
        "daily-am-orchestrator",
        "intent-alignment-tracker",
        "agency-ux-researcher",
        "knowledge-to-paper",
        "iac-review-agent",
        "email-research-dispatcher",
        "agency-devops-automator",
        "ui-suite",
        "hf-trending-intelligence",
        "semantic-guard",
        "rtk",
        "full-stack-planner",
        "security-expert",
        "agency-backend-architect",
        "agency-xr-cockpit-interaction-specialist",
        "role-data-scientist",
        "critical-review",
        "twitter-timeline-to-slack",
        "ralph-loop",
        "trading-market-news-analyst",
        "kwp-operations-compliance-tracking",
        "i18n-sync",
        "feedback-meeting-scheduler",
        "evals-skills",
        "compliance-gate",
        "agency-security-engineer",
        "hf-auto-finetune-planner",
        "hf-spaces",
        "skill-composer",
        "qa-test-expert",
        "autoskill-pipeline",
        "auto-research-distribute",
        "agency-agents-orchestrator"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 에어갭 모드. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"air-gap-orchestrator\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"paperclip-setup\", \"compliance-governance\", \"hf-endpoints\", \"test-suite\", \"release-commander\", \"workflow-sequential\", \"plans\", \"auto-research\", \"ai-workflow-integrator\", \"sefo-orchestrator\", \"paperclip-agents\", \"dependency-auditor\", \"sefo-federation\", \"jira-linear-adapter\", \"daily-am-orchestrator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0036",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 회의록 분석. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "meeting-digest"
      ],
      "distractor_skills": [
        "notion-meeting-sync",
        "paper-review",
        "deep-research-pipeline",
        "md-to-notion",
        "alphaxiv-paper-lookup",
        "github-sprint-digest",
        "nlm-arxiv-slides",
        "x-to-slack",
        "defuddle",
        "weekly-status-report",
        "pr-review-captain",
        "md-enhance-publish",
        "kwp-brand-voice-guideline-generation",
        "daily-stock-check",
        "role-ux-designer",
        "commit-to-issue",
        "bespin-news-digest",
        "transcribee",
        "role-ceo",
        "kwp-legal-meeting-briefing",
        "role-data-scientist",
        "twitter-timeline-to-slack",
        "role-dispatcher",
        "role-finance",
        "role-hr",
        "role-developer",
        "executive-briefing",
        "stock-csv-downloader",
        "pm-data-analytics",
        "role-trading-expert",
        "role-pm",
        "role-cso",
        "hf-buckets",
        "email-research-dispatcher",
        "kwp-human-resources-people-analytics",
        "md-to-slack-canvas",
        "gws-recipe-doc-from-template",
        "role-sales",
        "nlm-slides",
        "role-security-engineer",
        "gws-docs",
        "mission-control copy"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 회의록 분석. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"meeting-digest\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"notion-meeting-sync\", \"paper-review\", \"deep-research-pipeline\", \"md-to-notion\", \"alphaxiv-paper-lookup\", \"github-sprint-digest\", \"nlm-arxiv-slides\", \"x-to-slack\", \"defuddle\", \"weekly-status-report\", \"pr-review-captain\", \"md-enhance-publish\", \"kwp-brand-voice-guideline-generation\", \"daily-stock-check\", \"role-ux-designer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0037",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: company design standards apply. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "anthropic-brand-guidelines"
      ],
      "distractor_skills": [
        "anthropic-theme-factory",
        "anthropic-slack-gif-creator",
        "anthropic-canvas-design",
        "anthropic-claude-api",
        "anthropic-algorithmic-art",
        "anthropic-mcp-builder",
        "anthropic-web-artifacts-builder",
        "anthropic-webapp-testing",
        "anthropic-skill-creator",
        "kwp-marketing-brand-voice",
        "anthropic-frontend-design",
        "anthropic-internal-comms",
        "anthropic-docx",
        "anthropic-doc-coauthoring",
        "anthropic-pdf",
        "anthropic-pptx",
        "kwp-design-system",
        "anthropic-xlsx",
        "agency-ui-designer",
        "pm-marketing-growth",
        "kwp-brand-voice-brand-voice-enforcement",
        "kwp-brand-voice-guideline-generation",
        "sp-tdd",
        "sp-git-worktrees",
        "alphaear-logic-visualizer",
        "sp-debugging",
        "agency-accessibility-auditor",
        "kwp-sync",
        "kwp-data-data-context-extractor",
        "agency-visual-storyteller",
        "sp-core"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: company design standards apply. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-brand-guidelines\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"anthropic-theme-factory\", \"anthropic-slack-gif-creator\", \"anthropic-canvas-design\", \"anthropic-claude-api\", \"anthropic-algorithmic-art\", \"anthropic-mcp-builder\", \"anthropic-web-artifacts-builder\", \"anthropic-webapp-testing\", \"anthropic-skill-creator\", \"kwp-marketing-brand-voice\", \"anthropic-frontend-design\", \"anthropic-internal-comms\", \"anthropic-docx\", \"anthropic-doc-coauthoring\", \"anthropic-pdf\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0038",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Call Prep. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-sales-call-prep"
      ],
      "distractor_skills": [
        "kwp-sales-account-research",
        "kwp-sales-daily-briefing",
        "kwp-sales-draft-outreach",
        "kwp-legal-meeting-briefing",
        "kwp-sales-create-an-asset",
        "kwp-common-room-call-prep",
        "kwp-sales-competitive-intelligence",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-apollo-enrich-lead",
        "kwp-customer-support-customer-research",
        "kwp-common-room-account-research",
        "kwp-human-resources-people-analytics",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-org-planning",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-apollo-prospect",
        "kwp-operations-resource-planning",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-product-management-feature-spec",
        "kwp-customer-support-escalation",
        "kwp-finance-close-management",
        "kwp-design-design-handoff",
        "kwp-slack-slack-search",
        "kwp-design-system",
        "kwp-product-management-competitive-analysis",
        "kwp-product-management-user-research-synthesis",
        "kwp-marketing-performance-analytics",
        "kwp-finance-journal-entry-prep",
        "kwp-marketing-competitive-analysis",
        "kwp-legal-contract-review",
        "kwp-common-room-contact-research",
        "kwp-product-management-metrics-tracking",
        "kwp-data-data-validation",
        "kwp-data-data-context-extractor",
        "kwp-finance-variance-analysis",
        "kwp-finance-financial-statements",
        "kwp-brand-voice-guideline-generation",
        "kwp-operations-compliance-tracking",
        "kwp-finance-audit-support",
        "kwp-data-interactive-dashboard-builder",
        "kwp-marketing-content-creation",
        "kwp-enterprise-search-source-management",
        "kwp-marketing-campaign-planning",
        "kwp-legal-legal-risk-assessment",
        "kwp-human-resources-employee-handbook"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Call Prep. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-sales-call-prep\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-sales-account-research\", \"kwp-sales-daily-briefing\", \"kwp-sales-draft-outreach\", \"kwp-legal-meeting-briefing\", \"kwp-sales-create-an-asset\", \"kwp-common-room-call-prep\", \"kwp-sales-competitive-intelligence\", \"kwp-operations-vendor-management\", \"kwp-human-resources-interview-prep\", \"kwp-apollo-enrich-lead\", \"kwp-customer-support-customer-research\", \"kwp-common-room-account-research\", \"kwp-human-resources-people-analytics\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-human-resources-org-planning\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0039",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: run compliance check. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "compliance-gate"
      ],
      "distractor_skills": [
        "compliance-governance",
        "ci-quality-gate",
        "release-commander",
        "iac-review-agent",
        "ecc-verification-loop",
        "semantic-guard",
        "k8s-manifest-validator",
        "security-expert",
        "helm-validator",
        "terraform-reviewer",
        "dependency-auditor",
        "intent-alignment-tracker",
        "code-review-all",
        "infra-drift-detector",
        "deep-review",
        "ui-suite",
        "agency-data-consolidation-agent",
        "test-suite",
        "workflow-sequential",
        "auto-research",
        "trading-data-quality-checker",
        "ship",
        "kwp-engineering-code-review",
        "commit-to-issue",
        "ai-quality-evaluator",
        "portfolio-report-generator",
        "diagnose",
        "ux-expert",
        "release-ship",
        "agency-test-results-analyzer",
        "agency-api-tester",
        "agency-xr-immersive-developer",
        "agency-report-distribution-agent",
        "setup-doctor"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: run compliance check. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"compliance-gate\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"compliance-governance\", \"ci-quality-gate\", \"release-commander\", \"iac-review-agent\", \"ecc-verification-loop\", \"semantic-guard\", \"k8s-manifest-validator\", \"security-expert\", \"helm-validator\", \"terraform-reviewer\", \"dependency-auditor\", \"intent-alignment-tracker\", \"code-review-all\", \"infra-drift-detector\", \"deep-review\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0040",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Xr Interface Architect agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-xr-interface-architect"
      ],
      "distractor_skills": [
        "agency-xr-cockpit-interaction-specialist",
        "agency-xr-immersive-developer",
        "agency-brand-guardian",
        "agency-visionos-spatial-engineer",
        "agency-terminal-integration-specialist",
        "agency-behavioral-nudge-engine",
        "agency-app-store-optimizer",
        "agency-ux-architect",
        "agency-report-distribution-agent",
        "agency-macos-spatial-metal-engineer",
        "agency-senior-developer",
        "agency-rapid-prototyper",
        "agency-mobile-app-builder",
        "agency-data-consolidation-agent",
        "agency-cultural-intelligence-strategist",
        "agency-tiktok-strategist"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Xr Interface Architect agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-xr-interface-architect\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-xr-cockpit-interaction-specialist\", \"agency-xr-immersive-developer\", \"agency-brand-guardian\", \"agency-visionos-spatial-engineer\", \"agency-terminal-integration-specialist\", \"agency-behavioral-nudge-engine\", \"agency-app-store-optimizer\", \"agency-ux-architect\", \"agency-report-distribution-agent\", \"agency-macos-spatial-metal-engineer\", \"agency-senior-developer\", \"agency-rapid-prototyper\", \"agency-mobile-app-builder\", \"agency-data-consolidation-agent\", \"agency-cultural-intelligence-strategist\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0041",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 휴먼 인 더 루프. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "automation-strategist"
      ],
      "distractor_skills": [
        "pipeline-builder",
        "system-thinker",
        "technical-writer",
        "sre-devops-expert",
        "plans",
        "setup-doctor",
        "daily-strategy-post",
        "codebase-archaeologist",
        "ci-quality-gate",
        "daily-am-orchestrator",
        "auto-research-distribute",
        "context-engineer",
        "refactor-simulator",
        "onboarding-accelerator",
        "sprint-orchestrator",
        "eod-ship",
        "critical-review",
        "qa-test-expert",
        "problem-definition",
        "calendar-daily-briefing",
        "feedback-meeting-scheduler",
        "daily-stock-check",
        "performance-profiler",
        "auto-research",
        "executive-briefing",
        "decision-router",
        "security-expert",
        "cursor-automations",
        "commit-to-issue",
        "full-stack-planner",
        "k8s-manifest-validator",
        "alphaxiv-paper-lookup",
        "kwp-engineering-testing-strategy",
        "helm-validator",
        "kwp-human-resources-org-planning",
        "domain-commit",
        "code-review-all",
        "ai-workflow-integrator",
        "jira-linear-adapter",
        "trading-stanley-druckenmiller-investment",
        "prd-research-factory",
        "kwp-operations-risk-assessment",
        "kwp-engineering-system-design"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 휴먼 인 더 루프. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"automation-strategist\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"pipeline-builder\", \"system-thinker\", \"technical-writer\", \"sre-devops-expert\", \"plans\", \"setup-doctor\", \"daily-strategy-post\", \"codebase-archaeologist\", \"ci-quality-gate\", \"daily-am-orchestrator\", \"auto-research-distribute\", \"context-engineer\", \"refactor-simulator\", \"onboarding-accelerator\", \"sprint-orchestrator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0042",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: slow queries. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "performance-profiler"
      ],
      "distractor_skills": [
        "backend-expert",
        "sre-devops-expert",
        "db-expert",
        "k8s-manifest-validator",
        "ci-quality-gate",
        "technical-writer",
        "frontend-expert",
        "simplify",
        "commit-to-issue",
        "code-review-all",
        "kwp-engineering-code-review",
        "agency-backend-architect",
        "agency-devops-automator",
        "security-expert",
        "helm-validator",
        "pipeline-builder",
        "compliance-governance",
        "agency-frontend-developer",
        "problem-definition",
        "ux-expert",
        "daily-stock-check",
        "agency-performance-benchmarker",
        "kwp-design-design-system-management",
        "agency-infrastructure-maintainer",
        "domain-commit",
        "kwp-engineering-tech-debt",
        "onboarding-accelerator",
        "codebase-archaeologist",
        "agency-api-tester",
        "service-health-doctor",
        "agency-app-store-optimizer",
        "skill-optimizer",
        "calendar-daily-briefing",
        "kwp-design-accessibility-review",
        "agency-security-engineer",
        "qa-test-expert"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: slow queries. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"performance-profiler\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"backend-expert\", \"sre-devops-expert\", \"db-expert\", \"k8s-manifest-validator\", \"ci-quality-gate\", \"technical-writer\", \"frontend-expert\", \"simplify\", \"commit-to-issue\", \"code-review-all\", \"kwp-engineering-code-review\", \"agency-backend-architect\", \"agency-devops-automator\", \"security-expert\", \"helm-validator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0043",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Slack Messaging Best Practices. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-slack-slack-messaging"
      ],
      "distractor_skills": [
        "kwp-slack-slack-search",
        "kwp-common-room-compose-outreach",
        "kwp-common-room-contact-research",
        "kwp-common-room-account-research",
        "kwp-data-data-context-extractor",
        "kwp-common-room-call-prep",
        "kwp-design-design-handoff",
        "kwp-design-system",
        "kwp-apollo-sequence-load",
        "kwp-common-room-prospect",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-enterprise-search-search-strategy",
        "kwp-finance-close-management",
        "kwp-enterprise-search-source-management",
        "kwp-apollo-enrich-lead"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Slack Messaging Best Practices. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-slack-slack-messaging\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-slack-slack-search\", \"kwp-common-room-compose-outreach\", \"kwp-common-room-contact-research\", \"kwp-common-room-account-research\", \"kwp-data-data-context-extractor\", \"kwp-common-room-call-prep\", \"kwp-design-design-handoff\", \"kwp-design-system\", \"kwp-apollo-sequence-load\", \"kwp-common-room-prospect\", \"kwp-operations-vendor-management\", \"kwp-human-resources-interview-prep\", \"kwp-enterprise-search-search-strategy\", \"kwp-finance-close-management\", \"kwp-enterprise-search-source-management\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0044",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: a new ticket. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-customer-support-ticket-triage"
      ],
      "distractor_skills": [
        "kwp-legal-nda-triage",
        "kwp-customer-support-knowledge-management",
        "kwp-customer-support-escalation",
        "kwp-data-data-exploration",
        "kwp-legal-legal-risk-assessment",
        "kwp-customer-support-response-drafting",
        "kwp-human-resources-org-planning",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-legal-compliance",
        "kwp-enterprise-search-source-management",
        "kwp-product-management-user-research-synthesis",
        "kwp-product-management-competitive-analysis",
        "kwp-finance-reconciliation",
        "kwp-design-system",
        "kwp-slack-slack-search",
        "kwp-common-room-call-prep",
        "kwp-product-management-feature-spec",
        "kwp-finance-audit-support",
        "kwp-customer-support-customer-research",
        "kwp-finance-close-management",
        "kwp-slack-slack-messaging",
        "kwp-cowork-plugin-customizer",
        "kwp-apollo-prospect",
        "kwp-operations-resource-planning",
        "kwp-legal-contract-review",
        "kwp-marketing-competitive-analysis",
        "kwp-marketing-brand-voice",
        "kwp-human-resources-people-analytics",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-sales-competitive-intelligence",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-data-statistical-analysis",
        "kwp-data-data-validation",
        "kwp-apollo-sequence-load",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-finance-variance-analysis",
        "kwp-finance-financial-statements",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-legal-canned-responses",
        "kwp-data-interactive-dashboard-builder",
        "kwp-data-data-context-extractor",
        "kwp-legal-meeting-briefing",
        "kwp-sales-daily-briefing",
        "kwp-sales-account-research",
        "kwp-product-management-roadmap-management",
        "kwp-marketing-performance-analytics"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: a new ticket. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-customer-support-ticket-triage\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-legal-nda-triage\", \"kwp-customer-support-knowledge-management\", \"kwp-customer-support-escalation\", \"kwp-data-data-exploration\", \"kwp-legal-legal-risk-assessment\", \"kwp-customer-support-response-drafting\", \"kwp-human-resources-org-planning\", \"kwp-operations-vendor-management\", \"kwp-human-resources-interview-prep\", \"kwp-legal-compliance\", \"kwp-enterprise-search-source-management\", \"kwp-product-management-user-research-synthesis\", \"kwp-product-management-competitive-analysis\", \"kwp-finance-reconciliation\", \"kwp-design-system\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0045",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Kiwoom API. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "tab-kiwoom"
      ],
      "distractor_skills": [
        "tab-report-generate",
        "tab-stock-sync",
        "tab-llm-agents",
        "tab-dashboard-summary",
        "tab-fundamental-sync",
        "tab-strategy-comparison",
        "tab-glossary",
        "tab-market-breadth",
        "tab-turtle-refresh",
        "tab-patterns",
        "tab-dualma-refresh",
        "tab-analysis-run",
        "tab-screening",
        "tab-news-fetch",
        "tab-genai-features",
        "tab-market-environment",
        "stock-csv-downloader",
        "tab-bollinger-refresh",
        "tab-event-detect",
        "tab-hot-stock-discovery",
        "weekly-stock-update",
        "alphaear-predictor",
        "tab-technical-analysis",
        "alphaear-reporter",
        "trading-market-environment-analysis",
        "hf-jobs",
        "alphaear-news",
        "daily-stock-check",
        "alphaear-sentiment",
        "hf-evaluation",
        "gws-recipe-audit-sharing",
        "alphaear-signal-tracker",
        "trading-trade-hypothesis-ideator",
        "hf-trackio",
        "mirofish-financial-sim",
        "kwp-human-resources-people-analytics",
        "alphaear-search",
        "trading-sector-analyst",
        "trading-market-breadth-analyzer",
        "trading-breadth-chart-analyst",
        "kwp-data-data-validation",
        "role-trading-expert",
        "trading-theme-detector",
        "kwp-finance-variance-analysis"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Kiwoom API. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"tab-kiwoom\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"tab-report-generate\", \"tab-stock-sync\", \"tab-llm-agents\", \"tab-dashboard-summary\", \"tab-fundamental-sync\", \"tab-strategy-comparison\", \"tab-glossary\", \"tab-market-breadth\", \"tab-turtle-refresh\", \"tab-patterns\", \"tab-dualma-refresh\", \"tab-analysis-run\", \"tab-screening\", \"tab-news-fetch\", \"tab-genai-features\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0046",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: CSO perspective. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "role-cso"
      ],
      "distractor_skills": [
        "role-ceo",
        "role-pm",
        "role-sales",
        "role-ux-designer",
        "role-data-scientist",
        "role-hr",
        "role-trading-expert",
        "role-developer",
        "role-finance",
        "role-security-engineer",
        "role-cto",
        "deep-research-pipeline",
        "role-dispatcher",
        "pm-market-research",
        "pm-product-strategy",
        "pm-go-to-market",
        "pm-marketing-growth"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: CSO perspective. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"role-cso\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"role-ceo\", \"role-pm\", \"role-sales\", \"role-ux-designer\", \"role-data-scientist\", \"role-hr\", \"role-trading-expert\", \"role-developer\", \"role-finance\", \"role-security-engineer\", \"role-cto\", \"deep-research-pipeline\", \"role-dispatcher\", \"pm-market-research\", \"pm-product-strategy\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0047",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: reviewing an analysis for errors. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-data-data-validation"
      ],
      "distractor_skills": [
        "kwp-finance-audit-support",
        "kwp-product-management-competitive-analysis",
        "kwp-human-resources-people-analytics",
        "kwp-data-statistical-analysis",
        "kwp-human-resources-interview-prep",
        "kwp-finance-variance-analysis",
        "kwp-finance-financial-statements",
        "kwp-data-data-context-extractor",
        "kwp-data-data-exploration",
        "kwp-marketing-performance-analytics",
        "kwp-marketing-brand-voice",
        "kwp-operations-vendor-management",
        "kwp-slack-slack-search",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-data-interactive-dashboard-builder",
        "kwp-finance-close-management",
        "kwp-design-design-handoff",
        "kwp-design-system"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: reviewing an analysis for errors. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-data-data-validation\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-finance-audit-support\", \"kwp-product-management-competitive-analysis\", \"kwp-human-resources-people-analytics\", \"kwp-data-statistical-analysis\", \"kwp-human-resources-interview-prep\", \"kwp-finance-variance-analysis\", \"kwp-finance-financial-statements\", \"kwp-data-data-context-extractor\", \"kwp-data-data-exploration\", \"kwp-marketing-performance-analytics\", \"kwp-marketing-brand-voice\", \"kwp-operations-vendor-management\", \"kwp-slack-slack-search\", \"kwp-bio-research-single-cell-rna-qc\", \"kwp-data-interactive-dashboard-builder\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0048",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: tab-technical-analysis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "tab-technical-analysis"
      ],
      "distractor_skills": [
        "tab-turtle-refresh",
        "tab-bollinger-refresh",
        "tab-market-breadth",
        "tab-strategy-comparison",
        "tab-dualma-refresh",
        "tab-glossary",
        "tab-genai-features",
        "tab-stock-sync",
        "tab-screening",
        "tab-dashboard-summary",
        "tab-fundamental-sync",
        "tab-news-fetch",
        "tab-report-generate",
        "tab-market-environment",
        "tab-event-detect",
        "tab-llm-agents",
        "tab-patterns",
        "tab-analysis-run",
        "tab-hot-stock-discovery",
        "tab-kiwoom",
        "daily-stock-check",
        "morning-ship",
        "weekly-stock-update",
        "hf-jobs",
        "daily-strategy-post",
        "alphaear-predictor",
        "kwp-data-statistical-analysis",
        "gws-recipe-events-from-sheet",
        "alphaear-stock",
        "i18n-sync",
        "hf-cli",
        "agency-visionos-spatial-engineer",
        "agency-performance-benchmarker",
        "hf-spaces",
        "today",
        "sp-writing-plans",
        "gws-recipe-personalized-emails"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: tab-technical-analysis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"tab-technical-analysis\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"tab-turtle-refresh\", \"tab-bollinger-refresh\", \"tab-market-breadth\", \"tab-strategy-comparison\", \"tab-dualma-refresh\", \"tab-glossary\", \"tab-genai-features\", \"tab-stock-sync\", \"tab-screening\", \"tab-dashboard-summary\", \"tab-fundamental-sync\", \"tab-news-fetch\", \"tab-report-generate\", \"tab-market-environment\", \"tab-event-detect\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0049",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: validate K8s manifests. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "k8s-manifest-validator"
      ],
      "distractor_skills": [
        "helm-validator",
        "terraform-reviewer",
        "ci-quality-gate",
        "iac-review-agent",
        "sre-devops-expert",
        "infra-drift-detector",
        "performance-profiler",
        "code-review-all",
        "pipeline-builder",
        "setup-doctor",
        "security-expert",
        "commit-to-issue",
        "compliance-gate",
        "technical-writer",
        "backend-expert",
        "auto-research-distribute",
        "refactor-simulator",
        "alphaxiv-paper-lookup",
        "simplify",
        "problem-definition",
        "infra-runbook-generator",
        "cursor-automations",
        "agency-devops-automator",
        "daily-stock-check",
        "auto-research",
        "onboarding-accelerator",
        "domain-commit",
        "semantic-guard",
        "related-papers-scout",
        "agency-backend-architect",
        "calendar-daily-briefing",
        "jira-linear-adapter",
        "daily-am-orchestrator"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: validate K8s manifests. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"k8s-manifest-validator\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"helm-validator\", \"terraform-reviewer\", \"ci-quality-gate\", \"iac-review-agent\", \"sre-devops-expert\", \"infra-drift-detector\", \"performance-profiler\", \"code-review-all\", \"pipeline-builder\", \"setup-doctor\", \"security-expert\", \"commit-to-issue\", \"compliance-gate\", \"technical-writer\", \"backend-expert\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0050",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: write PRD. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "pm-execution"
      ],
      "distractor_skills": [
        "pm-product-strategy",
        "pm-product-discovery",
        "pm-market-research",
        "pm-marketing-growth",
        "pm-go-to-market",
        "pm-data-analytics",
        "pm-toolkit",
        "role-pm",
        "agency-sprint-prioritizer",
        "e2e-testing",
        "kwp-product-management-feature-spec",
        "agency-executive-summary-generator",
        "qa-test-expert",
        "kwp-product-management-roadmap-management",
        "technical-writer",
        "pm-sentinel",
        "mission-control copy",
        "pr-review-captain",
        "kwp-engineering-testing-strategy",
        "evals-skills",
        "skill-optimizer",
        "gws-docs",
        "hf-hub",
        "nlm-launch-kit",
        "role-ceo",
        "kwp-operations-change-management",
        "mission-control",
        "gws-workflows",
        "gws-sheets",
        "hf-collections",
        "kwp-human-resources-interview-prep",
        "kwp-marketing-campaign-planning",
        "notion-meeting-sync",
        "gws-calendar",
        "role-cso",
        "kwp-product-management-competitive-analysis",
        "domain-commit",
        "plans",
        "hf-models",
        "ci-quality-gate",
        "agency-roster",
        "ship",
        "kwp-product-management-metrics-tracking"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: write PRD. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"pm-execution\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"pm-product-strategy\", \"pm-product-discovery\", \"pm-market-research\", \"pm-marketing-growth\", \"pm-go-to-market\", \"pm-data-analytics\", \"pm-toolkit\", \"role-pm\", \"agency-sprint-prioritizer\", \"e2e-testing\", \"kwp-product-management-feature-spec\", \"agency-executive-summary-generator\", \"qa-test-expert\", \"kwp-product-management-roadmap-management\", \"technical-writer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0051",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: needs unified compliance validation before deployment. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "compliance-gate"
      ],
      "distractor_skills": [
        "compliance-governance",
        "ci-quality-gate",
        "release-commander",
        "iac-review-agent",
        "ecc-verification-loop",
        "semantic-guard",
        "k8s-manifest-validator",
        "security-expert",
        "helm-validator",
        "terraform-reviewer",
        "dependency-auditor",
        "intent-alignment-tracker",
        "code-review-all",
        "infra-drift-detector",
        "deep-review",
        "ui-suite",
        "agency-data-consolidation-agent",
        "test-suite",
        "workflow-sequential",
        "auto-research",
        "trading-data-quality-checker",
        "ship",
        "kwp-engineering-code-review",
        "commit-to-issue",
        "ai-quality-evaluator",
        "portfolio-report-generator",
        "diagnose",
        "ux-expert",
        "release-ship",
        "agency-test-results-analyzer",
        "agency-api-tester",
        "agency-xr-immersive-developer",
        "agency-report-distribution-agent",
        "setup-doctor",
        "agency-ai-engineer",
        "skill-optimizer",
        "hf-trending-intelligence",
        "workflow-parallel",
        "agency-image-prompt-engineer",
        "agency-accessibility-auditor"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: needs unified compliance validation before deployment. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"compliance-gate\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"compliance-governance\", \"ci-quality-gate\", \"release-commander\", \"iac-review-agent\", \"ecc-verification-loop\", \"semantic-guard\", \"k8s-manifest-validator\", \"security-expert\", \"helm-validator\", \"terraform-reviewer\", \"dependency-auditor\", \"intent-alignment-tracker\", \"code-review-all\", \"infra-drift-detector\", \"deep-review\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0052",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 영업 분석. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "role-sales"
      ],
      "distractor_skills": [
        "role-cso",
        "role-data-scientist",
        "role-ux-designer",
        "role-hr",
        "role-developer",
        "role-finance",
        "role-ceo",
        "role-pm",
        "role-security-engineer",
        "role-cto",
        "role-trading-expert",
        "deep-research-pipeline",
        "role-dispatcher",
        "executive-briefing",
        "pm-market-research",
        "paper-review",
        "pm-go-to-market",
        "kwp-operations-compliance-tracking",
        "agency-data-consolidation-agent",
        "kwp-engineering-documentation",
        "kwp-brand-voice-guideline-generation",
        "prd-research-factory",
        "gws-recipe-doc-from-template",
        "kwp-sales-call-prep",
        "commit-to-issue",
        "full-stack-planner",
        "agency-agents-orchestrator",
        "kwp-design-user-research",
        "kwp-product-management-competitive-analysis",
        "kwp-human-resources-people-analytics",
        "kwp-sales-draft-outreach",
        "agency-xr-immersive-developer",
        "x-to-slack",
        "skill-optimizer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 영업 분석. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"role-sales\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"role-cso\", \"role-data-scientist\", \"role-ux-designer\", \"role-hr\", \"role-developer\", \"role-finance\", \"role-ceo\", \"role-pm\", \"role-security-engineer\", \"role-cto\", \"role-trading-expert\", \"deep-research-pipeline\", \"role-dispatcher\", \"executive-briefing\", \"pm-market-research\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0053",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: analyzing interview notes. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-product-management-user-research-synthesis"
      ],
      "distractor_skills": [
        "kwp-human-resources-interview-prep",
        "kwp-product-management-feature-spec",
        "kwp-legal-canned-responses",
        "kwp-marketing-competitive-analysis",
        "kwp-human-resources-people-analytics",
        "kwp-customer-support-response-drafting",
        "kwp-data-statistical-analysis",
        "kwp-finance-variance-analysis",
        "kwp-operations-vendor-management",
        "kwp-data-interactive-dashboard-builder",
        "kwp-customer-support-escalation",
        "kwp-design-design-handoff",
        "kwp-slack-slack-search",
        "kwp-legal-meeting-briefing",
        "kwp-customer-support-customer-research",
        "kwp-design-system",
        "kwp-product-management-competitive-analysis",
        "kwp-finance-reconciliation",
        "kwp-product-management-roadmap-management",
        "kwp-common-room-contact-research",
        "kwp-common-room-account-research",
        "kwp-design-user-research",
        "kwp-data-data-context-extractor",
        "kwp-sales-competitive-intelligence",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-org-planning",
        "kwp-data-data-validation",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-sales-call-prep",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-finance-audit-support",
        "kwp-brand-voice-guideline-generation",
        "kwp-legal-compliance",
        "kwp-data-data-exploration",
        "kwp-customer-support-knowledge-management",
        "kwp-finance-close-management",
        "kwp-sales-account-research",
        "kwp-marketing-performance-analytics",
        "kwp-marketing-campaign-planning",
        "kwp-customer-support-ticket-triage",
        "kwp-cowork-plugin-customizer",
        "kwp-operations-resource-planning",
        "kwp-data-data-visualization",
        "kwp-marketing-brand-voice"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: analyzing interview notes. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-product-management-user-research-synthesis\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-human-resources-interview-prep\", \"kwp-product-management-feature-spec\", \"kwp-legal-canned-responses\", \"kwp-marketing-competitive-analysis\", \"kwp-human-resources-people-analytics\", \"kwp-customer-support-response-drafting\", \"kwp-data-statistical-analysis\", \"kwp-finance-variance-analysis\", \"kwp-operations-vendor-management\", \"kwp-data-interactive-dashboard-builder\", \"kwp-customer-support-escalation\", \"kwp-design-design-handoff\", \"kwp-slack-slack-search\", \"kwp-legal-meeting-briefing\", \"kwp-customer-support-customer-research\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0054",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: knowledge-to-paper. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "knowledge-to-paper"
      ],
      "distractor_skills": [
        "knowledge-daily-aggregator",
        "auto-research",
        "auto-research-distribute",
        "daily-strategy-post",
        "hf-trending-intelligence",
        "paper-auto-classifier",
        "cognee",
        "bespin-news-digest",
        "md-enhance-publish",
        "notebooklm-research",
        "email-research-dispatcher",
        "morning",
        "trading-edge-candidate-agent",
        "morning-ship",
        "nlm-slides",
        "daily-am-orchestrator",
        "notion-meeting-sync",
        "unified-intel-intake"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: knowledge-to-paper. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"knowledge-to-paper\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"knowledge-daily-aggregator\", \"auto-research\", \"auto-research-distribute\", \"daily-strategy-post\", \"hf-trending-intelligence\", \"paper-auto-classifier\", \"cognee\", \"bespin-news-digest\", \"md-enhance-publish\", \"notebooklm-research\", \"email-research-dispatcher\", \"morning\", \"trading-edge-candidate-agent\", \"morning-ship\", \"nlm-slides\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0055",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: What should ALWAYS be filtered out of queries? (test data, fraud, internal users, etc.). Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-data-data-context-extractor"
      ],
      "distractor_skills": [
        "kwp-common-room-contact-research",
        "kwp-common-room-account-research",
        "kwp-human-resources-people-analytics",
        "kwp-slack-slack-search",
        "kwp-cowork-plugin-customizer",
        "kwp-common-room-compose-outreach",
        "kwp-data-statistical-analysis",
        "kwp-data-data-validation",
        "kwp-finance-financial-statements",
        "kwp-legal-canned-responses",
        "kwp-design-design-handoff",
        "kwp-apollo-enrich-lead",
        "kwp-design-system",
        "kwp-data-data-exploration",
        "kwp-customer-support-knowledge-management",
        "kwp-product-management-competitive-analysis",
        "kwp-finance-reconciliation",
        "kwp-common-room-weekly-prep-brief",
        "kwp-sales-create-an-asset",
        "kwp-sales-draft-outreach",
        "kwp-slack-slack-messaging",
        "kwp-common-room-call-prep",
        "kwp-cowork-create-plugin",
        "kwp-human-resources-compensation-benchmarking"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: What should ALWAYS be filtered out of queries? (test data, fraud, internal users, etc.). Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-data-data-context-extractor\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-common-room-contact-research\", \"kwp-common-room-account-research\", \"kwp-human-resources-people-analytics\", \"kwp-slack-slack-search\", \"kwp-cowork-plugin-customizer\", \"kwp-common-room-compose-outreach\", \"kwp-data-statistical-analysis\", \"kwp-data-data-validation\", \"kwp-finance-financial-statements\", \"kwp-legal-canned-responses\", \"kwp-design-design-handoff\", \"kwp-apollo-enrich-lead\", \"kwp-design-system\", \"kwp-data-data-exploration\", \"kwp-customer-support-knowledge-management\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0056",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: reviewing AI-generated code. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "ecc-agentic-engineering"
      ],
      "distractor_skills": [
        "ecc-search-first",
        "ecc-autonomous-loops",
        "ecc-coding-standards",
        "ecc-strategic-compact",
        "ecc-eval-harness",
        "ecc-configure",
        "ecc-continuous-learning",
        "ecc-verification-loop",
        "sefo-orchestrator",
        "agency-visionos-spatial-engineer",
        "agency-security-engineer",
        "sp-finish-branch",
        "kwp-finance-close-management",
        "ecc-iterative-retrieval",
        "agency-ux-architect",
        "agency-lsp-index-engineer",
        "sp-subagent-dev",
        "plans",
        "agency-xr-cockpit-interaction-specialist",
        "sp-tdd",
        "paperclip-setup",
        "kwp-productivity-task-management",
        "sp-request-review",
        "kwp-product-management-roadmap-management",
        "kwp-operations-resource-planning",
        "hf-jobs",
        "agency-rapid-prototyper",
        "sp-receive-review",
        "paperclip-tasks",
        "agency-senior-project-manager",
        "agency-macos-spatial-metal-engineer",
        "sp-brainstorming",
        "kwp-cowork-plugin-customizer",
        "agency-xr-interface-architect",
        "agency-devops-automator",
        "agency-ai-engineer",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-marketing-brand-voice",
        "anthropic-claude-api",
        "agency-workflow-optimizer",
        "sp-verification",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-org-planning",
        "kwp-data-data-validation",
        "agency-xr-immersive-developer",
        "agency-frontend-developer",
        "issue-to-code",
        "kwp-design-system"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: reviewing AI-generated code. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"ecc-agentic-engineering\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"ecc-search-first\", \"ecc-autonomous-loops\", \"ecc-coding-standards\", \"ecc-strategic-compact\", \"ecc-eval-harness\", \"ecc-configure\", \"ecc-continuous-learning\", \"ecc-verification-loop\", \"sefo-orchestrator\", \"agency-visionos-spatial-engineer\", \"agency-security-engineer\", \"sp-finish-branch\", \"kwp-finance-close-management\", \"ecc-iterative-retrieval\", \"agency-ux-architect\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0057",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: outage. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "incident-to-improvement"
      ],
      "distractor_skills": [
        "incident-lifecycle-manager",
        "kwp-engineering-incident-response",
        "role-cto",
        "test-suite",
        "deep-research-pipeline",
        "infra-runbook-generator",
        "ui-suite",
        "diagnose",
        "release-commander",
        "pr-to-issue-linker",
        "deep-review",
        "e2e-overhaul",
        "refactor-simulator",
        "onboarding-accelerator",
        "autoskill-pipeline",
        "problem-definition",
        "infra-drift-detector",
        "iac-review-agent",
        "release-ship",
        "skill-optimizer",
        "hf-trending-intelligence",
        "helm-validator",
        "bespin-news-digest",
        "issue-to-code",
        "role-security-engineer",
        "qa-test-expert"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: outage. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"incident-to-improvement\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"incident-lifecycle-manager\", \"kwp-engineering-incident-response\", \"role-cto\", \"test-suite\", \"deep-research-pipeline\", \"infra-runbook-generator\", \"ui-suite\", \"diagnose\", \"release-commander\", \"pr-to-issue-linker\", \"deep-review\", \"e2e-overhaul\", \"refactor-simulator\", \"onboarding-accelerator\", \"autoskill-pipeline\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0058",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Agents Orchestrator agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-agents-orchestrator"
      ],
      "distractor_skills": [
        "agency-report-distribution-agent",
        "agency-brand-guardian",
        "agency-xr-interface-architect",
        "agency-data-consolidation-agent",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-behavioral-nudge-engine",
        "agency-terminal-integration-specialist",
        "agency-workflow-optimizer",
        "agency-rapid-prototyper"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Agents Orchestrator agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-agents-orchestrator\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-report-distribution-agent\", \"agency-brand-guardian\", \"agency-xr-interface-architect\", \"agency-data-consolidation-agent\", \"agency-xr-immersive-developer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-behavioral-nudge-engine\", \"agency-terminal-integration-specialist\", \"agency-workflow-optimizer\", \"agency-rapid-prototyper\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0059",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Source Management. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-enterprise-search-source-management"
      ],
      "distractor_skills": [
        "kwp-enterprise-search-knowledge-synthesis",
        "kwp-enterprise-search-search-strategy",
        "kwp-slack-slack-search",
        "kwp-design-system",
        "kwp-customer-support-customer-research",
        "kwp-sales-account-research",
        "kwp-human-resources-people-analytics",
        "kwp-data-data-context-extractor",
        "kwp-sales-competitive-intelligence",
        "kwp-operations-vendor-management"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Source Management. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-enterprise-search-source-management\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-enterprise-search-knowledge-synthesis\", \"kwp-enterprise-search-search-strategy\", \"kwp-slack-slack-search\", \"kwp-design-system\", \"kwp-customer-support-customer-research\", \"kwp-sales-account-research\", \"kwp-human-resources-people-analytics\", \"kwp-data-data-context-extractor\", \"kwp-sales-competitive-intelligence\", \"kwp-operations-vendor-management\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0060",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: NLM 노트북. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "notebooklm"
      ],
      "distractor_skills": [
        "notebooklm-studio",
        "notebooklm-research",
        "gws-drive",
        "nlm-arxiv-slides",
        "nlm-video",
        "nlm-slides",
        "gws-docs",
        "gws-sheets",
        "gws-chat",
        "nlm-deep-learn",
        "nlm-strategy-hub",
        "public-apis",
        "nlm-discovery-lab",
        "nlm-launch-kit",
        "daiso-mcp",
        "paperclip-tasks",
        "hf-buckets",
        "gws-recipe-doc-from-template",
        "x-to-slack",
        "gws-workspace",
        "hf-collections",
        "gws-recipe-audit-sharing",
        "tab-glossary"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: NLM 노트북. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"notebooklm\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"notebooklm-studio\", \"notebooklm-research\", \"gws-drive\", \"nlm-arxiv-slides\", \"nlm-video\", \"nlm-slides\", \"gws-docs\", \"gws-sheets\", \"gws-chat\", \"nlm-deep-learn\", \"nlm-strategy-hub\", \"public-apis\", \"nlm-discovery-lab\", \"nlm-launch-kit\", \"daiso-mcp\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0061",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: post-mortem. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "incident-to-improvement"
      ],
      "distractor_skills": [
        "incident-lifecycle-manager",
        "kwp-engineering-incident-response",
        "role-cto",
        "test-suite",
        "deep-research-pipeline",
        "infra-runbook-generator",
        "ui-suite",
        "diagnose",
        "release-commander",
        "pr-to-issue-linker",
        "deep-review",
        "e2e-overhaul",
        "refactor-simulator",
        "onboarding-accelerator",
        "autoskill-pipeline",
        "problem-definition",
        "infra-drift-detector",
        "iac-review-agent",
        "release-ship",
        "skill-optimizer",
        "hf-trending-intelligence",
        "helm-validator",
        "bespin-news-digest",
        "issue-to-code",
        "role-security-engineer",
        "qa-test-expert",
        "code-review-all",
        "role-developer",
        "intent-alignment-tracker",
        "backend-expert",
        "service-health-doctor",
        "agency-devops-automator",
        "auto-research"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: post-mortem. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"incident-to-improvement\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"incident-lifecycle-manager\", \"kwp-engineering-incident-response\", \"role-cto\", \"test-suite\", \"deep-research-pipeline\", \"infra-runbook-generator\", \"ui-suite\", \"diagnose\", \"release-commander\", \"pr-to-issue-linker\", \"deep-review\", \"e2e-overhaul\", \"refactor-simulator\", \"onboarding-accelerator\", \"autoskill-pipeline\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0062",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: multi-engine aggregation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "alphaear-search"
      ],
      "distractor_skills": [
        "alphaear-news",
        "alphaear-stock",
        "alphaear-deepear-lite",
        "alphaear-sentiment",
        "alphaear-predictor",
        "stock-csv-downloader",
        "alphaear-signal-tracker",
        "alphaear-logic-visualizer",
        "alphaear-reporter",
        "daily-stock-check",
        "trading-market-news-analyst",
        "hf-papers",
        "trading-market-environment-analysis",
        "cognee",
        "tab-stock-sync",
        "weekly-stock-update",
        "trading-technical-analyst",
        "trading-breadth-chart-analyst",
        "today",
        "tab-news-fetch",
        "trading-sector-analyst",
        "paper-auto-classifier",
        "email-research-dispatcher",
        "deep-research-pipeline",
        "recall",
        "daiso-mcp",
        "agency-ux-researcher",
        "kwp-human-resources-people-analytics",
        "mirofish-financial-sim",
        "kwp-engineering-documentation",
        "gws-recipe-events-from-sheet",
        "gws-recipe-doc-from-template",
        "gws-drive",
        "agency-frontend-developer",
        "local-dev-runner",
        "deep-review",
        "daily-strategy-post"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: multi-engine aggregation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-search\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"alphaear-news\", \"alphaear-stock\", \"alphaear-deepear-lite\", \"alphaear-sentiment\", \"alphaear-predictor\", \"stock-csv-downloader\", \"alphaear-signal-tracker\", \"alphaear-logic-visualizer\", \"alphaear-reporter\", \"daily-stock-check\", \"trading-market-news-analyst\", \"hf-papers\", \"trading-market-environment-analysis\", \"cognee\", \"tab-stock-sync\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0063",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: run a dependency sweep. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "dependency-auditor"
      ],
      "distractor_skills": [
        "ci-quality-gate",
        "compliance-governance",
        "security-expert",
        "qa-test-expert",
        "semantic-guard",
        "kwp-engineering-code-review",
        "sefo-governance",
        "skill-optimizer",
        "deep-review",
        "test-suite",
        "agency-security-engineer",
        "dependency-radar",
        "i18n-sync",
        "ux-expert",
        "code-review-all",
        "role-security-engineer",
        "compliance-gate",
        "github-sprint-digest",
        "kwp-design-accessibility-review",
        "kwp-operations-risk-assessment",
        "setup-doctor",
        "ui-suite",
        "agency-devops-automator",
        "sre-devops-expert",
        "stock-csv-downloader",
        "kwp-engineering-tech-debt",
        "e2e-testing",
        "service-health-doctor",
        "diagnose",
        "local-dev-runner",
        "release-commander",
        "commit-to-issue",
        "helm-validator",
        "alphaear-stock",
        "ai-quality-evaluator",
        "gws-recipe-audit-sharing",
        "terraform-reviewer",
        "kwp-design-user-research",
        "kwp-design-design-system-management",
        "agency-ux-researcher",
        "k8s-manifest-validator",
        "daily-stock-check",
        "simplify",
        "design-architect"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: run a dependency sweep. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"dependency-auditor\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"ci-quality-gate\", \"compliance-governance\", \"security-expert\", \"qa-test-expert\", \"semantic-guard\", \"kwp-engineering-code-review\", \"sefo-governance\", \"skill-optimizer\", \"deep-review\", \"test-suite\", \"agency-security-engineer\", \"dependency-radar\", \"i18n-sync\", \"ux-expert\", \"code-review-all\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0064",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: analyzing channel metrics (email. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-marketing-performance-analytics"
      ],
      "distractor_skills": [
        "kwp-marketing-competitive-analysis",
        "kwp-marketing-campaign-planning",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-people-analytics",
        "kwp-data-statistical-analysis",
        "kwp-product-management-competitive-analysis",
        "kwp-marketing-content-creation",
        "kwp-data-data-validation",
        "kwp-finance-variance-analysis",
        "kwp-finance-financial-statements",
        "kwp-data-interactive-dashboard-builder",
        "kwp-finance-close-management",
        "kwp-data-data-exploration",
        "kwp-sales-daily-briefing",
        "kwp-marketing-brand-voice",
        "kwp-customer-support-response-drafting",
        "kwp-operations-resource-planning",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-product-management-feature-spec",
        "kwp-operations-process-optimization",
        "kwp-customer-support-escalation",
        "kwp-design-design-handoff",
        "kwp-apollo-enrich-lead",
        "kwp-slack-slack-search",
        "kwp-design-system",
        "kwp-finance-reconciliation",
        "kwp-legal-legal-risk-assessment",
        "kwp-common-room-account-research",
        "kwp-legal-contract-review",
        "kwp-data-data-visualization"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: analyzing channel metrics (email. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-marketing-performance-analytics\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-marketing-competitive-analysis\", \"kwp-marketing-campaign-planning\", \"kwp-product-management-metrics-tracking\", \"kwp-human-resources-people-analytics\", \"kwp-data-statistical-analysis\", \"kwp-product-management-competitive-analysis\", \"kwp-marketing-content-creation\", \"kwp-data-data-validation\", \"kwp-finance-variance-analysis\", \"kwp-finance-financial-statements\", \"kwp-data-interactive-dashboard-builder\", \"kwp-finance-close-management\", \"kwp-data-data-exploration\", \"kwp-sales-daily-briefing\", \"kwp-marketing-brand-voice\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0065",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: configuration. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "paperclip-setup"
      ],
      "distractor_skills": [
        "paperclip-agents",
        "paperclip-tasks",
        "paperclip-control",
        "gws-workspace",
        "local-dev-runner",
        "slack-agent",
        "hf-hub",
        "sefo-orchestrator",
        "cursor-automations",
        "sefo-federation",
        "kwp-cowork-plugin-customizer",
        "agency-devops-automator",
        "agency-xr-cockpit-interaction-specialist",
        "hf-models",
        "agency-roster",
        "kwp-operations-change-management",
        "sefo-governance",
        "air-gap-orchestrator",
        "kwp-productivity-task-management",
        "kwp-operations-vendor-management",
        "kwp-cowork-cowork-plugin-customizer",
        "agency-infrastructure-maintainer",
        "gws-recipe-audit-sharing",
        "compliance-governance",
        "mirofish-financial-sim",
        "agency-xr-interface-architect",
        "anthropic-claude-api",
        "kwp-cowork-create-cowork-plugin",
        "sp-subagent-dev",
        "domain-commit",
        "agency-senior-project-manager",
        "agency-lsp-index-engineer",
        "agency-studio-producer",
        "kwp-design-system",
        "agency-xr-immersive-developer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: configuration. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"paperclip-setup\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"paperclip-agents\", \"paperclip-tasks\", \"paperclip-control\", \"gws-workspace\", \"local-dev-runner\", \"slack-agent\", \"hf-hub\", \"sefo-orchestrator\", \"cursor-automations\", \"sefo-federation\", \"kwp-cowork-plugin-customizer\", \"agency-devops-automator\", \"agency-xr-cockpit-interaction-specialist\", \"hf-models\", \"agency-roster\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0066",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: run autoskill evolution. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "autoskill-evolve"
      ],
      "distractor_skills": [
        "autoskill-extractor",
        "autoskill-judge",
        "autoskill-merger",
        "autoskill-pipeline",
        "workflow-miner",
        "kwp-sync",
        "agency-roster",
        "skill-optimizer",
        "sp-writing-skills",
        "skill-composer",
        "anthropic-skill-creator",
        "skill-guide-generator",
        "nlm-slides",
        "paperclip-agents",
        "intent-alignment-tracker",
        "domain-commit",
        "gws-recipe-events-from-sheet",
        "nlm-video",
        "ship",
        "agency-app-store-optimizer",
        "agency-feedback-synthesizer",
        "gws-recipe-personalized-emails",
        "auto-research-distribute",
        "agency-agents-orchestrator",
        "md-enhance-publish",
        "kwp-brand-voice-guideline-generation",
        "gws-recipe-save-email-to-doc",
        "agency-data-consolidation-agent",
        "kwp-cowork-create-cowork-plugin",
        "setup-doctor",
        "agency-xr-interface-architect",
        "agency-devops-automator",
        "kwp-cowork-create-plugin",
        "deep-research-pipeline",
        "ui-suite",
        "release-ship"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: run autoskill evolution. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"autoskill-evolve\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"autoskill-extractor\", \"autoskill-judge\", \"autoskill-merger\", \"autoskill-pipeline\", \"workflow-miner\", \"kwp-sync\", \"agency-roster\", \"skill-optimizer\", \"sp-writing-skills\", \"skill-composer\", \"anthropic-skill-creator\", \"skill-guide-generator\", \"nlm-slides\", \"paperclip-agents\", \"intent-alignment-tracker\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0067",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Screen Description Generator. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "screen-description"
      ],
      "distractor_skills": [
        "design-review",
        "implement-screen",
        "swagger-api-doc-generator",
        "overlay-layout-patterns",
        "fsd-development",
        "fe-pipeline",
        "github-workflow-automation",
        "figma-to-tds",
        "notion-docs-sync",
        "local-dev-setup",
        "google-daily",
        "sp-writing-plans",
        "anthropic-claude-api",
        "sp-subagent-dev",
        "sp-parallel-agents",
        "kwp-slack-slack-messaging",
        "sp-writing-skills",
        "sp-tdd",
        "sp-executing-plans",
        "kwp-slack-slack-search",
        "sp-debugging",
        "sp-finish-branch",
        "tab-turtle-refresh",
        "tab-fundamental-sync",
        "sp-request-review",
        "kwp-data-data-context-extractor",
        "agency-devops-automator",
        "tab-event-detect",
        "tab-dualma-refresh",
        "tab-bollinger-refresh",
        "tab-analysis-run",
        "gws-docs",
        "agency-brand-guardian",
        "tab-technical-analysis",
        "tab-llm-agents",
        "tab-hot-stock-discovery",
        "kwp-design-system",
        "auto-research-distribute",
        "agency-agents-orchestrator",
        "tab-screening",
        "tab-market-breadth",
        "tab-dashboard-summary",
        "kwp-common-room-contact-research",
        "kwp-common-room-account-research",
        "kwp-apollo-sequence-load"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Screen Description Generator. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"screen-description\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"design-review\", \"implement-screen\", \"swagger-api-doc-generator\", \"overlay-layout-patterns\", \"fsd-development\", \"fe-pipeline\", \"github-workflow-automation\", \"figma-to-tds\", \"notion-docs-sync\", \"local-dev-setup\", \"google-daily\", \"sp-writing-plans\", \"anthropic-claude-api\", \"sp-subagent-dev\", \"sp-parallel-agents\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0068",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Write internal communications using company formats. Use for 3P updates (Progress, Plans, Problems), company newsletters, FAQ responses, status reports, leadership updates, project updates, incident r…. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "anthropic-internal-comms"
      ],
      "distractor_skills": [
        "anthropic-theme-factory",
        "kwp-product-management-stakeholder-comms",
        "anthropic-slack-gif-creator",
        "anthropic-brand-guidelines",
        "anthropic-web-artifacts-builder",
        "anthropic-xlsx",
        "anthropic-algorithmic-art",
        "anthropic-doc-coauthoring",
        "anthropic-claude-api",
        "anthropic-skill-creator",
        "anthropic-canvas-design",
        "anthropic-docx",
        "anthropic-webapp-testing",
        "anthropic-mcp-builder",
        "agency-executive-summary-generator",
        "pm-go-to-market",
        "anthropic-frontend-design",
        "anthropic-pptx",
        "pm-toolkit",
        "role-ceo",
        "anthropic-pdf",
        "sp-subagent-dev",
        "agency-tiktok-strategist",
        "gws-docs",
        "kwp-common-room-account-research",
        "agency-reddit-community-builder",
        "sp-executing-plans",
        "sales-meeting-pipeline",
        "kwp-productivity-task-management",
        "agency-xiaohongshu-specialist",
        "kwp-common-room-compose-outreach",
        "pm-marketing-growth",
        "kwp-product-management-roadmap-management"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Write internal communications using company formats. Use for 3P updates (Progress, Plans, Problems), company newsletters, FAQ responses, status reports, leadership updates, project updates, incident r…. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-internal-comms\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"anthropic-theme-factory\", \"kwp-product-management-stakeholder-comms\", \"anthropic-slack-gif-creator\", \"anthropic-brand-guidelines\", \"anthropic-web-artifacts-builder\", \"anthropic-xlsx\", \"anthropic-algorithmic-art\", \"anthropic-doc-coauthoring\", \"anthropic-claude-api\", \"anthropic-skill-creator\", \"anthropic-canvas-design\", \"anthropic-docx\", \"anthropic-webapp-testing\", \"anthropic-mcp-builder\", \"agency-executive-summary-generator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0069",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 딥 리서치 파이프라인. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "deep-research-pipeline"
      ],
      "distractor_skills": [
        "paper-review",
        "role-ceo",
        "deep-review",
        "role-cso",
        "critical-review",
        "role-sales",
        "role-developer",
        "role-ux-designer",
        "role-dispatcher",
        "email-research-dispatcher",
        "role-data-scientist",
        "role-cto",
        "role-finance",
        "ui-suite",
        "hf-trending-intelligence",
        "role-pm",
        "auto-research",
        "meeting-digest",
        "role-hr",
        "role-trading-expert",
        "ship",
        "test-suite",
        "md-enhance-publish",
        "role-security-engineer",
        "autoskill-pipeline",
        "kwp-human-resources-people-analytics",
        "weekly-status-report",
        "twitter-timeline-to-slack",
        "nlm-arxiv-slides"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 딥 리서치 파이프라인. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"deep-research-pipeline\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"paper-review\", \"role-ceo\", \"deep-review\", \"role-cso\", \"critical-review\", \"role-sales\", \"role-developer\", \"role-ux-designer\", \"role-dispatcher\", \"email-research-dispatcher\", \"role-data-scientist\", \"role-cto\", \"role-finance\", \"ui-suite\", \"hf-trending-intelligence\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0070",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: check inbox. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "gws-gmail"
      ],
      "distractor_skills": [
        "gws-chat",
        "gws-calendar",
        "gws-workflows",
        "gws-recipe-personalized-emails",
        "gws-workspace",
        "gws-sheets",
        "gws-drive",
        "gws-docs",
        "gws-email-reply",
        "gws-recipe-save-email-to-doc",
        "gws-recipe-audit-sharing",
        "gws-recipe-events-from-sheet",
        "gws-recipe-doc-from-template",
        "calendar-daily-briefing",
        "gmail-daily-triage",
        "email-auto-reply",
        "proactive-meeting-scheduler",
        "ai-chief-of-staff",
        "hf-models",
        "kwp-operations-vendor-management",
        "hf-buckets",
        "hf-discussions",
        "hf-hub",
        "hf-papers",
        "hf-endpoints",
        "hf-repos",
        "google-daily",
        "skill-optimizer",
        "kwp-engineering-code-review",
        "kwp-design-accessibility-review",
        "transcribee",
        "kwp-finance-close-management",
        "anthropic-claude-api",
        "i18n-sync",
        "hf-collections"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: check inbox. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"gws-gmail\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"gws-chat\", \"gws-calendar\", \"gws-workflows\", \"gws-recipe-personalized-emails\", \"gws-workspace\", \"gws-sheets\", \"gws-drive\", \"gws-docs\", \"gws-email-reply\", \"gws-recipe-save-email-to-doc\", \"gws-recipe-audit-sharing\", \"gws-recipe-events-from-sheet\", \"gws-recipe-doc-from-template\", \"calendar-daily-briefing\", \"gmail-daily-triage\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0071",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: fetch stock CSVs. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "stock-csv-downloader"
      ],
      "distractor_skills": [
        "weekly-stock-update",
        "daily-stock-check",
        "alphaear-stock",
        "tab-stock-sync",
        "trading-sector-analyst",
        "alphaear-search",
        "alphaear-news",
        "alphaear-deepear-lite",
        "trading-market-environment-analysis",
        "trading-technical-analyst",
        "tab-kiwoom",
        "today",
        "tab-report-generate",
        "trading-us-stock-analysis",
        "trading-theme-detector",
        "alphaear-reporter",
        "trading-breadth-chart-analyst"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: fetch stock CSVs. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"stock-csv-downloader\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"weekly-stock-update\", \"daily-stock-check\", \"alphaear-stock\", \"tab-stock-sync\", \"trading-sector-analyst\", \"alphaear-search\", \"alphaear-news\", \"alphaear-deepear-lite\", \"trading-market-environment-analysis\", \"trading-technical-analyst\", \"tab-kiwoom\", \"today\", \"tab-report-generate\", \"trading-us-stock-analysis\", \"trading-theme-detector\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0072",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Guideline Generation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-brand-voice-guideline-generation"
      ],
      "distractor_skills": [
        "kwp-brand-voice-discover-brand",
        "kwp-brand-voice-brand-voice-enforcement",
        "kwp-marketing-brand-voice",
        "kwp-human-resources-people-analytics",
        "kwp-data-data-context-extractor",
        "kwp-cowork-create-plugin",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-product-management-user-research-synthesis",
        "kwp-cowork-create-cowork-plugin",
        "kwp-sales-call-prep",
        "kwp-operations-process-optimization",
        "kwp-productivity-task-management",
        "kwp-customer-support-knowledge-management",
        "kwp-design-design-handoff",
        "kwp-product-management-competitive-analysis",
        "kwp-design-system",
        "kwp-design-design-system-management",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-compensation-benchmarking"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Guideline Generation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-brand-voice-guideline-generation\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-brand-voice-discover-brand\", \"kwp-brand-voice-brand-voice-enforcement\", \"kwp-marketing-brand-voice\", \"kwp-human-resources-people-analytics\", \"kwp-data-data-context-extractor\", \"kwp-cowork-create-plugin\", \"kwp-operations-vendor-management\", \"kwp-human-resources-interview-prep\", \"kwp-product-management-user-research-synthesis\", \"kwp-cowork-create-cowork-plugin\", \"kwp-sales-call-prep\", \"kwp-operations-process-optimization\", \"kwp-productivity-task-management\", \"kwp-customer-support-knowledge-management\", \"kwp-design-design-handoff\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0073",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: error models. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "backend-expert"
      ],
      "distractor_skills": [
        "db-expert",
        "performance-profiler",
        "agency-backend-architect",
        "kwp-engineering-code-review",
        "sre-devops-expert",
        "agency-devops-automator",
        "kwp-engineering-system-design",
        "frontend-expert",
        "kwp-engineering-tech-debt",
        "agency-security-engineer",
        "service-health-doctor",
        "compliance-governance",
        "kwp-engineering-incident-response",
        "kwp-design-design-system-management",
        "ux-expert",
        "local-dev-runner",
        "ecc-coding-standards",
        "technical-writer",
        "agency-xr-immersive-developer",
        "agency-frontend-developer",
        "agency-brand-guardian",
        "ci-quality-gate",
        "agency-api-tester",
        "k8s-manifest-validator",
        "agency-visionos-spatial-engineer",
        "agency-app-store-optimizer",
        "anthropic-claude-api",
        "infra-drift-detector",
        "deep-review",
        "kwp-design-accessibility-review",
        "kwp-design-system",
        "agency-xr-interface-architect",
        "qa-test-expert",
        "evals-skills",
        "agency-ux-architect",
        "agency-rapid-prototyper",
        "agency-infrastructure-maintainer",
        "e2e-testing",
        "agency-xr-cockpit-interaction-specialist",
        "agency-tool-evaluator",
        "agency-legal-compliance-checker",
        "terraform-reviewer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: error models. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"backend-expert\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"db-expert\", \"performance-profiler\", \"agency-backend-architect\", \"kwp-engineering-code-review\", \"sre-devops-expert\", \"agency-devops-automator\", \"kwp-engineering-system-design\", \"frontend-expert\", \"kwp-engineering-tech-debt\", \"agency-security-engineer\", \"service-health-doctor\", \"compliance-governance\", \"kwp-engineering-incident-response\", \"kwp-design-design-system-management\", \"ux-expert\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0074",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: a security review. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "security-expert"
      ],
      "distractor_skills": [
        "compliance-governance",
        "ci-quality-gate",
        "semantic-guard",
        "role-security-engineer",
        "code-review-all",
        "technical-writer",
        "agency-security-engineer",
        "gws-recipe-audit-sharing",
        "dependency-auditor",
        "performance-profiler",
        "commit-to-issue",
        "ux-expert",
        "sre-devops-expert",
        "kwp-operations-compliance-tracking",
        "k8s-manifest-validator",
        "daily-stock-check",
        "domain-commit",
        "compliance-gate",
        "helm-validator",
        "calendar-daily-briefing",
        "simplify",
        "problem-definition",
        "agency-legal-compliance-checker",
        "auto-research-distribute",
        "plans",
        "daily-strategy-post",
        "auto-research",
        "onboarding-accelerator",
        "release-commander",
        "eod-ship",
        "critical-review"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: a security review. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"security-expert\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"compliance-governance\", \"ci-quality-gate\", \"semantic-guard\", \"role-security-engineer\", \"code-review-all\", \"technical-writer\", \"agency-security-engineer\", \"gws-recipe-audit-sharing\", \"dependency-auditor\", \"performance-profiler\", \"commit-to-issue\", \"ux-expert\", \"sre-devops-expert\", \"kwp-operations-compliance-tracking\", \"k8s-manifest-validator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0075",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: ad-hoc historical price queries. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "alphaear-stock"
      ],
      "distractor_skills": [
        "stock-csv-downloader",
        "alphaear-search",
        "alphaear-deepear-lite",
        "weekly-stock-update",
        "daily-stock-check",
        "alphaear-news",
        "tab-stock-sync",
        "alphaear-predictor",
        "alphaear-sentiment",
        "trading-sector-analyst",
        "alphaear-reporter",
        "alphaear-signal-tracker",
        "trading-technical-analyst",
        "trading-us-stock-analysis",
        "trading-market-environment-analysis",
        "trading-stanley-druckenmiller-investment",
        "kwp-sync",
        "agency-xr-interface-architect",
        "today",
        "agency-ux-architect",
        "kwp-engineering-tech-debt",
        "trading-breadth-chart-analyst",
        "gws-recipe-doc-from-template",
        "gws-drive",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "hf-models",
        "agency-test-results-analyzer",
        "kwp-data-data-context-extractor",
        "agency-terminal-integration-specialist",
        "agency-brand-guardian",
        "tab-report-generate",
        "trading-theme-detector",
        "gws-workspace"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: ad-hoc historical price queries. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-stock\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"stock-csv-downloader\", \"alphaear-search\", \"alphaear-deepear-lite\", \"weekly-stock-update\", \"daily-stock-check\", \"alphaear-news\", \"tab-stock-sync\", \"alphaear-predictor\", \"alphaear-sentiment\", \"trading-sector-analyst\", \"alphaear-reporter\", \"alphaear-signal-tracker\", \"trading-technical-analyst\", \"trading-us-stock-analysis\", \"trading-market-environment-analysis\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0076",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Agents Orchestrator agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-agents-orchestrator"
      ],
      "distractor_skills": [
        "agency-report-distribution-agent",
        "agency-brand-guardian",
        "agency-xr-interface-architect",
        "agency-data-consolidation-agent",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-behavioral-nudge-engine",
        "agency-terminal-integration-specialist",
        "agency-workflow-optimizer",
        "agency-rapid-prototyper",
        "agency-mobile-app-builder",
        "agency-visionos-spatial-engineer",
        "agency-senior-developer",
        "agency-app-store-optimizer",
        "agency-senior-project-manager"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Agents Orchestrator agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-agents-orchestrator\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-report-distribution-agent\", \"agency-brand-guardian\", \"agency-xr-interface-architect\", \"agency-data-consolidation-agent\", \"agency-xr-immersive-developer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-behavioral-nudge-engine\", \"agency-terminal-integration-specialist\", \"agency-workflow-optimizer\", \"agency-rapid-prototyper\", \"agency-mobile-app-builder\", \"agency-visionos-spatial-engineer\", \"agency-senior-developer\", \"agency-app-store-optimizer\", \"agency-senior-project-manager\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0077",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: troubleshooting docs. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-customer-support-knowledge-management"
      ],
      "distractor_skills": [
        "kwp-customer-support-ticket-triage",
        "kwp-human-resources-interview-prep",
        "kwp-design-system",
        "kwp-data-data-context-extractor",
        "kwp-operations-vendor-management",
        "kwp-customer-support-escalation",
        "kwp-slack-slack-search",
        "kwp-customer-support-customer-research",
        "kwp-human-resources-people-analytics",
        "kwp-customer-support-response-drafting",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-org-planning",
        "kwp-legal-nda-triage",
        "kwp-finance-financial-statements",
        "kwp-product-management-feature-spec",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-finance-audit-support",
        "kwp-data-interactive-dashboard-builder",
        "kwp-finance-close-management",
        "kwp-enterprise-search-source-management",
        "kwp-design-design-handoff",
        "kwp-product-management-user-research-synthesis",
        "kwp-product-management-roadmap-management",
        "kwp-legal-legal-risk-assessment",
        "kwp-finance-journal-entry-prep",
        "kwp-product-management-competitive-analysis",
        "kwp-finance-reconciliation",
        "kwp-cowork-plugin-customizer",
        "kwp-apollo-prospect",
        "kwp-operations-resource-planning",
        "kwp-legal-contract-review",
        "kwp-data-data-visualization",
        "kwp-marketing-competitive-analysis",
        "kwp-marketing-brand-voice",
        "kwp-cowork-create-plugin",
        "kwp-sales-draft-outreach",
        "kwp-sales-competitive-intelligence",
        "kwp-product-management-metrics-tracking",
        "kwp-data-statistical-analysis",
        "kwp-data-data-validation",
        "kwp-brand-voice-guideline-generation",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-finance-variance-analysis",
        "kwp-common-room-compose-outreach",
        "kwp-common-room-call-prep",
        "kwp-legal-canned-responses",
        "kwp-enterprise-search-knowledge-synthesis"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: troubleshooting docs. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-customer-support-knowledge-management\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-customer-support-ticket-triage\", \"kwp-human-resources-interview-prep\", \"kwp-design-system\", \"kwp-data-data-context-extractor\", \"kwp-operations-vendor-management\", \"kwp-customer-support-escalation\", \"kwp-slack-slack-search\", \"kwp-customer-support-customer-research\", \"kwp-human-resources-people-analytics\", \"kwp-customer-support-response-drafting\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-human-resources-org-planning\", \"kwp-legal-nda-triage\", \"kwp-finance-financial-statements\", \"kwp-product-management-feature-spec\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0078",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: references agency-sprint-prioritizer. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-sprint-prioritizer"
      ],
      "distractor_skills": [
        "agency-experiment-tracker",
        "agency-ux-researcher",
        "agency-frontend-developer",
        "agency-growth-hacker",
        "agency-wechat-official-account-manager",
        "agency-devops-automator",
        "agency-brand-guardian",
        "agency-reddit-community-builder",
        "agency-data-analytics-reporter",
        "agency-analytics-reporter",
        "agency-tool-evaluator",
        "agency-performance-benchmarker",
        "agency-app-store-optimizer",
        "agency-studio-operations"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: references agency-sprint-prioritizer. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-sprint-prioritizer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-experiment-tracker\", \"agency-ux-researcher\", \"agency-frontend-developer\", \"agency-growth-hacker\", \"agency-wechat-official-account-manager\", \"agency-devops-automator\", \"agency-brand-guardian\", \"agency-reddit-community-builder\", \"agency-data-analytics-reporter\", \"agency-analytics-reporter\", \"agency-tool-evaluator\", \"agency-performance-benchmarker\", \"agency-app-store-optimizer\", \"agency-studio-operations\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0079",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 전체 코드 리뷰. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "code-review-all"
      ],
      "distractor_skills": [
        "deep-review",
        "ci-quality-gate",
        "simplify",
        "critical-review",
        "security-expert",
        "auto-research",
        "commit-to-issue",
        "sre-devops-expert",
        "problem-definition",
        "k8s-manifest-validator",
        "performance-profiler",
        "ui-suite",
        "helm-validator",
        "ship",
        "refactor-simulator",
        "agency-test-results-analyzer",
        "agency-api-tester",
        "test-suite",
        "setup-doctor",
        "domain-commit",
        "eod-ship",
        "onboarding-accelerator",
        "codebase-archaeologist",
        "agency-infrastructure-maintainer",
        "agency-xr-immersive-developer",
        "kwp-engineering-tech-debt",
        "calendar-daily-briefing",
        "diagnose",
        "technical-writer",
        "related-papers-scout",
        "kwp-engineering-code-review",
        "agency-app-store-optimizer",
        "jira-linear-adapter"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 전체 코드 리뷰. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"code-review-all\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"deep-review\", \"ci-quality-gate\", \"simplify\", \"critical-review\", \"security-expert\", \"auto-research\", \"commit-to-issue\", \"sre-devops-expert\", \"problem-definition\", \"k8s-manifest-validator\", \"performance-profiler\", \"ui-suite\", \"helm-validator\", \"ship\", \"refactor-simulator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0080",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: date ranges). 3. If too few results. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-slack-slack-search"
      ],
      "distractor_skills": [
        "kwp-slack-slack-messaging",
        "kwp-design-system",
        "kwp-data-data-context-extractor",
        "kwp-common-room-compose-outreach",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-finance-close-management",
        "kwp-enterprise-search-source-management",
        "kwp-design-design-handoff",
        "kwp-apollo-enrich-lead",
        "kwp-operations-change-management",
        "kwp-enterprise-search-knowledge-synthesis"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: date ranges). 3. If too few results. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-slack-slack-search\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-slack-slack-messaging\", \"kwp-design-system\", \"kwp-data-data-context-extractor\", \"kwp-common-room-compose-outreach\", \"kwp-operations-vendor-management\", \"kwp-human-resources-interview-prep\", \"kwp-finance-close-management\", \"kwp-enterprise-search-source-management\", \"kwp-design-design-handoff\", \"kwp-apollo-enrich-lead\", \"kwp-operations-change-management\", \"kwp-enterprise-search-knowledge-synthesis\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0081",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 0DTE trade. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "trading-options-theta"
      ],
      "distractor_skills": [
        "trading-options-strategy-advisor",
        "trading-sector-analyst",
        "trading-position-sizer",
        "trading-market-environment-analysis",
        "trading-market-top-detector",
        "trading-breadth-chart-analyst",
        "trading-theme-detector",
        "trading-stanley-druckenmiller-investment",
        "trading-trade-hypothesis-ideator",
        "trading-market-breadth-analyzer",
        "trading-edge-candidate-agent",
        "trading-uptrend-analyzer",
        "trading-us-stock-analysis",
        "trading-technical-analyst",
        "trading-market-news-analyst",
        "trading-backtest-expert",
        "trading-us-market-bubble-detector",
        "trading-data-quality-checker",
        "trading-finviz-screener",
        "mirofish-financial-sim",
        "trading-scenario-analyzer",
        "stock-csv-downloader",
        "daily-stock-check",
        "role-trading-expert",
        "alphaear-sentiment",
        "weekly-stock-update",
        "mirofish",
        "today",
        "alphaear-stock",
        "alphaear-deepear-lite",
        "tab-strategy-comparison",
        "daily-am-orchestrator",
        "tab-kiwoom",
        "morning-ship",
        "e2e-testing"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 0DTE trade. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"trading-options-theta\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"trading-options-strategy-advisor\", \"trading-sector-analyst\", \"trading-position-sizer\", \"trading-market-environment-analysis\", \"trading-market-top-detector\", \"trading-breadth-chart-analyst\", \"trading-theme-detector\", \"trading-stanley-druckenmiller-investment\", \"trading-trade-hypothesis-ideator\", \"trading-market-breadth-analyzer\", \"trading-edge-candidate-agent\", \"trading-uptrend-analyzer\", \"trading-us-stock-analysis\", \"trading-technical-analyst\", \"trading-market-news-analyst\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0082",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 노트북 삭제. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "notebooklm"
      ],
      "distractor_skills": [
        "notebooklm-studio",
        "notebooklm-research",
        "gws-drive",
        "nlm-arxiv-slides",
        "nlm-video",
        "nlm-slides",
        "gws-docs",
        "gws-sheets",
        "gws-chat",
        "nlm-deep-learn",
        "nlm-strategy-hub",
        "public-apis",
        "nlm-discovery-lab",
        "nlm-launch-kit",
        "daiso-mcp",
        "paperclip-tasks",
        "hf-buckets",
        "gws-recipe-doc-from-template",
        "x-to-slack",
        "gws-workspace",
        "hf-collections",
        "gws-recipe-audit-sharing",
        "tab-glossary",
        "gws-recipe-save-email-to-doc",
        "hf-endpoints",
        "defuddle",
        "gws-calendar",
        "gws-recipe-events-from-sheet",
        "hf-discussions",
        "gws-gmail",
        "tab-market-breadth",
        "transcribee",
        "hf-papers",
        "gws-workflows",
        "tab-news-fetch",
        "alphaxiv-paper-lookup",
        "paperclip-control",
        "tab-report-generate",
        "notion-meeting-sync",
        "hf-repos",
        "daily-am-orchestrator",
        "cognee",
        "knowledge-to-paper"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 노트북 삭제. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"notebooklm\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"notebooklm-studio\", \"notebooklm-research\", \"gws-drive\", \"nlm-arxiv-slides\", \"nlm-video\", \"nlm-slides\", \"gws-docs\", \"gws-sheets\", \"gws-chat\", \"nlm-deep-learn\", \"nlm-strategy-hub\", \"public-apis\", \"nlm-discovery-lab\", \"nlm-launch-kit\", \"daiso-mcp\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0083",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: review resume. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "pm-toolkit"
      ],
      "distractor_skills": [
        "pm-go-to-market",
        "pm-marketing-growth",
        "pm-data-analytics",
        "pm-market-research",
        "pm-product-discovery",
        "pm-product-strategy",
        "pm-execution",
        "i18n-sync",
        "pr-review-captain",
        "compliance-governance",
        "skill-optimizer",
        "agency-tiktok-strategist",
        "gws-recipe-audit-sharing",
        "agency-reddit-community-builder",
        "docs-tutor-setup",
        "anthropic-doc-coauthoring",
        "agency-xiaohongshu-specialist",
        "agency-legal-compliance-checker",
        "ux-expert",
        "gws-recipe-doc-from-template",
        "e2e-testing",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "technical-writer",
        "kwp-engineering-code-review",
        "agency-app-store-optimizer",
        "agency-technical-writer",
        "kwp-sync",
        "agency-instagram-curator",
        "agency-accessibility-auditor",
        "agency-xr-interface-architect",
        "agency-devops-automator",
        "prompt-transformer",
        "kwp-design-design-system-management",
        "intent-alignment-tracker",
        "agency-ux-architect",
        "agency-rapid-prototyper",
        "kwp-marketing-brand-voice",
        "pm-sentinel",
        "kwp-engineering-documentation",
        "gws-calendar",
        "agency-wechat-official-account-manager",
        "gws-recipe-events-from-sheet",
        "gws-drive",
        "agency-report-distribution-agent",
        "agency-frontend-developer",
        "semantic-guard",
        "role-developer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: review resume. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"pm-toolkit\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"pm-go-to-market\", \"pm-marketing-growth\", \"pm-data-analytics\", \"pm-market-research\", \"pm-product-discovery\", \"pm-product-strategy\", \"pm-execution\", \"i18n-sync\", \"pr-review-captain\", \"compliance-governance\", \"skill-optimizer\", \"agency-tiktok-strategist\", \"gws-recipe-audit-sharing\", \"agency-reddit-community-builder\", \"docs-tutor-setup\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0084",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: market participation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "trading-uptrend-analyzer"
      ],
      "distractor_skills": [
        "trading-market-breadth-analyzer",
        "trading-market-top-detector",
        "trading-breadth-chart-analyst",
        "trading-market-environment-analysis",
        "trading-sector-analyst",
        "trading-theme-detector",
        "trading-position-sizer",
        "trading-stanley-druckenmiller-investment",
        "trading-trade-hypothesis-ideator",
        "trading-options-strategy-advisor",
        "trading-finviz-screener",
        "trading-edge-candidate-agent",
        "trading-options-theta",
        "trading-technical-analyst",
        "trading-us-market-bubble-detector",
        "trading-us-stock-analysis",
        "trading-backtest-expert",
        "alphaear-sentiment",
        "trading-market-news-analyst",
        "alphaear-deepear-lite",
        "stock-csv-downloader"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: market participation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"trading-uptrend-analyzer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"trading-market-breadth-analyzer\", \"trading-market-top-detector\", \"trading-breadth-chart-analyst\", \"trading-market-environment-analysis\", \"trading-sector-analyst\", \"trading-theme-detector\", \"trading-position-sizer\", \"trading-stanley-druckenmiller-investment\", \"trading-trade-hypothesis-ideator\", \"trading-options-strategy-advisor\", \"trading-finviz-screener\", \"trading-edge-candidate-agent\", \"trading-options-theta\", \"trading-technical-analyst\", \"trading-us-market-bubble-detector\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0085",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Design Handoff. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-design-design-handoff"
      ],
      "distractor_skills": [
        "kwp-human-resources-interview-prep",
        "kwp-human-resources-people-analytics",
        "kwp-human-resources-org-planning",
        "kwp-slack-slack-search",
        "kwp-operations-vendor-management",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-cowork-create-plugin",
        "kwp-design-system",
        "kwp-customer-support-escalation",
        "kwp-data-data-context-extractor",
        "kwp-operations-resource-planning",
        "kwp-data-data-visualization",
        "kwp-product-management-stakeholder-comms",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-data-data-validation",
        "kwp-finance-variance-analysis",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-finance-audit-support",
        "kwp-enterprise-search-knowledge-synthesis",
        "kwp-finance-close-management",
        "kwp-product-management-competitive-analysis",
        "kwp-sales-daily-briefing",
        "kwp-sales-account-research",
        "kwp-product-management-user-research-synthesis",
        "kwp-marketing-performance-analytics",
        "kwp-legal-legal-risk-assessment",
        "kwp-finance-journal-entry-prep",
        "kwp-common-room-account-research",
        "kwp-design-design-system-management",
        "kwp-sales-draft-outreach",
        "kwp-sales-competitive-intelligence",
        "kwp-legal-nda-triage",
        "kwp-engineering-documentation",
        "kwp-finance-financial-statements",
        "kwp-sales-call-prep",
        "kwp-bio-research-single-cell-rna-qc"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Design Handoff. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-design-design-handoff\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-human-resources-interview-prep\", \"kwp-human-resources-people-analytics\", \"kwp-human-resources-org-planning\", \"kwp-slack-slack-search\", \"kwp-operations-vendor-management\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-cowork-create-plugin\", \"kwp-design-system\", \"kwp-customer-support-escalation\", \"kwp-data-data-context-extractor\", \"kwp-operations-resource-planning\", \"kwp-data-data-visualization\", \"kwp-product-management-stakeholder-comms\", \"kwp-product-management-metrics-tracking\", \"kwp-human-resources-recruiting-pipeline\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0086",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 자동화 ROI. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "automation-strategist"
      ],
      "distractor_skills": [
        "pipeline-builder",
        "system-thinker",
        "technical-writer",
        "sre-devops-expert",
        "plans",
        "setup-doctor",
        "daily-strategy-post",
        "codebase-archaeologist",
        "ci-quality-gate",
        "daily-am-orchestrator",
        "auto-research-distribute",
        "context-engineer",
        "refactor-simulator",
        "onboarding-accelerator",
        "sprint-orchestrator",
        "eod-ship",
        "critical-review",
        "qa-test-expert",
        "problem-definition",
        "calendar-daily-briefing",
        "feedback-meeting-scheduler",
        "daily-stock-check",
        "performance-profiler",
        "auto-research",
        "executive-briefing",
        "decision-router",
        "security-expert",
        "cursor-automations",
        "commit-to-issue",
        "full-stack-planner",
        "k8s-manifest-validator",
        "alphaxiv-paper-lookup",
        "kwp-engineering-testing-strategy",
        "helm-validator",
        "kwp-human-resources-org-planning",
        "domain-commit",
        "code-review-all",
        "ai-workflow-integrator",
        "jira-linear-adapter",
        "trading-stanley-druckenmiller-investment",
        "prd-research-factory",
        "kwp-operations-risk-assessment",
        "kwp-engineering-system-design"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 자동화 ROI. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"automation-strategist\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"pipeline-builder\", \"system-thinker\", \"technical-writer\", \"sre-devops-expert\", \"plans\", \"setup-doctor\", \"daily-strategy-post\", \"codebase-archaeologist\", \"ci-quality-gate\", \"daily-am-orchestrator\", \"auto-research-distribute\", \"context-engineer\", \"refactor-simulator\", \"onboarding-accelerator\", \"sprint-orchestrator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0087",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Web Application Testing. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "anthropic-webapp-testing"
      ],
      "distractor_skills": [
        "e2e-testing",
        "anthropic-web-artifacts-builder",
        "kwp-engineering-testing-strategy",
        "anthropic-frontend-design",
        "qa-test-expert",
        "frontend-expert",
        "anthropic-canvas-design",
        "anthropic-claude-api",
        "anthropic-skill-creator",
        "anthropic-theme-factory",
        "anthropic-slack-gif-creator",
        "anthropic-brand-guidelines"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Web Application Testing. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-webapp-testing\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"e2e-testing\", \"anthropic-web-artifacts-builder\", \"kwp-engineering-testing-strategy\", \"anthropic-frontend-design\", \"qa-test-expert\", \"frontend-expert\", \"anthropic-canvas-design\", \"anthropic-claude-api\", \"anthropic-skill-creator\", \"anthropic-theme-factory\", \"anthropic-slack-gif-creator\", \"anthropic-brand-guidelines\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0088",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: set up the development stack from scratch. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "local-dev-runner"
      ],
      "distractor_skills": [
        "service-health-doctor",
        "agency-devops-automator",
        "gws-workspace",
        "kwp-engineering-incident-response",
        "backend-expert",
        "agency-brand-guardian",
        "agency-performance-benchmarker",
        "agency-backend-architect",
        "sre-devops-expert",
        "e2e-testing",
        "agency-xr-immersive-developer",
        "agency-frontend-developer",
        "kwp-engineering-tech-debt",
        "paperclip-setup",
        "gws-docs",
        "ci-quality-gate",
        "agency-test-results-analyzer",
        "kwp-engineering-code-review",
        "agency-app-store-optimizer",
        "db-expert"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: set up the development stack from scratch. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"local-dev-runner\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"service-health-doctor\", \"agency-devops-automator\", \"gws-workspace\", \"kwp-engineering-incident-response\", \"backend-expert\", \"agency-brand-guardian\", \"agency-performance-benchmarker\", \"agency-backend-architect\", \"sre-devops-expert\", \"e2e-testing\", \"agency-xr-immersive-developer\", \"agency-frontend-developer\", \"kwp-engineering-tech-debt\", \"paperclip-setup\", \"gws-docs\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0089",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Evidence Collector agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-evidence-collector"
      ],
      "distractor_skills": [
        "agency-reality-checker",
        "agency-xr-cockpit-interaction-specialist",
        "agency-terminal-integration-specialist",
        "agency-xr-interface-architect",
        "agency-senior-developer",
        "agency-app-store-optimizer",
        "agency-xr-immersive-developer",
        "agency-report-distribution-agent",
        "agency-behavioral-nudge-engine",
        "agency-brand-guardian",
        "agency-ux-architect",
        "agency-rapid-prototyper",
        "agency-visionos-spatial-engineer",
        "agency-tool-evaluator",
        "agency-macos-spatial-metal-engineer",
        "agency-lsp-index-engineer",
        "agency-instagram-curator",
        "agency-test-results-analyzer",
        "agency-inclusive-visuals-specialist",
        "agency-api-tester",
        "agency-agents-orchestrator",
        "agency-workflow-optimizer",
        "agency-cultural-intelligence-strategist",
        "agency-mobile-app-builder",
        "agency-data-consolidation-agent",
        "agency-tiktok-strategist",
        "agency-senior-project-manager",
        "agency-reddit-community-builder",
        "agency-sales-data-extraction-agent",
        "agency-infrastructure-maintainer",
        "agency-xiaohongshu-specialist",
        "agency-legal-compliance-checker",
        "agency-whimsy-injector"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Evidence Collector agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-evidence-collector\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-reality-checker\", \"agency-xr-cockpit-interaction-specialist\", \"agency-terminal-integration-specialist\", \"agency-xr-interface-architect\", \"agency-senior-developer\", \"agency-app-store-optimizer\", \"agency-xr-immersive-developer\", \"agency-report-distribution-agent\", \"agency-behavioral-nudge-engine\", \"agency-brand-guardian\", \"agency-ux-architect\", \"agency-rapid-prototyper\", \"agency-visionos-spatial-engineer\", \"agency-tool-evaluator\", \"agency-macos-spatial-metal-engineer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0090",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: sync Drive sources. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "notebooklm"
      ],
      "distractor_skills": [
        "notebooklm-studio",
        "notebooklm-research",
        "gws-drive",
        "nlm-arxiv-slides",
        "nlm-video",
        "nlm-slides",
        "gws-docs",
        "gws-sheets",
        "gws-chat",
        "nlm-deep-learn"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: sync Drive sources. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"notebooklm\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"notebooklm-studio\", \"notebooklm-research\", \"gws-drive\", \"nlm-arxiv-slides\", \"nlm-video\", \"nlm-slides\", \"gws-docs\", \"gws-sheets\", \"gws-chat\", \"nlm-deep-learn\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0091",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: references agency-social-media-strategist. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-social-media-strategist"
      ],
      "distractor_skills": [
        "agency-content-creator",
        "agency-twitter-engager",
        "agency-brand-guardian",
        "agency-xr-interface-architect",
        "agency-zhihu-strategist",
        "agency-xr-immersive-developer",
        "agency-wechat-official-account-manager",
        "agency-mobile-app-builder",
        "agency-app-store-optimizer",
        "agency-tool-evaluator",
        "agency-xr-cockpit-interaction-specialist",
        "agency-terminal-integration-specialist",
        "agency-developer-advocate",
        "agency-xiaohongshu-specialist",
        "agency-growth-hacker",
        "agency-visionos-spatial-engineer",
        "agency-tiktok-strategist",
        "agency-reddit-community-builder",
        "agency-instagram-curator",
        "agency-test-results-analyzer",
        "agency-inclusive-visuals-specialist",
        "agency-api-tester",
        "agency-report-distribution-agent",
        "agency-behavioral-nudge-engine",
        "agency-workflow-optimizer",
        "agency-cultural-intelligence-strategist",
        "agency-ux-architect"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: references agency-social-media-strategist. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-social-media-strategist\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-content-creator\", \"agency-twitter-engager\", \"agency-brand-guardian\", \"agency-xr-interface-architect\", \"agency-zhihu-strategist\", \"agency-xr-immersive-developer\", \"agency-wechat-official-account-manager\", \"agency-mobile-app-builder\", \"agency-app-store-optimizer\", \"agency-tool-evaluator\", \"agency-xr-cockpit-interaction-specialist\", \"agency-terminal-integration-specialist\", \"agency-developer-advocate\", \"agency-xiaohongshu-specialist\", \"agency-growth-hacker\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0092",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: mail merge. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "gws-recipe-personalized-emails"
      ],
      "distractor_skills": [
        "gws-recipe-events-from-sheet",
        "gws-chat",
        "gws-gmail",
        "gws-drive",
        "gws-workspace",
        "gws-sheets",
        "gws-recipe-doc-from-template",
        "gws-recipe-save-email-to-doc",
        "gws-workflows",
        "gws-docs",
        "gws-calendar",
        "gws-recipe-audit-sharing",
        "gws-email-reply",
        "calendar-daily-briefing",
        "domain-commit",
        "gmail-daily-triage",
        "stock-csv-downloader",
        "kwp-human-resources-people-analytics",
        "twitter-timeline-to-slack",
        "email-research-dispatcher",
        "skill-optimizer",
        "email-auto-reply",
        "compliance-governance",
        "bespin-news-digest",
        "anthropic-claude-api",
        "agency-xr-interface-architect",
        "agency-devops-automator",
        "agency-ux-researcher",
        "agency-data-consolidation-agent",
        "smart-meeting-scheduler",
        "alphaear-search",
        "alphaear-news",
        "pm-data-analytics",
        "daily-stock-check",
        "proactive-meeting-scheduler"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: mail merge. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"gws-recipe-personalized-emails\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"gws-recipe-events-from-sheet\", \"gws-chat\", \"gws-gmail\", \"gws-drive\", \"gws-workspace\", \"gws-sheets\", \"gws-recipe-doc-from-template\", \"gws-recipe-save-email-to-doc\", \"gws-workflows\", \"gws-docs\", \"gws-calendar\", \"gws-recipe-audit-sharing\", \"gws-email-reply\", \"calendar-daily-briefing\", \"domain-commit\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0093",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: decision documentation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-product-management-stakeholder-comms"
      ],
      "distractor_skills": [
        "kwp-customer-support-escalation",
        "kwp-design-design-handoff",
        "kwp-marketing-campaign-planning",
        "kwp-customer-support-response-drafting",
        "kwp-product-management-metrics-tracking",
        "kwp-product-management-feature-spec",
        "kwp-finance-close-management",
        "kwp-slack-slack-search",
        "kwp-design-system",
        "kwp-product-management-competitive-analysis",
        "kwp-sales-create-an-asset",
        "kwp-human-resources-people-analytics",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-data-data-validation",
        "kwp-data-data-context-extractor",
        "kwp-finance-financial-statements",
        "kwp-common-room-compose-outreach",
        "kwp-finance-audit-support",
        "kwp-data-interactive-dashboard-builder",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-marketing-content-creation"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: decision documentation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-product-management-stakeholder-comms\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-customer-support-escalation\", \"kwp-design-design-handoff\", \"kwp-marketing-campaign-planning\", \"kwp-customer-support-response-drafting\", \"kwp-product-management-metrics-tracking\", \"kwp-product-management-feature-spec\", \"kwp-finance-close-management\", \"kwp-slack-slack-search\", \"kwp-design-system\", \"kwp-product-management-competitive-analysis\", \"kwp-sales-create-an-asset\", \"kwp-human-resources-people-analytics\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-data-data-validation\", \"kwp-data-data-context-extractor\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0094",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Autonomous Optimization Architect agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-autonomous-optimization-architect"
      ],
      "distractor_skills": [
        "agency-xr-interface-architect",
        "agency-terminal-integration-specialist",
        "agency-infrastructure-maintainer",
        "agency-app-store-optimizer",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-report-distribution-agent",
        "agency-behavioral-nudge-engine",
        "agency-agents-orchestrator",
        "agency-brand-guardian",
        "agency-ux-architect",
        "agency-data-consolidation-agent",
        "agency-visionos-spatial-engineer",
        "agency-tool-evaluator"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Autonomous Optimization Architect agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-autonomous-optimization-architect\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-xr-interface-architect\", \"agency-terminal-integration-specialist\", \"agency-infrastructure-maintainer\", \"agency-app-store-optimizer\", \"agency-xr-immersive-developer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-report-distribution-agent\", \"agency-behavioral-nudge-engine\", \"agency-agents-orchestrator\", \"agency-brand-guardian\", \"agency-ux-architect\", \"agency-data-consolidation-agent\", \"agency-visionos-spatial-engineer\", \"agency-tool-evaluator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0095",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: any commercial agreement where you need clause-by-clause analysis against standard positions. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-legal-contract-review"
      ],
      "distractor_skills": [
        "kwp-legal-compliance",
        "kwp-operations-vendor-management",
        "kwp-design-system",
        "kwp-legal-meeting-briefing",
        "kwp-customer-support-customer-research",
        "kwp-customer-support-response-drafting",
        "kwp-data-data-validation",
        "kwp-legal-nda-triage",
        "kwp-legal-canned-responses",
        "kwp-slack-slack-search",
        "kwp-marketing-content-creation",
        "kwp-product-management-competitive-analysis",
        "kwp-cowork-plugin-customizer",
        "kwp-apollo-prospect",
        "kwp-sales-daily-briefing",
        "kwp-sales-account-research",
        "kwp-marketing-performance-analytics",
        "kwp-legal-legal-risk-assessment",
        "kwp-marketing-brand-voice",
        "kwp-human-resources-people-analytics",
        "kwp-data-data-context-extractor",
        "kwp-data-statistical-analysis",
        "kwp-sales-call-prep",
        "kwp-finance-variance-analysis",
        "kwp-finance-financial-statements",
        "kwp-finance-audit-support",
        "kwp-data-interactive-dashboard-builder",
        "kwp-customer-support-escalation"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: any commercial agreement where you need clause-by-clause analysis against standard positions. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-legal-contract-review\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-legal-compliance\", \"kwp-operations-vendor-management\", \"kwp-design-system\", \"kwp-legal-meeting-briefing\", \"kwp-customer-support-customer-research\", \"kwp-customer-support-response-drafting\", \"kwp-data-data-validation\", \"kwp-legal-nda-triage\", \"kwp-legal-canned-responses\", \"kwp-slack-slack-search\", \"kwp-marketing-content-creation\", \"kwp-product-management-competitive-analysis\", \"kwp-cowork-plugin-customizer\", \"kwp-apollo-prospect\", \"kwp-sales-daily-briefing\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0096",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: and you need to decide how to integrate the work - guides completion of development work by presenting structured options for merge. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "sp-finish-branch"
      ],
      "distractor_skills": [
        "sp-verification",
        "sp-executing-plans",
        "sp-receive-review",
        "sp-subagent-dev",
        "sp-tdd",
        "sp-debugging",
        "sp-request-review",
        "sp-core",
        "sp-brainstorming",
        "sp-writing-skills",
        "sp-git-worktrees",
        "sp-writing-plans",
        "agency-brand-guardian",
        "kwp-human-resources-interview-prep",
        "agency-visionos-spatial-engineer",
        "anthropic-claude-api",
        "agency-agents-orchestrator",
        "agency-xr-interface-architect",
        "agency-devops-automator",
        "kwp-legal-contract-review",
        "agency-ux-architect",
        "agency-rapid-prototyper",
        "agency-mobile-app-builder",
        "kwp-engineering-testing-strategy",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-frontend-developer",
        "agency-api-tester",
        "agency-terminal-integration-specialist",
        "ship",
        "agency-workflow-optimizer",
        "technical-writer",
        "kwp-engineering-code-review",
        "agency-senior-developer",
        "agency-app-store-optimizer",
        "kwp-marketing-content-creation",
        "kwp-operations-vendor-management",
        "kwp-design-accessibility-review",
        "kwp-customer-support-knowledge-management",
        "kwp-customer-support-customer-research",
        "agency-security-engineer",
        "frontend-expert",
        "agency-instagram-curator",
        "qa-test-expert",
        "kwp-engineering-system-design"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: and you need to decide how to integrate the work - guides completion of development work by presenting structured options for merge. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"sp-finish-branch\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"sp-verification\", \"sp-executing-plans\", \"sp-receive-review\", \"sp-subagent-dev\", \"sp-tdd\", \"sp-debugging\", \"sp-request-review\", \"sp-core\", \"sp-brainstorming\", \"sp-writing-skills\", \"sp-git-worktrees\", \"sp-writing-plans\", \"agency-brand-guardian\", \"kwp-human-resources-interview-prep\", \"agency-visionos-spatial-engineer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0097",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: duplicate a template doc. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "gws-recipe-doc-from-template"
      ],
      "distractor_skills": [
        "gws-recipe-save-email-to-doc",
        "gws-docs",
        "gws-recipe-events-from-sheet",
        "gws-drive",
        "gws-workspace",
        "gws-sheets",
        "gws-recipe-personalized-emails",
        "gws-chat",
        "gws-recipe-audit-sharing",
        "gws-calendar",
        "gws-gmail",
        "gws-workflows",
        "docs-tutor-setup",
        "technical-writer",
        "commit-to-issue",
        "kwp-engineering-documentation",
        "domain-commit",
        "calendar-daily-briefing",
        "anthropic-docx",
        "kwp-human-resources-interview-prep",
        "ship",
        "anthropic-claude-api",
        "skill-guide-generator",
        "nlm-video",
        "docx-template-engine",
        "gws-email-reply",
        "nlm-slides",
        "kwp-human-resources-people-analytics",
        "skill-seekers",
        "ppt-template-engine",
        "notebooklm-studio",
        "hf-repos"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: duplicate a template doc. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"gws-recipe-doc-from-template\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"gws-recipe-save-email-to-doc\", \"gws-docs\", \"gws-recipe-events-from-sheet\", \"gws-drive\", \"gws-workspace\", \"gws-sheets\", \"gws-recipe-personalized-emails\", \"gws-chat\", \"gws-recipe-audit-sharing\", \"gws-calendar\", \"gws-gmail\", \"gws-workflows\", \"docs-tutor-setup\", \"technical-writer\", \"commit-to-issue\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0098",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: ingest documents. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "cognee"
      ],
      "distractor_skills": [
        "alphaear-search",
        "mirofish-graph-explorer",
        "hf-papers",
        "scrapling",
        "knowledge-to-paper",
        "kwp-brand-voice-guideline-generation",
        "knowledge-daily-aggregator",
        "anthropic-pdf",
        "defuddle",
        "paper-archive",
        "transcribee",
        "hf-collections",
        "docs-tutor-setup",
        "pipeline-builder",
        "twitter-timeline-to-slack",
        "agency-data-consolidation-agent",
        "anthropic-canvas-design",
        "paper-auto-classifier",
        "md-enhance-publish",
        "alphaxiv-paper-lookup",
        "gws-docs",
        "agency-ai-engineer",
        "gws-workspace",
        "commit-to-issue",
        "hf-trending-intelligence",
        "gws-sheets",
        "alphaear-stock",
        "agency-feedback-synthesizer",
        "nlm-arxiv-slides",
        "anthropic-docx",
        "stock-csv-downloader",
        "agency-rapid-prototyper",
        "anthropic-xlsx",
        "anthropic-claude-api",
        "related-papers-scout",
        "trading-finviz-screener",
        "agency-lsp-index-engineer",
        "kwp-human-resources-people-analytics",
        "email-research-dispatcher",
        "semantic-guard",
        "gws-recipe-events-from-sheet",
        "gws-recipe-doc-from-template",
        "gws-drive",
        "agency-xr-immersive-developer",
        "agency-frontend-developer",
        "skill-optimizer",
        "agency-terminal-integration-specialist",
        "hf-hub"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: ingest documents. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"cognee\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"alphaear-search\", \"mirofish-graph-explorer\", \"hf-papers\", \"scrapling\", \"knowledge-to-paper\", \"kwp-brand-voice-guideline-generation\", \"knowledge-daily-aggregator\", \"anthropic-pdf\", \"defuddle\", \"paper-archive\", \"transcribee\", \"hf-collections\", \"docs-tutor-setup\", \"pipeline-builder\", \"twitter-timeline-to-slack\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0099",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Contact Research. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-common-room-contact-research"
      ],
      "distractor_skills": [
        "kwp-common-room-account-research",
        "kwp-common-room-compose-outreach",
        "kwp-data-data-context-extractor",
        "kwp-common-room-prospect",
        "kwp-apollo-enrich-lead",
        "kwp-sales-account-research",
        "kwp-common-room-call-prep",
        "kwp-sales-draft-outreach",
        "kwp-common-room-weekly-prep-brief",
        "kwp-slack-slack-search",
        "kwp-design-system",
        "kwp-slack-slack-messaging",
        "kwp-legal-canned-responses",
        "kwp-legal-meeting-briefing",
        "kwp-customer-support-customer-research",
        "kwp-marketing-content-creation",
        "kwp-finance-reconciliation",
        "kwp-cowork-plugin-customizer",
        "kwp-product-management-user-research-synthesis",
        "kwp-marketing-competitive-analysis",
        "kwp-human-resources-people-analytics",
        "kwp-customer-support-response-drafting",
        "kwp-sales-competitive-intelligence",
        "kwp-human-resources-org-planning",
        "kwp-data-statistical-analysis",
        "kwp-data-data-validation",
        "kwp-sales-call-prep",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-data-interactive-dashboard-builder",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-legal-compliance",
        "kwp-enterprise-search-search-strategy",
        "kwp-data-data-exploration"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Contact Research. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-common-room-contact-research\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-common-room-account-research\", \"kwp-common-room-compose-outreach\", \"kwp-data-data-context-extractor\", \"kwp-common-room-prospect\", \"kwp-apollo-enrich-lead\", \"kwp-sales-account-research\", \"kwp-common-room-call-prep\", \"kwp-sales-draft-outreach\", \"kwp-common-room-weekly-prep-brief\", \"kwp-slack-slack-search\", \"kwp-design-system\", \"kwp-slack-slack-messaging\", \"kwp-legal-canned-responses\", \"kwp-legal-meeting-briefing\", \"kwp-customer-support-customer-research\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0100",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: run a dependency sweep. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "dependency-auditor"
      ],
      "distractor_skills": [
        "ci-quality-gate",
        "compliance-governance",
        "security-expert",
        "qa-test-expert",
        "semantic-guard",
        "kwp-engineering-code-review",
        "sefo-governance",
        "skill-optimizer",
        "deep-review",
        "test-suite",
        "agency-security-engineer",
        "dependency-radar",
        "i18n-sync",
        "ux-expert",
        "code-review-all",
        "role-security-engineer",
        "compliance-gate",
        "github-sprint-digest",
        "kwp-design-accessibility-review",
        "kwp-operations-risk-assessment",
        "setup-doctor"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: run a dependency sweep. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"dependency-auditor\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"ci-quality-gate\", \"compliance-governance\", \"security-expert\", \"qa-test-expert\", \"semantic-guard\", \"kwp-engineering-code-review\", \"sefo-governance\", \"skill-optimizer\", \"deep-review\", \"test-suite\", \"agency-security-engineer\", \"dependency-radar\", \"i18n-sync\", \"ux-expert\", \"code-review-all\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0101",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Technical Documentation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-engineering-documentation"
      ],
      "distractor_skills": [
        "kwp-engineering-system-design",
        "kwp-design-user-research",
        "kwp-engineering-testing-strategy",
        "kwp-design-ux-writing",
        "kwp-operations-compliance-tracking",
        "kwp-design-design-system-management",
        "kwp-engineering-tech-debt",
        "kwp-operations-risk-assessment",
        "kwp-engineering-code-review",
        "kwp-design-accessibility-review",
        "kwp-engineering-incident-response",
        "kwp-design-design-critique",
        "technical-writer",
        "kwp-productivity-task-management",
        "agency-technical-writer",
        "kwp-human-resources-compensation-benchmarking"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Technical Documentation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-engineering-documentation\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-engineering-system-design\", \"kwp-design-user-research\", \"kwp-engineering-testing-strategy\", \"kwp-design-ux-writing\", \"kwp-operations-compliance-tracking\", \"kwp-design-design-system-management\", \"kwp-engineering-tech-debt\", \"kwp-operations-risk-assessment\", \"kwp-engineering-code-review\", \"kwp-design-accessibility-review\", \"kwp-engineering-incident-response\", \"kwp-design-design-critique\", \"technical-writer\", \"kwp-productivity-task-management\", \"agency-technical-writer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0102",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: UX Writing. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-design-ux-writing"
      ],
      "distractor_skills": [
        "kwp-engineering-testing-strategy",
        "kwp-engineering-system-design",
        "kwp-design-user-research",
        "kwp-design-accessibility-review",
        "kwp-engineering-tech-debt",
        "kwp-engineering-documentation",
        "kwp-operations-compliance-tracking",
        "kwp-operations-risk-assessment",
        "kwp-design-design-system-management",
        "kwp-design-design-critique",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-engineering-code-review",
        "kwp-engineering-incident-response",
        "kwp-productivity-task-management",
        "kwp-operations-resource-planning",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-human-resources-org-planning",
        "kwp-bio-research-scientific-problem-selection",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-design-design-handoff",
        "kwp-brand-voice-discover-brand",
        "kwp-brand-voice-brand-voice-enforcement",
        "kwp-design-system",
        "kwp-human-resources-people-analytics",
        "kwp-common-room-account-research",
        "kwp-human-resources-recruiting-pipeline"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: UX Writing. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-design-ux-writing\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-engineering-testing-strategy\", \"kwp-engineering-system-design\", \"kwp-design-user-research\", \"kwp-design-accessibility-review\", \"kwp-engineering-tech-debt\", \"kwp-engineering-documentation\", \"kwp-operations-compliance-tracking\", \"kwp-operations-risk-assessment\", \"kwp-design-design-system-management\", \"kwp-design-design-critique\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-engineering-code-review\", \"kwp-engineering-incident-response\", \"kwp-productivity-task-management\", \"kwp-operations-resource-planning\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0103",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 마켓 브레드스. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "trading-uptrend-analyzer"
      ],
      "distractor_skills": [
        "trading-market-breadth-analyzer",
        "trading-market-top-detector",
        "trading-breadth-chart-analyst",
        "trading-market-environment-analysis",
        "trading-sector-analyst",
        "trading-theme-detector",
        "trading-position-sizer",
        "trading-stanley-druckenmiller-investment",
        "trading-trade-hypothesis-ideator",
        "trading-options-strategy-advisor",
        "trading-finviz-screener",
        "trading-edge-candidate-agent",
        "trading-options-theta",
        "trading-technical-analyst",
        "trading-us-market-bubble-detector",
        "trading-us-stock-analysis",
        "trading-backtest-expert"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 마켓 브레드스. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"trading-uptrend-analyzer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"trading-market-breadth-analyzer\", \"trading-market-top-detector\", \"trading-breadth-chart-analyst\", \"trading-market-environment-analysis\", \"trading-sector-analyst\", \"trading-theme-detector\", \"trading-position-sizer\", \"trading-stanley-druckenmiller-investment\", \"trading-trade-hypothesis-ideator\", \"trading-options-strategy-advisor\", \"trading-finviz-screener\", \"trading-edge-candidate-agent\", \"trading-options-theta\", \"trading-technical-analyst\", \"trading-us-market-bubble-detector\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0104",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Tech Debt Management. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-engineering-tech-debt"
      ],
      "distractor_skills": [
        "kwp-engineering-code-review",
        "kwp-design-accessibility-review",
        "kwp-design-design-system-management",
        "kwp-engineering-system-design",
        "kwp-design-ux-writing",
        "kwp-operations-risk-assessment",
        "kwp-design-design-critique",
        "kwp-engineering-testing-strategy",
        "kwp-design-user-research",
        "kwp-engineering-documentation",
        "kwp-operations-compliance-tracking",
        "kwp-engineering-incident-response",
        "kwp-operations-vendor-management",
        "kwp-productivity-task-management",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-human-resources-interview-prep",
        "kwp-operations-change-management",
        "kwp-design-system",
        "agency-xr-immersive-developer",
        "kwp-operations-resource-planning",
        "agency-brand-guardian",
        "kwp-operations-process-optimization",
        "kwp-sales-daily-briefing",
        "kwp-cowork-plugin-customizer",
        "kwp-brand-voice-discover-brand"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Tech Debt Management. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-engineering-tech-debt\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-engineering-code-review\", \"kwp-design-accessibility-review\", \"kwp-design-design-system-management\", \"kwp-engineering-system-design\", \"kwp-design-ux-writing\", \"kwp-operations-risk-assessment\", \"kwp-design-design-critique\", \"kwp-engineering-testing-strategy\", \"kwp-design-user-research\", \"kwp-engineering-documentation\", \"kwp-operations-compliance-tracking\", \"kwp-engineering-incident-response\", \"kwp-operations-vendor-management\", \"kwp-productivity-task-management\", \"kwp-human-resources-org-planning\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0105",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: testing responsive design. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "playwright-runner"
      ],
      "distractor_skills": [
        "e2e-testing",
        "scrapling",
        "anthropic-webapp-testing",
        "kwp-engineering-testing-strategy",
        "qa-test-expert",
        "agent-browser",
        "agency-xr-immersive-developer",
        "hf-jobs",
        "agency-ux-researcher",
        "kwp-product-management-metrics-tracking",
        "anthropic-slack-gif-creator",
        "frontend-expert",
        "agency-accessibility-auditor",
        "agency-experiment-tracker",
        "kwp-human-resources-interview-prep",
        "kwp-design-user-research",
        "kwp-data-data-visualization",
        "ecc-eval-harness",
        "agency-ux-architect",
        "agency-devops-automator",
        "ecc-coding-standards",
        "ci-quality-gate",
        "agency-test-results-analyzer",
        "agency-api-tester",
        "e2e-overhaul",
        "ecc-continuous-learning",
        "commit-to-issue",
        "kwp-human-resources-org-planning",
        "kwp-data-data-validation",
        "agency-xr-cockpit-interaction-specialist",
        "anthropic-claude-api",
        "slack-agent",
        "kwp-finance-financial-statements",
        "kwp-design-system",
        "sre-devops-expert",
        "dependency-auditor",
        "anthropic-mcp-builder",
        "kwp-finance-audit-support"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: testing responsive design. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"playwright-runner\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"e2e-testing\", \"scrapling\", \"anthropic-webapp-testing\", \"kwp-engineering-testing-strategy\", \"qa-test-expert\", \"agent-browser\", \"agency-xr-immersive-developer\", \"hf-jobs\", \"agency-ux-researcher\", \"kwp-product-management-metrics-tracking\", \"anthropic-slack-gif-creator\", \"frontend-expert\", \"agency-accessibility-auditor\", \"agency-experiment-tracker\", \"kwp-human-resources-interview-prep\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0106",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: when you need channel-specific formatting. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-marketing-content-creation"
      ],
      "distractor_skills": [
        "kwp-marketing-performance-analytics",
        "kwp-marketing-campaign-planning",
        "kwp-marketing-brand-voice",
        "kwp-customer-support-response-drafting",
        "kwp-slack-slack-search",
        "kwp-customer-support-customer-research",
        "kwp-design-system",
        "kwp-common-room-contact-research",
        "kwp-marketing-competitive-analysis",
        "kwp-legal-contract-review",
        "kwp-common-room-compose-outreach",
        "kwp-legal-meeting-briefing",
        "kwp-sales-daily-briefing",
        "kwp-cowork-plugin-customizer",
        "kwp-sales-draft-outreach",
        "kwp-product-management-stakeholder-comms",
        "kwp-data-data-context-extractor",
        "kwp-sales-call-prep",
        "kwp-product-management-feature-spec",
        "kwp-customer-support-escalation",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-finance-close-management",
        "kwp-apollo-enrich-lead",
        "kwp-sales-create-an-asset",
        "kwp-sales-account-research",
        "kwp-finance-journal-entry-prep",
        "kwp-product-management-competitive-analysis",
        "kwp-finance-reconciliation",
        "kwp-common-room-weekly-prep-brief",
        "kwp-operations-resource-planning",
        "kwp-human-resources-people-analytics",
        "kwp-data-sql-queries",
        "kwp-common-room-account-research",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-org-planning",
        "kwp-data-statistical-analysis",
        "kwp-data-data-validation",
        "kwp-finance-variance-analysis"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: when you need channel-specific formatting. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-marketing-content-creation\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-marketing-performance-analytics\", \"kwp-marketing-campaign-planning\", \"kwp-marketing-brand-voice\", \"kwp-customer-support-response-drafting\", \"kwp-slack-slack-search\", \"kwp-customer-support-customer-research\", \"kwp-design-system\", \"kwp-common-room-contact-research\", \"kwp-marketing-competitive-analysis\", \"kwp-legal-contract-review\", \"kwp-common-room-compose-outreach\", \"kwp-legal-meeting-briefing\", \"kwp-sales-daily-briefing\", \"kwp-cowork-plugin-customizer\", \"kwp-sales-draft-outreach\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0107",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: data impact analysis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "role-data-scientist"
      ],
      "distractor_skills": [
        "role-hr",
        "role-developer",
        "role-finance",
        "role-ux-designer",
        "role-sales",
        "role-security-engineer",
        "role-trading-expert",
        "role-pm",
        "role-cso",
        "role-cto",
        "role-ceo",
        "deep-research-pipeline",
        "role-dispatcher",
        "pm-data-analytics",
        "paper-review",
        "executive-briefing",
        "kwp-operations-compliance-tracking",
        "commit-to-issue",
        "qa-test-expert",
        "intent-alignment-tracker",
        "agency-data-consolidation-agent",
        "kwp-human-resources-people-analytics",
        "agency-test-results-analyzer",
        "hf-evaluation",
        "trading-market-news-analyst",
        "agency-data-analytics-reporter",
        "agency-analytics-reporter",
        "kwp-design-accessibility-review",
        "sentence-polisher",
        "kwp-engineering-system-design",
        "agency-agents-orchestrator"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: data impact analysis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"role-data-scientist\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"role-hr\", \"role-developer\", \"role-finance\", \"role-ux-designer\", \"role-sales\", \"role-security-engineer\", \"role-trading-expert\", \"role-pm\", \"role-cso\", \"role-cto\", \"role-ceo\", \"deep-research-pipeline\", \"role-dispatcher\", \"pm-data-analytics\", \"paper-review\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0108",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: stop paperclip. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "paperclip-setup"
      ],
      "distractor_skills": [
        "paperclip-agents",
        "paperclip-tasks",
        "paperclip-control",
        "gws-workspace",
        "local-dev-runner",
        "slack-agent",
        "hf-hub",
        "sefo-orchestrator",
        "cursor-automations",
        "sefo-federation",
        "kwp-cowork-plugin-customizer",
        "agency-devops-automator",
        "agency-xr-cockpit-interaction-specialist",
        "hf-models",
        "agency-roster",
        "kwp-operations-change-management",
        "sefo-governance",
        "air-gap-orchestrator",
        "kwp-productivity-task-management",
        "kwp-operations-vendor-management",
        "kwp-cowork-cowork-plugin-customizer",
        "agency-infrastructure-maintainer",
        "gws-recipe-audit-sharing",
        "compliance-governance",
        "mirofish-financial-sim",
        "agency-xr-interface-architect",
        "anthropic-claude-api",
        "kwp-cowork-create-cowork-plugin",
        "sp-subagent-dev",
        "domain-commit",
        "agency-senior-project-manager",
        "agency-lsp-index-engineer",
        "agency-studio-producer",
        "kwp-design-system",
        "agency-xr-immersive-developer",
        "agency-frontend-developer",
        "ship",
        "public-apis",
        "workflow-parallel"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: stop paperclip. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"paperclip-setup\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"paperclip-agents\", \"paperclip-tasks\", \"paperclip-control\", \"gws-workspace\", \"local-dev-runner\", \"slack-agent\", \"hf-hub\", \"sefo-orchestrator\", \"cursor-automations\", \"sefo-federation\", \"kwp-cowork-plugin-customizer\", \"agency-devops-automator\", \"agency-xr-cockpit-interaction-specialist\", \"hf-models\", \"agency-roster\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0109",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 볼린저 밴드 갱신. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "tab-bollinger-refresh"
      ],
      "distractor_skills": [
        "tab-turtle-refresh",
        "tab-dualma-refresh",
        "tab-strategy-comparison",
        "tab-market-breadth",
        "tab-technical-analysis",
        "tab-llm-agents",
        "tab-report-generate",
        "tab-market-environment",
        "tab-fundamental-sync",
        "tab-hot-stock-discovery",
        "tab-screening"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 볼린저 밴드 갱신. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"tab-bollinger-refresh\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"tab-turtle-refresh\", \"tab-dualma-refresh\", \"tab-strategy-comparison\", \"tab-market-breadth\", \"tab-technical-analysis\", \"tab-llm-agents\", \"tab-report-generate\", \"tab-market-environment\", \"tab-fundamental-sync\", \"tab-hot-stock-discovery\", \"tab-screening\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0110",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Visionos Spatial Engineer agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-visionos-spatial-engineer"
      ],
      "distractor_skills": [
        "agency-macos-spatial-metal-engineer",
        "agency-xr-interface-architect",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-terminal-integration-specialist",
        "agency-brand-guardian",
        "agency-ux-architect",
        "agency-mobile-app-builder",
        "agency-senior-developer",
        "agency-app-store-optimizer",
        "agency-lsp-index-engineer",
        "agency-report-distribution-agent",
        "agency-behavioral-nudge-engine",
        "agency-rapid-prototyper",
        "agency-data-consolidation-agent",
        "agency-tool-evaluator",
        "agency-senior-project-manager",
        "agency-security-engineer",
        "agency-evidence-collector",
        "agency-test-results-analyzer",
        "agency-sales-data-extraction-agent",
        "agency-inclusive-visuals-specialist",
        "agency-api-tester",
        "agency-agents-orchestrator",
        "agency-workflow-optimizer",
        "agency-cultural-intelligence-strategist"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Visionos Spatial Engineer agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-visionos-spatial-engineer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-macos-spatial-metal-engineer\", \"agency-xr-interface-architect\", \"agency-xr-immersive-developer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-terminal-integration-specialist\", \"agency-brand-guardian\", \"agency-ux-architect\", \"agency-mobile-app-builder\", \"agency-senior-developer\", \"agency-app-store-optimizer\", \"agency-lsp-index-engineer\", \"agency-report-distribution-agent\", \"agency-behavioral-nudge-engine\", \"agency-rapid-prototyper\", \"agency-data-consolidation-agent\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0111",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: authenticate with Hugging Face. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "hf-hub"
      ],
      "distractor_skills": [
        "hf-collections",
        "hf-models",
        "hf-discussions",
        "hf-repos",
        "hf-buckets",
        "hf-spaces",
        "hf-endpoints",
        "hf-cli",
        "hf-papers",
        "gws-workspace",
        "hf-datasets",
        "hf-dataset-viewer",
        "hf-trending-intelligence",
        "hf-model-trainer",
        "hf-evaluation",
        "hf-model-arena",
        "hf-auto-finetune-planner",
        "hf-trackio",
        "hf-jobs",
        "paperclip-setup",
        "gws-gmail",
        "gws-chat",
        "gws-drive",
        "local-dev-runner",
        "gws-docs",
        "skill-optimizer",
        "kwp-operations-vendor-management",
        "gws-sheets",
        "alphaear-stock",
        "gws-recipe-audit-sharing",
        "anthropic-claude-api",
        "paperclip-agents",
        "agency-infrastructure-maintainer",
        "agency-devops-automator",
        "pm-product-strategy",
        "domain-commit",
        "gws-calendar",
        "gws-recipe-doc-from-template",
        "agency-xr-immersive-developer",
        "agency-studio-producer",
        "agency-terminal-integration-specialist",
        "kwp-operations-change-management",
        "kwp-engineering-code-review",
        "agency-project-shepherd",
        "kwp-sync",
        "kwp-design-accessibility-review",
        "sefo-federation",
        "agency-legal-compliance-checker",
        "trading-market-environment-analysis",
        "stock-csv-downloader"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: authenticate with Hugging Face. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"hf-hub\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"hf-collections\", \"hf-models\", \"hf-discussions\", \"hf-repos\", \"hf-buckets\", \"hf-spaces\", \"hf-endpoints\", \"hf-cli\", \"hf-papers\", \"gws-workspace\", \"hf-datasets\", \"hf-dataset-viewer\", \"hf-trending-intelligence\", \"hf-model-trainer\", \"hf-evaluation\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0112",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: paperclip agent create. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "paperclip-agents"
      ],
      "distractor_skills": [
        "paperclip-tasks",
        "paperclip-control",
        "paperclip-setup",
        "agency-roster",
        "cursor-automations",
        "autoskill-merger",
        "sefo-governance",
        "hf-hub",
        "sefo-orchestrator",
        "full-stack-planner",
        "autoskill-extractor",
        "autoskill-evolve",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "skill-optimizer",
        "ship",
        "workflow-parallel",
        "gws-workspace",
        "mission-control copy",
        "kwp-productivity-task-management",
        "kwp-human-resources-interview-prep",
        "prompt-transformer",
        "context-engineer",
        "agency-infrastructure-maintainer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: paperclip agent create. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"paperclip-agents\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"paperclip-tasks\", \"paperclip-control\", \"paperclip-setup\", \"agency-roster\", \"cursor-automations\", \"autoskill-merger\", \"sefo-governance\", \"hf-hub\", \"sefo-orchestrator\", \"full-stack-planner\", \"autoskill-extractor\", \"autoskill-evolve\", \"agency-xr-immersive-developer\", \"agency-xr-cockpit-interaction-specialist\", \"skill-optimizer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0113",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: request routing. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "feature-intake-router"
      ],
      "distractor_skills": [
        "prd-research-factory",
        "proposal-factory",
        "lead-lifecycle-manager",
        "incident-lifecycle-manager",
        "tech-support-brain",
        "docs-freshness-guardian",
        "auto-research-distribute",
        "alphaxiv-paper-lookup",
        "commit-to-issue",
        "problem-definition",
        "calendar-daily-briefing",
        "full-stack-planner",
        "jira-linear-adapter",
        "daily-stock-check",
        "auto-research",
        "sprint-orchestrator",
        "domain-commit",
        "simplify",
        "related-papers-scout",
        "dependency-radar",
        "scrapling",
        "feedback-meeting-scheduler",
        "daily-am-orchestrator",
        "technical-writer",
        "performance-profiler",
        "daily-strategy-post",
        "decision-router",
        "helm-validator",
        "code-review-all",
        "ci-quality-gate",
        "sales-weekly-dashboard",
        "sales-meeting-pipeline",
        "security-expert",
        "sre-devops-expert",
        "role-pm",
        "planning-meeting-sentinel"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: request routing. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"feature-intake-router\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"prd-research-factory\", \"proposal-factory\", \"lead-lifecycle-manager\", \"incident-lifecycle-manager\", \"tech-support-brain\", \"docs-freshness-guardian\", \"auto-research-distribute\", \"alphaxiv-paper-lookup\", \"commit-to-issue\", \"problem-definition\", \"calendar-daily-briefing\", \"full-stack-planner\", \"jira-linear-adapter\", \"daily-stock-check\", \"auto-research\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0114",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Knowledge Synthesis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-enterprise-search-knowledge-synthesis"
      ],
      "distractor_skills": [
        "kwp-enterprise-search-source-management",
        "kwp-enterprise-search-search-strategy",
        "kwp-slack-slack-search",
        "kwp-human-resources-interview-prep",
        "kwp-design-design-handoff",
        "kwp-human-resources-people-analytics",
        "kwp-data-data-context-extractor",
        "kwp-finance-variance-analysis",
        "kwp-operations-vendor-management",
        "kwp-finance-close-management",
        "kwp-apollo-enrich-lead",
        "kwp-customer-support-customer-research",
        "kwp-sales-create-an-asset",
        "kwp-sales-account-research",
        "kwp-product-management-competitive-analysis",
        "kwp-marketing-performance-analytics",
        "kwp-design-system",
        "kwp-apollo-prospect",
        "kwp-common-room-account-research",
        "kwp-sales-draft-outreach",
        "kwp-sales-competitive-intelligence",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-org-planning",
        "kwp-data-statistical-analysis",
        "kwp-data-data-validation",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-finance-financial-statements",
        "kwp-product-management-feature-spec",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-finance-audit-support",
        "kwp-data-interactive-dashboard-builder",
        "kwp-customer-support-escalation",
        "kwp-legal-meeting-briefing",
        "kwp-customer-support-knowledge-management",
        "kwp-sales-daily-briefing",
        "kwp-product-management-user-research-synthesis",
        "kwp-marketing-campaign-planning",
        "kwp-legal-legal-risk-assessment",
        "kwp-finance-journal-entry-prep",
        "kwp-finance-reconciliation",
        "kwp-cowork-plugin-customizer",
        "kwp-operations-resource-planning"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Knowledge Synthesis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-enterprise-search-knowledge-synthesis\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-enterprise-search-source-management\", \"kwp-enterprise-search-search-strategy\", \"kwp-slack-slack-search\", \"kwp-human-resources-interview-prep\", \"kwp-design-design-handoff\", \"kwp-human-resources-people-analytics\", \"kwp-data-data-context-extractor\", \"kwp-finance-variance-analysis\", \"kwp-operations-vendor-management\", \"kwp-finance-close-management\", \"kwp-apollo-enrich-lead\", \"kwp-customer-support-customer-research\", \"kwp-sales-create-an-asset\", \"kwp-sales-account-research\", \"kwp-product-management-competitive-analysis\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0115",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: battlecard. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "pm-go-to-market"
      ],
      "distractor_skills": [
        "pm-marketing-growth",
        "pm-market-research",
        "pm-product-strategy",
        "pm-toolkit",
        "pm-product-discovery",
        "pm-execution",
        "pm-data-analytics",
        "role-cso",
        "kwp-marketing-competitive-analysis",
        "kwp-marketing-campaign-planning",
        "agency-trend-researcher",
        "nlm-launch-kit",
        "kwp-human-resources-interview-prep",
        "agency-tiktok-strategist",
        "role-sales",
        "agency-xiaohongshu-specialist",
        "gws-recipe-doc-from-template",
        "e2e-testing",
        "agency-reddit-community-builder",
        "qa-test-expert",
        "kwp-product-management-competitive-analysis",
        "pm-sentinel",
        "kwp-engineering-testing-strategy",
        "anthropic-internal-comms",
        "trading-breadth-chart-analyst",
        "gws-recipe-events-from-sheet"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: battlecard. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"pm-go-to-market\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"pm-marketing-growth\", \"pm-market-research\", \"pm-product-strategy\", \"pm-toolkit\", \"pm-product-discovery\", \"pm-execution\", \"pm-data-analytics\", \"role-cso\", \"kwp-marketing-competitive-analysis\", \"kwp-marketing-campaign-planning\", \"agency-trend-researcher\", \"nlm-launch-kit\", \"kwp-human-resources-interview-prep\", \"agency-tiktok-strategist\", \"role-sales\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0116",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: reviewing code quality. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "ecc-coding-standards"
      ],
      "distractor_skills": [
        "ecc-search-first",
        "backend-expert",
        "ecc-configure",
        "ecc-verification-loop",
        "ecc-agentic-engineering",
        "kwp-engineering-tech-debt",
        "frontend-expert",
        "agency-frontend-developer",
        "code-review-all",
        "ecc-eval-harness",
        "agency-test-results-analyzer",
        "agency-api-tester",
        "deep-review",
        "kwp-engineering-code-review",
        "ecc-continuous-learning"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: reviewing code quality. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"ecc-coding-standards\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"ecc-search-first\", \"backend-expert\", \"ecc-configure\", \"ecc-verification-loop\", \"ecc-agentic-engineering\", \"kwp-engineering-tech-debt\", \"frontend-expert\", \"agency-frontend-developer\", \"code-review-all\", \"ecc-eval-harness\", \"agency-test-results-analyzer\", \"agency-api-tester\", \"deep-review\", \"kwp-engineering-code-review\", \"ecc-continuous-learning\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0117",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: auto-classify papers. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "paper-auto-classifier"
      ],
      "distractor_skills": [
        "hf-papers",
        "hf-trending-intelligence",
        "paper-review",
        "related-papers-scout",
        "email-research-dispatcher",
        "unified-intel-intake",
        "paper-archive",
        "auto-research",
        "knowledge-to-paper",
        "auto-research-distribute",
        "alphaear-search",
        "alphaxiv-paper-lookup",
        "pr-to-issue-linker",
        "cognee",
        "bespin-news-digest",
        "standup-digest",
        "tech-support-brain",
        "docs-tutor",
        "daily-am-orchestrator",
        "trading-finviz-screener",
        "notebooklm-research",
        "deep-review",
        "deep-research-pipeline",
        "github-sprint-digest",
        "role-ux-designer",
        "hf-hub",
        "workflow-miner",
        "issue-to-code",
        "trading-market-news-analyst",
        "twitter-timeline-to-slack",
        "weekly-status-report"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: auto-classify papers. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"paper-auto-classifier\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"hf-papers\", \"hf-trending-intelligence\", \"paper-review\", \"related-papers-scout\", \"email-research-dispatcher\", \"unified-intel-intake\", \"paper-archive\", \"auto-research\", \"knowledge-to-paper\", \"auto-research-distribute\", \"alphaear-search\", \"alphaxiv-paper-lookup\", \"pr-to-issue-linker\", \"cognee\", \"bespin-news-digest\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0118",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: any meeting where legal context. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-legal-meeting-briefing"
      ],
      "distractor_skills": [
        "kwp-human-resources-interview-prep",
        "kwp-sales-call-prep",
        "kwp-product-management-metrics-tracking",
        "kwp-finance-audit-support",
        "kwp-finance-close-management",
        "kwp-design-system",
        "kwp-customer-support-customer-research",
        "kwp-product-management-competitive-analysis",
        "kwp-common-room-account-research",
        "kwp-legal-legal-risk-assessment",
        "kwp-finance-journal-entry-prep",
        "kwp-marketing-competitive-analysis",
        "kwp-legal-contract-review",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-data-data-validation",
        "kwp-common-room-call-prep",
        "kwp-finance-variance-analysis",
        "kwp-finance-financial-statements",
        "kwp-operations-vendor-management",
        "kwp-product-management-feature-spec",
        "kwp-operations-compliance-tracking"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: any meeting where legal context. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-legal-meeting-briefing\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-human-resources-interview-prep\", \"kwp-sales-call-prep\", \"kwp-product-management-metrics-tracking\", \"kwp-finance-audit-support\", \"kwp-finance-close-management\", \"kwp-design-system\", \"kwp-customer-support-customer-research\", \"kwp-product-management-competitive-analysis\", \"kwp-common-room-account-research\", \"kwp-legal-legal-risk-assessment\", \"kwp-finance-journal-entry-prep\", \"kwp-marketing-competitive-analysis\", \"kwp-legal-contract-review\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-data-data-validation\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0119",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: UI 변경 요청 시. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "fe-pipeline"
      ],
      "distractor_skills": [
        "implement-screen",
        "design-review",
        "swagger-api-doc-generator",
        "figma-to-tds",
        "screen-description",
        "github-workflow-automation",
        "fsd-development",
        "technical-writer",
        "overlay-layout-patterns",
        "sre-devops-expert",
        "performance-profiler",
        "notion-docs-sync",
        "k8s-manifest-validator",
        "auto-research-distribute",
        "ci-quality-gate",
        "security-expert",
        "commit-to-issue",
        "feature-intake-router",
        "proposal-factory",
        "prd-research-factory",
        "alphaxiv-paper-lookup",
        "auto-research",
        "calendar-daily-briefing",
        "helm-validator",
        "anthropic-claude-api",
        "code-review-all",
        "simplify",
        "related-papers-scout",
        "problem-definition",
        "domain-commit",
        "automation-strategist",
        "eod-ship"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: UI 변경 요청 시. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"fe-pipeline\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"implement-screen\", \"design-review\", \"swagger-api-doc-generator\", \"figma-to-tds\", \"screen-description\", \"github-workflow-automation\", \"fsd-development\", \"technical-writer\", \"overlay-layout-patterns\", \"sre-devops-expert\", \"performance-profiler\", \"notion-docs-sync\", \"k8s-manifest-validator\", \"auto-research-distribute\", \"ci-quality-gate\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0120",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 주식 데이터 동기화. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "tab-stock-sync"
      ],
      "distractor_skills": [
        "tab-fundamental-sync",
        "tab-dualma-refresh",
        "tab-llm-agents",
        "weekly-stock-update",
        "tab-kiwoom",
        "tab-genai-features",
        "tab-strategy-comparison",
        "tab-report-generate",
        "tab-turtle-refresh",
        "tab-market-breadth",
        "tab-dashboard-summary",
        "tab-news-fetch",
        "tab-bollinger-refresh",
        "stock-csv-downloader",
        "tab-technical-analysis",
        "tab-hot-stock-discovery",
        "tab-screening",
        "tab-patterns",
        "tab-market-environment",
        "alphaear-stock",
        "tab-glossary",
        "daily-stock-check",
        "tab-event-detect",
        "tab-analysis-run",
        "trading-sector-analyst",
        "today",
        "alphaear-search",
        "trading-technical-analyst",
        "alphaear-news",
        "trading-us-stock-analysis",
        "trading-backtest-expert",
        "alphaear-deepear-lite",
        "hf-jobs",
        "trading-market-breadth-analyzer",
        "alphaear-predictor",
        "trading-us-market-bubble-detector",
        "trading-scenario-analyzer",
        "trading-breadth-chart-analyst",
        "morning",
        "alphaear-reporter",
        "hf-cli",
        "alphaear-signal-tracker",
        "trading-stanley-druckenmiller-investment",
        "kwp-sync",
        "trading-trade-hypothesis-ideator",
        "hf-trackio"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 주식 데이터 동기화. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"tab-stock-sync\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"tab-fundamental-sync\", \"tab-dualma-refresh\", \"tab-llm-agents\", \"weekly-stock-update\", \"tab-kiwoom\", \"tab-genai-features\", \"tab-strategy-comparison\", \"tab-report-generate\", \"tab-turtle-refresh\", \"tab-market-breadth\", \"tab-dashboard-summary\", \"tab-news-fetch\", \"tab-bollinger-refresh\", \"stock-csv-downloader\", \"tab-technical-analysis\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0121",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: produce one. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "anthropic-pdf"
      ],
      "distractor_skills": [
        "anthropic-pptx",
        "anthropic-docx",
        "anthropic-xlsx",
        "anthropic-canvas-design",
        "anthropic-doc-coauthoring",
        "anthropic-claude-api",
        "gws-docs",
        "anthropic-skill-creator",
        "gws-sheets",
        "anthropic-slack-gif-creator",
        "skill-optimizer",
        "anthropic-theme-factory",
        "anthropic-algorithmic-art",
        "anthropic-brand-guidelines",
        "cognee",
        "anthropic-mcp-builder",
        "gws-recipe-events-from-sheet",
        "anthropic-frontend-design",
        "md-to-slack-canvas",
        "kwp-sync",
        "office-template-enforcer",
        "prompt-transformer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: produce one. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-pdf\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"anthropic-pptx\", \"anthropic-docx\", \"anthropic-xlsx\", \"anthropic-canvas-design\", \"anthropic-doc-coauthoring\", \"anthropic-claude-api\", \"gws-docs\", \"anthropic-skill-creator\", \"gws-sheets\", \"anthropic-slack-gif-creator\", \"skill-optimizer\", \"anthropic-theme-factory\", \"anthropic-algorithmic-art\", \"anthropic-brand-guidelines\", \"cognee\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0122",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: test generation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "qa-test-expert"
      ],
      "distractor_skills": [
        "kwp-engineering-testing-strategy",
        "e2e-testing",
        "frontend-expert",
        "test-suite",
        "ci-quality-gate",
        "sre-devops-expert",
        "dependency-auditor",
        "anthropic-webapp-testing",
        "agency-test-results-analyzer",
        "kwp-human-resources-org-planning",
        "skill-optimizer",
        "db-expert",
        "e2e-overhaul",
        "deep-review",
        "commit-to-issue",
        "backend-expert",
        "ui-suite",
        "pr-review-captain",
        "kwp-engineering-tech-debt",
        "agency-api-tester",
        "technical-writer",
        "performance-profiler"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: test generation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"qa-test-expert\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-engineering-testing-strategy\", \"e2e-testing\", \"frontend-expert\", \"test-suite\", \"ci-quality-gate\", \"sre-devops-expert\", \"dependency-auditor\", \"anthropic-webapp-testing\", \"agency-test-results-analyzer\", \"kwp-human-resources-org-planning\", \"skill-optimizer\", \"db-expert\", \"e2e-overhaul\", \"deep-review\", \"commit-to-issue\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0123",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: any API discovery/lookup task. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "public-apis"
      ],
      "distractor_skills": [
        "daiso-mcp",
        "hf-papers",
        "gws-drive",
        "hf-models",
        "notebooklm-research",
        "gws-chat",
        "anthropic-claude-api",
        "notebooklm",
        "sefo-orchestrator",
        "ai-chief-of-staff",
        "anthropic-mcp-builder",
        "sefo-federation",
        "hf-endpoints",
        "related-papers-scout",
        "trading-finviz-screener",
        "gws-calendar",
        "paperclip-tasks",
        "alphaear-search",
        "gws-recipe-events-from-sheet",
        "paper-archive",
        "paperclip-setup",
        "hf-discussions",
        "gws-workspace",
        "mirofish-graph-explorer",
        "kwp-cowork-cowork-plugin-customizer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: any API discovery/lookup task. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"public-apis\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"daiso-mcp\", \"hf-papers\", \"gws-drive\", \"hf-models\", \"notebooklm-research\", \"gws-chat\", \"anthropic-claude-api\", \"notebooklm\", \"sefo-orchestrator\", \"ai-chief-of-staff\", \"anthropic-mcp-builder\", \"sefo-federation\", \"hf-endpoints\", \"related-papers-scout\", \"trading-finviz-screener\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0124",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: a holistic design review. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "design-architect"
      ],
      "distractor_skills": [
        "ux-expert",
        "ui-suite",
        "frontend-expert",
        "kwp-design-design-system-management",
        "kwp-design-design-critique",
        "prompt-transformer",
        "ci-quality-gate",
        "agency-test-results-analyzer",
        "skill-optimizer",
        "deep-review",
        "kwp-design-accessibility-review",
        "qa-test-expert",
        "role-ux-designer",
        "kwp-design-user-research",
        "design-review",
        "kwp-engineering-tech-debt",
        "e2e-testing",
        "agency-xr-immersive-developer",
        "agency-frontend-developer",
        "kwp-cowork-create-cowork-plugin",
        "db-expert",
        "agency-visionos-spatial-engineer",
        "hf-papers"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: a holistic design review. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"design-architect\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"ux-expert\", \"ui-suite\", \"frontend-expert\", \"kwp-design-design-system-management\", \"kwp-design-design-critique\", \"prompt-transformer\", \"ci-quality-gate\", \"agency-test-results-analyzer\", \"skill-optimizer\", \"deep-review\", \"kwp-design-accessibility-review\", \"qa-test-expert\", \"role-ux-designer\", \"kwp-design-user-research\", \"design-review\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0125",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: should I automate this. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "automation-strategist"
      ],
      "distractor_skills": [
        "pipeline-builder",
        "system-thinker",
        "technical-writer",
        "sre-devops-expert",
        "plans",
        "setup-doctor",
        "daily-strategy-post",
        "codebase-archaeologist",
        "ci-quality-gate",
        "daily-am-orchestrator",
        "auto-research-distribute",
        "context-engineer",
        "refactor-simulator",
        "onboarding-accelerator",
        "sprint-orchestrator",
        "eod-ship",
        "critical-review",
        "qa-test-expert",
        "problem-definition",
        "calendar-daily-briefing",
        "feedback-meeting-scheduler",
        "daily-stock-check",
        "performance-profiler",
        "auto-research",
        "executive-briefing"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: should I automate this. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"automation-strategist\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"pipeline-builder\", \"system-thinker\", \"technical-writer\", \"sre-devops-expert\", \"plans\", \"setup-doctor\", \"daily-strategy-post\", \"codebase-archaeologist\", \"ci-quality-gate\", \"daily-am-orchestrator\", \"auto-research-distribute\", \"context-engineer\", \"refactor-simulator\", \"onboarding-accelerator\", \"sprint-orchestrator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0126",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: local document store (daily_news DB). Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "alphaear-search"
      ],
      "distractor_skills": [
        "alphaear-news",
        "alphaear-stock",
        "alphaear-deepear-lite",
        "alphaear-sentiment",
        "alphaear-predictor",
        "stock-csv-downloader",
        "alphaear-signal-tracker",
        "alphaear-logic-visualizer",
        "alphaear-reporter",
        "daily-stock-check",
        "trading-market-news-analyst",
        "hf-papers",
        "trading-market-environment-analysis",
        "cognee",
        "tab-stock-sync",
        "weekly-stock-update"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: local document store (daily_news DB). Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-search\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"alphaear-news\", \"alphaear-stock\", \"alphaear-deepear-lite\", \"alphaear-sentiment\", \"alphaear-predictor\", \"stock-csv-downloader\", \"alphaear-signal-tracker\", \"alphaear-logic-visualizer\", \"alphaear-reporter\", \"daily-stock-check\", \"trading-market-news-analyst\", \"hf-papers\", \"trading-market-environment-analysis\", \"cognee\", \"tab-stock-sync\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0127",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: advance-decline health. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "trading-market-breadth-analyzer"
      ],
      "distractor_skills": [
        "trading-uptrend-analyzer",
        "trading-sector-analyst",
        "trading-theme-detector",
        "trading-breadth-chart-analyst",
        "trading-market-top-detector",
        "trading-market-environment-analysis",
        "trading-position-sizer",
        "trading-technical-analyst",
        "trading-trade-hypothesis-ideator",
        "trading-stanley-druckenmiller-investment",
        "trading-us-stock-analysis",
        "trading-us-market-bubble-detector",
        "trading-finviz-screener",
        "trading-options-strategy-advisor",
        "trading-options-theta"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: advance-decline health. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"trading-market-breadth-analyzer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"trading-uptrend-analyzer\", \"trading-sector-analyst\", \"trading-theme-detector\", \"trading-breadth-chart-analyst\", \"trading-market-top-detector\", \"trading-market-environment-analysis\", \"trading-position-sizer\", \"trading-technical-analyst\", \"trading-trade-hypothesis-ideator\", \"trading-stanley-druckenmiller-investment\", \"trading-us-stock-analysis\", \"trading-us-market-bubble-detector\", \"trading-finviz-screener\", \"trading-options-strategy-advisor\", \"trading-options-theta\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0128",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: fix my UI. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "ui-suite"
      ],
      "distractor_skills": [
        "test-suite",
        "deep-review",
        "simplify",
        "ship",
        "diagnose",
        "design-architect",
        "skill-optimizer",
        "ux-expert",
        "code-review-all",
        "deep-research-pipeline",
        "release-commander",
        "frontend-expert",
        "ecc-verification-loop",
        "release-ship",
        "ci-quality-gate",
        "auto-research",
        "critical-review",
        "qa-test-expert",
        "kwp-engineering-code-review",
        "commit-to-issue",
        "agency-security-engineer",
        "hf-papers",
        "i18n-sync",
        "dependency-auditor",
        "ai-workflow-integrator",
        "kwp-design-design-system-management",
        "backend-expert",
        "kwp-engineering-tech-debt",
        "domain-commit",
        "agency-frontend-developer",
        "service-health-doctor",
        "incident-to-improvement",
        "ecc-coding-standards",
        "agency-test-results-analyzer",
        "compliance-gate",
        "performance-profiler",
        "agency-backend-architect",
        "evals-skills",
        "agency-visionos-spatial-engineer",
        "kwp-sync",
        "hf-trending-intelligence"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: fix my UI. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"ui-suite\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"test-suite\", \"deep-review\", \"simplify\", \"ship\", \"diagnose\", \"design-architect\", \"skill-optimizer\", \"ux-expert\", \"code-review-all\", \"deep-research-pipeline\", \"release-commander\", \"frontend-expert\", \"ecc-verification-loop\", \"release-ship\", \"ci-quality-gate\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0129",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Ai Engineer agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-ai-engineer"
      ],
      "distractor_skills": [
        "agency-data-engineer",
        "agency-xr-immersive-developer",
        "agency-brand-guardian",
        "agency-terminal-integration-specialist",
        "agency-data-consolidation-agent",
        "agency-app-store-optimizer",
        "agency-tiktok-strategist",
        "agency-macos-spatial-metal-engineer",
        "agency-lsp-index-engineer",
        "agency-instagram-curator",
        "agency-xr-cockpit-interaction-specialist",
        "agency-api-tester",
        "agency-infrastructure-maintainer",
        "agency-xr-interface-architect",
        "agency-ux-architect",
        "agency-mobile-app-builder",
        "agency-growth-hacker",
        "agency-visionos-spatial-engineer",
        "agency-tool-evaluator",
        "agency-experiment-tracker",
        "agency-reddit-community-builder",
        "agency-test-results-analyzer",
        "agency-studio-operations",
        "agency-developer-advocate",
        "agency-workflow-optimizer",
        "agency-legal-compliance-checker",
        "agency-rapid-prototyper",
        "agency-devops-automator",
        "agency-senior-developer",
        "agency-trend-researcher",
        "agency-senior-project-manager",
        "agency-feedback-synthesizer",
        "agency-security-engineer",
        "agency-sales-data-extraction-agent",
        "agency-inclusive-visuals-specialist",
        "agency-finance-tracker",
        "agency-report-distribution-agent",
        "agency-twitter-engager",
        "agency-behavioral-nudge-engine",
        "agency-agents-orchestrator",
        "agency-visual-storyteller",
        "agency-xiaohongshu-specialist",
        "agency-project-shepherd",
        "agency-image-prompt-engineer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Ai Engineer agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-ai-engineer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-data-engineer\", \"agency-xr-immersive-developer\", \"agency-brand-guardian\", \"agency-terminal-integration-specialist\", \"agency-data-consolidation-agent\", \"agency-app-store-optimizer\", \"agency-tiktok-strategist\", \"agency-macos-spatial-metal-engineer\", \"agency-lsp-index-engineer\", \"agency-instagram-curator\", \"agency-xr-cockpit-interaction-specialist\", \"agency-api-tester\", \"agency-infrastructure-maintainer\", \"agency-xr-interface-architect\", \"agency-ux-architect\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0130",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: sprint-orchestrator. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "sprint-orchestrator"
      ],
      "distractor_skills": [
        "jira-linear-adapter",
        "github-sprint-digest",
        "feedback-meeting-scheduler",
        "standup-digest",
        "pr-to-issue-linker",
        "commit-to-issue",
        "calendar-daily-briefing",
        "weekly-status-report",
        "simplify",
        "domain-commit",
        "daily-am-orchestrator",
        "eod-ship",
        "onboarding-accelerator",
        "portfolio-report-generator",
        "context-engineer",
        "critical-review",
        "daily-stock-check",
        "plans",
        "daily-strategy-post",
        "release-ship",
        "setup-doctor",
        "technical-writer",
        "cursor-automations",
        "pipeline-builder",
        "automation-strategist",
        "auto-research-distribute",
        "auto-research",
        "ppt-template-engine",
        "executive-briefing",
        "security-expert",
        "code-review-all",
        "related-papers-scout",
        "problem-definition",
        "refactor-simulator",
        "codebase-archaeologist",
        "performance-profiler",
        "feature-intake-router",
        "decision-router",
        "issue-to-code",
        "ci-quality-gate"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: sprint-orchestrator. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"sprint-orchestrator\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"jira-linear-adapter\", \"github-sprint-digest\", \"feedback-meeting-scheduler\", \"standup-digest\", \"pr-to-issue-linker\", \"commit-to-issue\", \"calendar-daily-briefing\", \"weekly-status-report\", \"simplify\", \"domain-commit\", \"daily-am-orchestrator\", \"eod-ship\", \"onboarding-accelerator\", \"portfolio-report-generator\", \"context-engineer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0131",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 노션 페이지로 변환. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "md-to-notion"
      ],
      "distractor_skills": [
        "md-to-slack-canvas",
        "md-enhance-publish",
        "meeting-digest",
        "gws-docs",
        "docs-tutor-setup",
        "hf-buckets",
        "deep-research-pipeline",
        "notion-meeting-sync",
        "paper-review",
        "release-ship",
        "nlm-video",
        "github-sprint-digest",
        "gws-drive",
        "issue-to-code",
        "commit-to-issue",
        "cursor-sync",
        "ai-chief-of-staff",
        "transcribee",
        "weekly-status-report",
        "gws-sheets",
        "defuddle",
        "kwp-brand-voice-guideline-generation",
        "visual-explainer",
        "knowledge-daily-aggregator",
        "agency-data-consolidation-agent",
        "anthropic-pdf",
        "onboarding-accelerator",
        "nlm-slides",
        "nlm-arxiv-slides",
        "demo-forge"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 노션 페이지로 변환. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"md-to-notion\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"md-to-slack-canvas\", \"md-enhance-publish\", \"meeting-digest\", \"gws-docs\", \"docs-tutor-setup\", \"hf-buckets\", \"deep-research-pipeline\", \"notion-meeting-sync\", \"paper-review\", \"release-ship\", \"nlm-video\", \"github-sprint-digest\", \"gws-drive\", \"issue-to-code\", \"commit-to-issue\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0132",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 기술적 분석 계산. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "tab-technical-analysis"
      ],
      "distractor_skills": [
        "tab-turtle-refresh",
        "tab-bollinger-refresh",
        "tab-market-breadth",
        "tab-strategy-comparison",
        "tab-dualma-refresh",
        "tab-glossary",
        "tab-genai-features",
        "tab-stock-sync",
        "tab-screening",
        "tab-dashboard-summary",
        "tab-fundamental-sync",
        "tab-news-fetch",
        "tab-report-generate",
        "tab-market-environment",
        "tab-event-detect",
        "tab-llm-agents",
        "tab-patterns",
        "tab-analysis-run",
        "tab-hot-stock-discovery",
        "tab-kiwoom",
        "daily-stock-check",
        "morning-ship",
        "weekly-stock-update",
        "hf-jobs",
        "daily-strategy-post",
        "alphaear-predictor",
        "kwp-data-statistical-analysis",
        "gws-recipe-events-from-sheet",
        "alphaear-stock",
        "i18n-sync",
        "hf-cli",
        "agency-visionos-spatial-engineer",
        "agency-performance-benchmarker",
        "hf-spaces",
        "today",
        "sp-writing-plans",
        "gws-recipe-personalized-emails",
        "bespin-news-digest",
        "email-research-dispatcher",
        "autoskill-pipeline",
        "sp-subagent-dev",
        "sp-parallel-agents",
        "trading-market-environment-analysis",
        "stock-csv-downloader",
        "hf-endpoints",
        "agency-ux-architect"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 기술적 분석 계산. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"tab-technical-analysis\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"tab-turtle-refresh\", \"tab-bollinger-refresh\", \"tab-market-breadth\", \"tab-strategy-comparison\", \"tab-dualma-refresh\", \"tab-glossary\", \"tab-genai-features\", \"tab-stock-sync\", \"tab-screening\", \"tab-dashboard-summary\", \"tab-fundamental-sync\", \"tab-news-fetch\", \"tab-report-generate\", \"tab-market-environment\", \"tab-event-detect\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0133",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Accessibility Review. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-design-accessibility-review"
      ],
      "distractor_skills": [
        "kwp-design-user-research",
        "kwp-design-design-system-management",
        "kwp-engineering-code-review",
        "kwp-engineering-tech-debt",
        "kwp-design-ux-writing",
        "kwp-design-design-critique",
        "kwp-operations-compliance-tracking",
        "kwp-operations-risk-assessment",
        "kwp-engineering-system-design",
        "kwp-engineering-incident-response",
        "kwp-engineering-testing-strategy",
        "kwp-operations-vendor-management",
        "kwp-engineering-documentation",
        "kwp-productivity-task-management",
        "ux-expert",
        "kwp-operations-process-optimization",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-interview-prep",
        "kwp-operations-change-management",
        "agency-xr-immersive-developer",
        "agency-accessibility-auditor"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Accessibility Review. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-design-accessibility-review\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-design-user-research\", \"kwp-design-design-system-management\", \"kwp-engineering-code-review\", \"kwp-engineering-tech-debt\", \"kwp-design-ux-writing\", \"kwp-design-design-critique\", \"kwp-operations-compliance-tracking\", \"kwp-operations-risk-assessment\", \"kwp-engineering-system-design\", \"kwp-engineering-incident-response\", \"kwp-engineering-testing-strategy\", \"kwp-operations-vendor-management\", \"kwp-engineering-documentation\", \"kwp-productivity-task-management\", \"ux-expert\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0134",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: setting up real-time dashboards on HF Spaces. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "hf-trackio"
      ],
      "distractor_skills": [
        "hf-evaluation",
        "hf-endpoints",
        "hf-datasets",
        "hf-cli",
        "hf-jobs",
        "hf-dataset-viewer",
        "hf-models",
        "hf-collections",
        "hf-model-trainer",
        "hf-papers",
        "hf-spaces",
        "hf-buckets",
        "hf-hub",
        "hf-discussions",
        "hf-model-arena",
        "kwp-product-management-metrics-tracking",
        "alphaear-news",
        "hf-repos",
        "hf-auto-finetune-planner",
        "tab-report-generate",
        "tab-kiwoom",
        "anthropic-mcp-builder",
        "hf-trending-intelligence",
        "trading-us-stock-analysis"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: setting up real-time dashboards on HF Spaces. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"hf-trackio\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"hf-evaluation\", \"hf-endpoints\", \"hf-datasets\", \"hf-cli\", \"hf-jobs\", \"hf-dataset-viewer\", \"hf-models\", \"hf-collections\", \"hf-model-trainer\", \"hf-papers\", \"hf-spaces\", \"hf-buckets\", \"hf-hub\", \"hf-discussions\", \"hf-model-arena\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0135",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: references agency-xr-cockpit-interaction-specialist. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-xr-cockpit-interaction-specialist"
      ],
      "distractor_skills": [
        "agency-xr-interface-architect",
        "agency-xr-immersive-developer",
        "agency-terminal-integration-specialist",
        "agency-behavioral-nudge-engine",
        "agency-app-store-optimizer",
        "agency-ux-architect",
        "agency-report-distribution-agent",
        "agency-macos-spatial-metal-engineer",
        "agency-lsp-index-engineer",
        "agency-brand-guardian",
        "agency-api-tester",
        "agency-visionos-spatial-engineer",
        "agency-senior-developer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: references agency-xr-cockpit-interaction-specialist. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-xr-cockpit-interaction-specialist\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-xr-interface-architect\", \"agency-xr-immersive-developer\", \"agency-terminal-integration-specialist\", \"agency-behavioral-nudge-engine\", \"agency-app-store-optimizer\", \"agency-ux-architect\", \"agency-report-distribution-agent\", \"agency-macos-spatial-metal-engineer\", \"agency-lsp-index-engineer\", \"agency-brand-guardian\", \"agency-api-tester\", \"agency-visionos-spatial-engineer\", \"agency-senior-developer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0136",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: When people here say 'user' or 'customer', what exactly do they mean? Are there different types?. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-data-data-context-extractor"
      ],
      "distractor_skills": [
        "kwp-common-room-contact-research",
        "kwp-common-room-account-research",
        "kwp-human-resources-people-analytics",
        "kwp-slack-slack-search",
        "kwp-cowork-plugin-customizer",
        "kwp-common-room-compose-outreach",
        "kwp-data-statistical-analysis",
        "kwp-data-data-validation",
        "kwp-finance-financial-statements",
        "kwp-legal-canned-responses",
        "kwp-design-design-handoff",
        "kwp-apollo-enrich-lead",
        "kwp-design-system",
        "kwp-data-data-exploration",
        "kwp-customer-support-knowledge-management",
        "kwp-product-management-competitive-analysis",
        "kwp-finance-reconciliation",
        "kwp-common-room-weekly-prep-brief",
        "kwp-sales-create-an-asset",
        "kwp-sales-draft-outreach",
        "kwp-slack-slack-messaging",
        "kwp-common-room-call-prep",
        "kwp-cowork-create-plugin",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-finance-variance-analysis",
        "kwp-common-room-prospect",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-operations-process-optimization",
        "kwp-enterprise-search-knowledge-synthesis",
        "kwp-data-interactive-dashboard-builder",
        "kwp-finance-close-management",
        "kwp-enterprise-search-source-management",
        "kwp-brand-voice-guideline-generation",
        "kwp-legal-compliance",
        "kwp-enterprise-search-search-strategy",
        "kwp-customer-support-customer-research",
        "kwp-apollo-prospect",
        "kwp-sales-account-research",
        "kwp-product-management-user-research-synthesis",
        "kwp-marketing-performance-analytics",
        "kwp-bio-research-nextflow-development",
        "kwp-marketing-competitive-analysis"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: When people here say 'user' or 'customer', what exactly do they mean? Are there different types?. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-data-data-context-extractor\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-common-room-contact-research\", \"kwp-common-room-account-research\", \"kwp-human-resources-people-analytics\", \"kwp-slack-slack-search\", \"kwp-cowork-plugin-customizer\", \"kwp-common-room-compose-outreach\", \"kwp-data-statistical-analysis\", \"kwp-data-data-validation\", \"kwp-finance-financial-statements\", \"kwp-legal-canned-responses\", \"kwp-design-design-handoff\", \"kwp-apollo-enrich-lead\", \"kwp-design-system\", \"kwp-data-data-exploration\", \"kwp-customer-support-knowledge-management\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0137",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: UX Writing. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-design-ux-writing"
      ],
      "distractor_skills": [
        "kwp-engineering-testing-strategy",
        "kwp-engineering-system-design",
        "kwp-design-user-research",
        "kwp-design-accessibility-review",
        "kwp-engineering-tech-debt",
        "kwp-engineering-documentation",
        "kwp-operations-compliance-tracking",
        "kwp-operations-risk-assessment",
        "kwp-design-design-system-management",
        "kwp-design-design-critique",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-engineering-code-review",
        "kwp-engineering-incident-response",
        "kwp-productivity-task-management",
        "kwp-operations-resource-planning",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-human-resources-org-planning",
        "kwp-bio-research-scientific-problem-selection",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-design-design-handoff",
        "kwp-brand-voice-discover-brand",
        "kwp-brand-voice-brand-voice-enforcement",
        "kwp-design-system",
        "kwp-human-resources-people-analytics",
        "kwp-common-room-account-research",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-product-management-feature-spec",
        "kwp-common-room-compose-outreach",
        "kwp-common-room-call-prep",
        "kwp-sales-daily-briefing"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: UX Writing. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-design-ux-writing\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-engineering-testing-strategy\", \"kwp-engineering-system-design\", \"kwp-design-user-research\", \"kwp-design-accessibility-review\", \"kwp-engineering-tech-debt\", \"kwp-engineering-documentation\", \"kwp-operations-compliance-tracking\", \"kwp-operations-risk-assessment\", \"kwp-design-design-system-management\", \"kwp-design-design-critique\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-engineering-code-review\", \"kwp-engineering-incident-response\", \"kwp-productivity-task-management\", \"kwp-operations-resource-planning\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0138",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: generate daily report. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "tab-analysis-run"
      ],
      "distractor_skills": [
        "tab-patterns",
        "tab-genai-features",
        "tab-event-detect",
        "tab-report-generate",
        "tab-screening",
        "tab-llm-agents",
        "tab-strategy-comparison",
        "tab-market-breadth",
        "tab-turtle-refresh",
        "tab-dualma-refresh",
        "tab-kiwoom",
        "tab-glossary",
        "tab-dashboard-summary",
        "tab-news-fetch",
        "tab-fundamental-sync",
        "tab-market-environment",
        "tab-stock-sync",
        "tab-bollinger-refresh",
        "tab-technical-analysis",
        "tab-hot-stock-discovery",
        "alphaear-reporter"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: generate daily report. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"tab-analysis-run\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"tab-patterns\", \"tab-genai-features\", \"tab-event-detect\", \"tab-report-generate\", \"tab-screening\", \"tab-llm-agents\", \"tab-strategy-comparison\", \"tab-market-breadth\", \"tab-turtle-refresh\", \"tab-dualma-refresh\", \"tab-kiwoom\", \"tab-glossary\", \"tab-dashboard-summary\", \"tab-news-fetch\", \"tab-fundamental-sync\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0139",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: sequential workflow. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "workflow-sequential"
      ],
      "distractor_skills": [
        "workflow-parallel",
        "workflow-eval-opt",
        "ai-workflow-integrator",
        "sp-parallel-agents",
        "mission-control",
        "workflow-miner",
        "test-suite",
        "compliance-gate",
        "deep-research-pipeline",
        "auto-research",
        "kwp-operations-process-optimization",
        "sefo-orchestrator",
        "mission-control copy",
        "code-review-all",
        "sp-subagent-dev",
        "paperclip-tasks",
        "ralph-loop"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: sequential workflow. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"workflow-sequential\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"workflow-parallel\", \"workflow-eval-opt\", \"ai-workflow-integrator\", \"sp-parallel-agents\", \"mission-control\", \"workflow-miner\", \"test-suite\", \"compliance-gate\", \"deep-research-pipeline\", \"auto-research\", \"kwp-operations-process-optimization\", \"sefo-orchestrator\", \"mission-control copy\", \"code-review-all\", \"sp-subagent-dev\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0140",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Search Strategy. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-enterprise-search-search-strategy"
      ],
      "distractor_skills": [
        "kwp-enterprise-search-knowledge-synthesis",
        "kwp-enterprise-search-source-management",
        "kwp-slack-slack-search",
        "kwp-data-data-context-extractor",
        "kwp-human-resources-interview-prep",
        "kwp-cowork-plugin-customizer",
        "kwp-design-system",
        "kwp-common-room-contact-research",
        "kwp-data-statistical-analysis",
        "kwp-finance-variance-analysis",
        "kwp-legal-canned-responses",
        "kwp-operations-vendor-management",
        "kwp-legal-meeting-briefing",
        "kwp-customer-support-customer-research",
        "kwp-finance-close-management",
        "kwp-design-design-handoff",
        "kwp-apollo-enrich-lead",
        "kwp-slack-slack-messaging",
        "kwp-sales-account-research",
        "kwp-product-management-user-research-synthesis",
        "kwp-marketing-performance-analytics",
        "kwp-product-management-competitive-analysis",
        "kwp-finance-reconciliation",
        "kwp-apollo-prospect",
        "kwp-marketing-competitive-analysis",
        "kwp-marketing-brand-voice",
        "kwp-human-resources-people-analytics",
        "kwp-customer-support-response-drafting",
        "kwp-common-room-account-research",
        "kwp-apollo-sequence-load",
        "kwp-sales-draft-outreach",
        "kwp-sales-competitive-intelligence",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-org-planning",
        "kwp-data-data-validation",
        "kwp-finance-financial-statements",
        "kwp-common-room-prospect",
        "kwp-marketing-content-creation"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Search Strategy. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-enterprise-search-search-strategy\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-enterprise-search-knowledge-synthesis\", \"kwp-enterprise-search-source-management\", \"kwp-slack-slack-search\", \"kwp-data-data-context-extractor\", \"kwp-human-resources-interview-prep\", \"kwp-cowork-plugin-customizer\", \"kwp-design-system\", \"kwp-common-room-contact-research\", \"kwp-data-statistical-analysis\", \"kwp-finance-variance-analysis\", \"kwp-legal-canned-responses\", \"kwp-operations-vendor-management\", \"kwp-legal-meeting-briefing\", \"kwp-customer-support-customer-research\", \"kwp-finance-close-management\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0141",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: meeting sentinel. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "planning-meeting-sentinel"
      ],
      "distractor_skills": [
        "planning-weekly-pulse",
        "lead-lifecycle-manager",
        "sales-meeting-pipeline",
        "docs-freshness-guardian",
        "sales-weekly-dashboard",
        "dependency-radar",
        "feature-intake-router",
        "prd-research-factory",
        "notion-meeting-sync",
        "gws-calendar"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: meeting sentinel. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"planning-meeting-sentinel\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"planning-weekly-pulse\", \"lead-lifecycle-manager\", \"sales-meeting-pipeline\", \"docs-freshness-guardian\", \"sales-weekly-dashboard\", \"dependency-radar\", \"feature-intake-router\", \"prd-research-factory\", \"notion-meeting-sync\", \"gws-calendar\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0142",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: paper review. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "paper-review"
      ],
      "distractor_skills": [
        "nlm-arxiv-slides",
        "meeting-digest",
        "deep-research-pipeline",
        "nlm-slides",
        "alphaxiv-paper-lookup",
        "paper-auto-classifier",
        "paper-archive",
        "notion-meeting-sync",
        "nlm-video",
        "role-ux-designer",
        "hf-trending-intelligence",
        "nlm-strategy-hub",
        "role-cso",
        "auto-research-distribute",
        "role-sales",
        "role-data-scientist",
        "role-ceo",
        "role-finance",
        "role-developer",
        "ship",
        "github-sprint-digest",
        "helm-validator",
        "role-cto"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: paper review. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"paper-review\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"nlm-arxiv-slides\", \"meeting-digest\", \"deep-research-pipeline\", \"nlm-slides\", \"alphaxiv-paper-lookup\", \"paper-auto-classifier\", \"paper-archive\", \"notion-meeting-sync\", \"nlm-video\", \"role-ux-designer\", \"hf-trending-intelligence\", \"nlm-strategy-hub\", \"role-cso\", \"auto-research-distribute\", \"role-sales\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0143",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: defining acceptance criteria. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-product-management-feature-spec"
      ],
      "distractor_skills": [
        "kwp-marketing-campaign-planning",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-product-management-competitive-analysis",
        "kwp-human-resources-people-analytics",
        "kwp-product-management-metrics-tracking",
        "kwp-customer-support-escalation",
        "kwp-product-management-user-research-synthesis",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-org-planning",
        "kwp-finance-financial-statements",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-slack-slack-search",
        "kwp-finance-close-management",
        "kwp-design-system"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: defining acceptance criteria. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-product-management-feature-spec\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-marketing-campaign-planning\", \"kwp-operations-vendor-management\", \"kwp-human-resources-interview-prep\", \"kwp-product-management-competitive-analysis\", \"kwp-human-resources-people-analytics\", \"kwp-product-management-metrics-tracking\", \"kwp-customer-support-escalation\", \"kwp-product-management-user-research-synthesis\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-human-resources-org-planning\", \"kwp-finance-financial-statements\", \"kwp-operations-process-optimization\", \"kwp-operations-change-management\", \"kwp-slack-slack-search\", \"kwp-finance-close-management\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0144",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Helm 차트 린트. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "helm-validator"
      ],
      "distractor_skills": [
        "k8s-manifest-validator",
        "terraform-reviewer",
        "ci-quality-gate",
        "iac-review-agent",
        "infra-drift-detector",
        "alphaxiv-paper-lookup",
        "commit-to-issue",
        "sre-devops-expert",
        "code-review-all",
        "performance-profiler",
        "auto-research",
        "pipeline-builder",
        "refactor-simulator",
        "setup-doctor",
        "plans",
        "security-expert",
        "compliance-gate",
        "feedback-meeting-scheduler",
        "simplify",
        "related-papers-scout",
        "problem-definition",
        "infra-runbook-generator",
        "auto-research-distribute",
        "ecc-verification-loop",
        "domain-commit",
        "critical-review",
        "semantic-guard",
        "calendar-daily-briefing",
        "technical-writer",
        "daily-am-orchestrator",
        "full-stack-planner",
        "ppt-template-engine",
        "agency-devops-automator",
        "eod-ship",
        "intent-alignment-tracker",
        "cursor-automations",
        "jira-linear-adapter",
        "scrapling",
        "daily-stock-check",
        "onboarding-accelerator",
        "codebase-archaeologist",
        "kwp-engineering-code-review",
        "cursor-sync",
        "executive-briefing",
        "sefo-governance",
        "paper-review",
        "compliance-governance",
        "ai-quality-evaluator",
        "dependency-auditor"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Helm 차트 린트. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"helm-validator\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"k8s-manifest-validator\", \"terraform-reviewer\", \"ci-quality-gate\", \"iac-review-agent\", \"infra-drift-detector\", \"alphaxiv-paper-lookup\", \"commit-to-issue\", \"sre-devops-expert\", \"code-review-all\", \"performance-profiler\", \"auto-research\", \"pipeline-builder\", \"refactor-simulator\", \"setup-doctor\", \"plans\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0145",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: running evals to test a skill. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "anthropic-skill-creator"
      ],
      "distractor_skills": [
        "sp-writing-skills",
        "autoskill-merger",
        "kwp-sync",
        "skill-optimizer",
        "ecc-configure",
        "skill-seekers",
        "anthropic-algorithmic-art",
        "autoskill-extractor",
        "agency-roster",
        "anthropic-pptx",
        "skill-composer",
        "anthropic-slack-gif-creator",
        "prompt-transformer",
        "autoskill-judge",
        "anthropic-webapp-testing",
        "anthropic-pdf",
        "anthropic-canvas-design",
        "anthropic-claude-api",
        "anthropic-brand-guidelines",
        "anthropic-docx",
        "ecc-eval-harness",
        "kwp-data-data-context-extractor",
        "anthropic-mcp-builder",
        "kwp-cowork-create-plugin",
        "kwp-finance-financial-statements",
        "skill-guide-generator",
        "anthropic-frontend-design",
        "anthropic-xlsx",
        "anthropic-theme-factory",
        "gws-recipe-doc-from-template",
        "agency-xr-immersive-developer",
        "autoskill-evolve",
        "agency-test-results-analyzer",
        "anthropic-web-artifacts-builder",
        "agency-workflow-optimizer",
        "agency-developer-advocate",
        "agency-app-store-optimizer",
        "sp-core",
        "frontend-expert",
        "anthropic-internal-comms",
        "hf-evaluation",
        "anthropic-doc-coauthoring",
        "agency-infrastructure-maintainer",
        "ecc-continuous-learning",
        "kwp-cowork-plugin-customizer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: running evals to test a skill. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-skill-creator\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"sp-writing-skills\", \"autoskill-merger\", \"kwp-sync\", \"skill-optimizer\", \"ecc-configure\", \"skill-seekers\", \"anthropic-algorithmic-art\", \"autoskill-extractor\", \"agency-roster\", \"anthropic-pptx\", \"skill-composer\", \"anthropic-slack-gif-creator\", \"prompt-transformer\", \"autoskill-judge\", \"anthropic-webapp-testing\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0146",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: create an automation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "cursor-automations"
      ],
      "distractor_skills": [
        "commit-to-issue",
        "pipeline-builder",
        "domain-commit",
        "setup-doctor",
        "agent-browser",
        "eod-ship",
        "context-engineer",
        "jira-linear-adapter",
        "daily-am-orchestrator",
        "full-stack-planner",
        "calendar-daily-briefing",
        "technical-writer",
        "feedback-meeting-scheduler",
        "simplify",
        "k8s-manifest-validator",
        "daily-stock-check",
        "plans",
        "paperclip-agents",
        "scrapling",
        "sprint-orchestrator",
        "ci-quality-gate",
        "cursor-sync",
        "decision-router",
        "paperclip-setup",
        "code-review-all",
        "sre-devops-expert",
        "onboarding-accelerator"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: create an automation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"cursor-automations\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"commit-to-issue\", \"pipeline-builder\", \"domain-commit\", \"setup-doctor\", \"agent-browser\", \"eod-ship\", \"context-engineer\", \"jira-linear-adapter\", \"daily-am-orchestrator\", \"full-stack-planner\", \"calendar-daily-briefing\", \"technical-writer\", \"feedback-meeting-scheduler\", \"simplify\", \"k8s-manifest-validator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0147",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: unified-intel-intake. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "unified-intel-intake"
      ],
      "distractor_skills": [
        "twitter-timeline-to-slack",
        "paper-auto-classifier",
        "bespin-news-digest",
        "iac-review-agent",
        "x-to-slack",
        "auto-research-distribute",
        "daily-am-orchestrator",
        "daily-strategy-post",
        "auto-research",
        "knowledge-to-paper",
        "hf-trending-intelligence",
        "autoskill-pipeline",
        "decision-router",
        "release-commander",
        "release-ship",
        "eod-ship",
        "sprint-orchestrator",
        "morning-ship",
        "role-dispatcher",
        "weekly-status-report",
        "agency-agents-orchestrator",
        "portfolio-report-generator",
        "agency-lsp-index-engineer",
        "deep-research-pipeline",
        "github-sprint-digest",
        "ship",
        "agency-workflow-optimizer",
        "kwp-engineering-code-review",
        "alphaxiv-paper-lookup",
        "standup-digest",
        "kwp-design-accessibility-review",
        "related-papers-scout",
        "agency-content-creator",
        "knowledge-daily-aggregator",
        "paper-review",
        "kwp-design-design-system-management",
        "kwp-design-design-critique",
        "intent-alignment-tracker",
        "agency-data-consolidation-agent",
        "pr-to-issue-linker",
        "plans",
        "email-research-dispatcher",
        "deep-review",
        "ui-suite",
        "agency-xr-interface-architect",
        "agency-studio-operations",
        "agency-devops-automator",
        "agency-developer-advocate",
        "ai-quality-evaluator"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: unified-intel-intake. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"unified-intel-intake\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"twitter-timeline-to-slack\", \"paper-auto-classifier\", \"bespin-news-digest\", \"iac-review-agent\", \"x-to-slack\", \"auto-research-distribute\", \"daily-am-orchestrator\", \"daily-strategy-post\", \"auto-research\", \"knowledge-to-paper\", \"hf-trending-intelligence\", \"autoskill-pipeline\", \"decision-router\", \"release-commander\", \"release-ship\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0148",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: list bucket contents. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "hf-buckets"
      ],
      "distractor_skills": [
        "hf-repos",
        "hf-discussions",
        "hf-models",
        "hf-collections",
        "hf-endpoints",
        "hf-cli",
        "hf-hub",
        "hf-spaces",
        "hf-papers",
        "hf-dataset-viewer",
        "gws-drive",
        "hf-datasets",
        "gws-docs",
        "gws-sheets",
        "hf-auto-finetune-planner",
        "gws-chat",
        "hf-model-trainer",
        "hf-evaluation",
        "hf-model-arena",
        "hf-trackio",
        "hf-jobs"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: list bucket contents. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"hf-buckets\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"hf-repos\", \"hf-discussions\", \"hf-models\", \"hf-collections\", \"hf-endpoints\", \"hf-cli\", \"hf-hub\", \"hf-spaces\", \"hf-papers\", \"hf-dataset-viewer\", \"gws-drive\", \"hf-datasets\", \"gws-docs\", \"gws-sheets\", \"hf-auto-finetune-planner\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0149",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: write paper from today's data. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "knowledge-to-paper"
      ],
      "distractor_skills": [
        "knowledge-daily-aggregator",
        "auto-research",
        "auto-research-distribute",
        "daily-strategy-post",
        "hf-trending-intelligence",
        "paper-auto-classifier",
        "cognee",
        "bespin-news-digest",
        "md-enhance-publish",
        "notebooklm-research",
        "email-research-dispatcher",
        "morning",
        "trading-edge-candidate-agent",
        "morning-ship",
        "nlm-slides",
        "daily-am-orchestrator",
        "notion-meeting-sync",
        "unified-intel-intake",
        "github-sprint-digest",
        "mirofish-graph-explorer",
        "ppt-template-engine",
        "twitter-timeline-to-slack",
        "mirofish"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: write paper from today's data. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"knowledge-to-paper\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"knowledge-daily-aggregator\", \"auto-research\", \"auto-research-distribute\", \"daily-strategy-post\", \"hf-trending-intelligence\", \"paper-auto-classifier\", \"cognee\", \"bespin-news-digest\", \"md-enhance-publish\", \"notebooklm-research\", \"email-research-dispatcher\", \"morning\", \"trading-edge-candidate-agent\", \"morning-ship\", \"nlm-slides\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0150",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: ideal customer profile. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "pm-go-to-market"
      ],
      "distractor_skills": [
        "pm-marketing-growth",
        "pm-market-research",
        "pm-product-strategy",
        "pm-toolkit",
        "pm-product-discovery",
        "pm-execution",
        "pm-data-analytics",
        "role-cso",
        "kwp-marketing-competitive-analysis",
        "kwp-marketing-campaign-planning",
        "agency-trend-researcher",
        "nlm-launch-kit",
        "kwp-human-resources-interview-prep",
        "agency-tiktok-strategist",
        "role-sales",
        "agency-xiaohongshu-specialist",
        "gws-recipe-doc-from-template",
        "e2e-testing",
        "agency-reddit-community-builder",
        "qa-test-expert",
        "kwp-product-management-competitive-analysis",
        "pm-sentinel",
        "kwp-engineering-testing-strategy",
        "anthropic-internal-comms",
        "trading-breadth-chart-analyst",
        "gws-recipe-events-from-sheet",
        "gws-drive",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-report-distribution-agent",
        "agency-twitter-engager",
        "gws-docs",
        "agency-app-store-optimizer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: ideal customer profile. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"pm-go-to-market\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"pm-marketing-growth\", \"pm-market-research\", \"pm-product-strategy\", \"pm-toolkit\", \"pm-product-discovery\", \"pm-execution\", \"pm-data-analytics\", \"role-cso\", \"kwp-marketing-competitive-analysis\", \"kwp-marketing-campaign-planning\", \"agency-trend-researcher\", \"nlm-launch-kit\", \"kwp-human-resources-interview-prep\", \"agency-tiktok-strategist\", \"role-sales\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0151",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: update stock prices from investing.com. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "stock-csv-downloader"
      ],
      "distractor_skills": [
        "weekly-stock-update",
        "daily-stock-check",
        "alphaear-stock",
        "tab-stock-sync",
        "trading-sector-analyst",
        "alphaear-search",
        "alphaear-news",
        "alphaear-deepear-lite",
        "trading-market-environment-analysis",
        "trading-technical-analyst",
        "tab-kiwoom",
        "today",
        "tab-report-generate",
        "trading-us-stock-analysis",
        "trading-theme-detector",
        "alphaear-reporter",
        "trading-breadth-chart-analyst",
        "gws-recipe-events-from-sheet",
        "weekly-status-report",
        "trading-backtest-expert",
        "trading-trade-hypothesis-ideator",
        "trading-market-breadth-analyzer",
        "tab-llm-agents",
        "trading-stanley-druckenmiller-investment",
        "kwp-sync",
        "alphaear-sentiment",
        "kwp-human-resources-people-analytics",
        "agency-data-consolidation-agent",
        "tab-fundamental-sync",
        "gws-drive",
        "trading-market-top-detector",
        "tab-market-environment",
        "calendar-daily-briefing",
        "alphaear-predictor",
        "trading-uptrend-analyzer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: update stock prices from investing.com. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"stock-csv-downloader\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"weekly-stock-update\", \"daily-stock-check\", \"alphaear-stock\", \"tab-stock-sync\", \"trading-sector-analyst\", \"alphaear-search\", \"alphaear-news\", \"alphaear-deepear-lite\", \"trading-market-environment-analysis\", \"trading-technical-analyst\", \"tab-kiwoom\", \"today\", \"tab-report-generate\", \"trading-us-stock-analysis\", \"trading-theme-detector\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0152",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Call Prep. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-common-room-call-prep"
      ],
      "distractor_skills": [
        "kwp-common-room-account-research",
        "kwp-common-room-weekly-prep-brief",
        "kwp-common-room-compose-outreach",
        "kwp-common-room-prospect",
        "kwp-common-room-contact-research",
        "kwp-sales-call-prep",
        "kwp-legal-meeting-briefing",
        "kwp-human-resources-employee-handbook",
        "kwp-data-data-context-extractor",
        "kwp-sales-draft-outreach",
        "kwp-slack-slack-messaging",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-apollo-enrich-lead",
        "kwp-slack-slack-search",
        "kwp-design-system",
        "kwp-apollo-prospect",
        "kwp-sales-daily-briefing",
        "kwp-sales-account-research",
        "kwp-finance-journal-entry-prep"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Call Prep. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-common-room-call-prep\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-common-room-account-research\", \"kwp-common-room-weekly-prep-brief\", \"kwp-common-room-compose-outreach\", \"kwp-common-room-prospect\", \"kwp-common-room-contact-research\", \"kwp-sales-call-prep\", \"kwp-legal-meeting-briefing\", \"kwp-human-resources-employee-handbook\", \"kwp-data-data-context-extractor\", \"kwp-sales-draft-outreach\", \"kwp-slack-slack-messaging\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-operations-vendor-management\", \"kwp-human-resources-interview-prep\", \"kwp-apollo-enrich-lead\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0153",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Risk Assessment. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-operations-risk-assessment"
      ],
      "distractor_skills": [
        "kwp-engineering-system-design",
        "kwp-engineering-tech-debt",
        "kwp-engineering-code-review",
        "kwp-engineering-testing-strategy",
        "kwp-design-ux-writing",
        "kwp-design-accessibility-review",
        "kwp-engineering-incident-response",
        "kwp-design-user-research",
        "kwp-design-design-system-management",
        "kwp-design-design-critique",
        "kwp-engineering-documentation",
        "kwp-operations-compliance-tracking",
        "kwp-operations-vendor-management",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-operations-process-optimization",
        "kwp-legal-legal-risk-assessment",
        "kwp-human-resources-people-analytics",
        "agency-security-engineer",
        "kwp-operations-change-management",
        "kwp-human-resources-interview-prep",
        "kwp-productivity-task-management",
        "kwp-operations-resource-planning",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-org-planning"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Risk Assessment. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-operations-risk-assessment\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-engineering-system-design\", \"kwp-engineering-tech-debt\", \"kwp-engineering-code-review\", \"kwp-engineering-testing-strategy\", \"kwp-design-ux-writing\", \"kwp-design-accessibility-review\", \"kwp-engineering-incident-response\", \"kwp-design-user-research\", \"kwp-design-design-system-management\", \"kwp-design-design-critique\", \"kwp-engineering-documentation\", \"kwp-operations-compliance-tracking\", \"kwp-operations-vendor-management\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-operations-process-optimization\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0154",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: requirements to PRD. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "prd-research-factory"
      ],
      "distractor_skills": [
        "proposal-factory",
        "feature-intake-router",
        "plans",
        "auto-research",
        "alphaxiv-paper-lookup",
        "full-stack-planner",
        "sales-meeting-pipeline",
        "role-cso",
        "critical-review",
        "commit-to-issue",
        "problem-definition",
        "daily-am-orchestrator",
        "daily-strategy-post",
        "nlm-launch-kit",
        "kwp-product-management-feature-spec",
        "scrapling",
        "lead-lifecycle-manager",
        "daily-stock-check",
        "ppt-template-engine",
        "executive-briefing",
        "decision-router",
        "eod-ship",
        "ci-quality-gate",
        "simplify",
        "sales-weekly-dashboard",
        "deep-research-pipeline",
        "feedback-meeting-scheduler"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: requirements to PRD. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"prd-research-factory\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"proposal-factory\", \"feature-intake-router\", \"plans\", \"auto-research\", \"alphaxiv-paper-lookup\", \"full-stack-planner\", \"sales-meeting-pipeline\", \"role-cso\", \"critical-review\", \"commit-to-issue\", \"problem-definition\", \"daily-am-orchestrator\", \"daily-strategy-post\", \"nlm-launch-kit\", \"kwp-product-management-feature-spec\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0155",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: when deciding whether an NDA needs full counsel review. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-legal-nda-triage"
      ],
      "distractor_skills": [
        "kwp-legal-legal-risk-assessment",
        "kwp-customer-support-escalation",
        "kwp-customer-support-ticket-triage",
        "kwp-legal-compliance",
        "kwp-data-data-exploration",
        "kwp-human-resources-people-analytics",
        "kwp-legal-contract-review",
        "kwp-operations-vendor-management",
        "kwp-legal-canned-responses",
        "kwp-design-design-handoff",
        "kwp-customer-support-knowledge-management",
        "kwp-slack-slack-search",
        "kwp-design-system",
        "kwp-product-management-competitive-analysis",
        "kwp-sales-competitive-intelligence",
        "kwp-data-data-context-extractor",
        "kwp-product-management-feature-spec",
        "kwp-operations-process-optimization",
        "kwp-human-resources-interview-prep",
        "kwp-legal-meeting-briefing",
        "kwp-finance-close-management",
        "kwp-enterprise-search-source-management",
        "kwp-apollo-enrich-lead",
        "kwp-sales-daily-briefing",
        "kwp-sales-create-an-asset",
        "kwp-sales-account-research",
        "kwp-marketing-performance-analytics",
        "kwp-finance-reconciliation",
        "kwp-cowork-plugin-customizer",
        "kwp-apollo-prospect",
        "kwp-operations-resource-planning",
        "kwp-marketing-competitive-analysis",
        "kwp-marketing-brand-voice",
        "kwp-customer-support-response-drafting",
        "kwp-cowork-create-plugin",
        "kwp-common-room-contact-research",
        "kwp-common-room-account-research"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: when deciding whether an NDA needs full counsel review. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-legal-nda-triage\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-legal-legal-risk-assessment\", \"kwp-customer-support-escalation\", \"kwp-customer-support-ticket-triage\", \"kwp-legal-compliance\", \"kwp-data-data-exploration\", \"kwp-human-resources-people-analytics\", \"kwp-legal-contract-review\", \"kwp-operations-vendor-management\", \"kwp-legal-canned-responses\", \"kwp-design-design-handoff\", \"kwp-customer-support-knowledge-management\", \"kwp-slack-slack-search\", \"kwp-design-system\", \"kwp-product-management-competitive-analysis\", \"kwp-sales-competitive-intelligence\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0156",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Requirements for Outputs. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "anthropic-xlsx"
      ],
      "distractor_skills": [
        "anthropic-pptx",
        "anthropic-docx",
        "anthropic-pdf",
        "anthropic-canvas-design",
        "anthropic-claude-api",
        "anthropic-algorithmic-art",
        "anthropic-web-artifacts-builder",
        "anthropic-doc-coauthoring",
        "anthropic-theme-factory",
        "anthropic-slack-gif-creator",
        "anthropic-skill-creator",
        "anthropic-webapp-testing",
        "anthropic-frontend-design",
        "anthropic-internal-comms",
        "anthropic-brand-guidelines",
        "gws-sheets",
        "anthropic-mcp-builder",
        "cognee",
        "kwp-data-interactive-dashboard-builder"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Requirements for Outputs. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-xlsx\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"anthropic-pptx\", \"anthropic-docx\", \"anthropic-pdf\", \"anthropic-canvas-design\", \"anthropic-claude-api\", \"anthropic-algorithmic-art\", \"anthropic-web-artifacts-builder\", \"anthropic-doc-coauthoring\", \"anthropic-theme-factory\", \"anthropic-slack-gif-creator\", \"anthropic-skill-creator\", \"anthropic-webapp-testing\", \"anthropic-frontend-design\", \"anthropic-internal-comms\", \"anthropic-brand-guidelines\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0157",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: escalations. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-customer-support-response-drafting"
      ],
      "distractor_skills": [
        "kwp-design-system",
        "kwp-legal-canned-responses",
        "kwp-slack-slack-search",
        "kwp-data-interactive-dashboard-builder",
        "kwp-customer-support-escalation",
        "kwp-legal-compliance",
        "kwp-customer-support-customer-research",
        "kwp-product-management-competitive-analysis",
        "kwp-marketing-content-creation",
        "kwp-product-management-user-research-synthesis"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: escalations. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-customer-support-response-drafting\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-design-system\", \"kwp-legal-canned-responses\", \"kwp-slack-slack-search\", \"kwp-data-interactive-dashboard-builder\", \"kwp-customer-support-escalation\", \"kwp-legal-compliance\", \"kwp-customer-support-customer-research\", \"kwp-product-management-competitive-analysis\", \"kwp-marketing-content-creation\", \"kwp-product-management-user-research-synthesis\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0158",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: K8s 매니페스트 검증. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "k8s-manifest-validator"
      ],
      "distractor_skills": [
        "helm-validator",
        "terraform-reviewer",
        "ci-quality-gate",
        "iac-review-agent",
        "sre-devops-expert",
        "infra-drift-detector",
        "performance-profiler",
        "code-review-all",
        "pipeline-builder",
        "setup-doctor",
        "security-expert",
        "commit-to-issue",
        "compliance-gate",
        "technical-writer",
        "backend-expert",
        "auto-research-distribute",
        "refactor-simulator",
        "alphaxiv-paper-lookup",
        "simplify",
        "problem-definition",
        "infra-runbook-generator",
        "cursor-automations",
        "agency-devops-automator",
        "daily-stock-check",
        "auto-research",
        "onboarding-accelerator",
        "domain-commit",
        "semantic-guard",
        "related-papers-scout",
        "agency-backend-architect",
        "calendar-daily-briefing",
        "jira-linear-adapter",
        "daily-am-orchestrator",
        "codebase-archaeologist",
        "plans",
        "daily-strategy-post",
        "cursor-sync",
        "critical-review",
        "executive-briefing",
        "eod-ship",
        "ux-expert"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: K8s 매니페스트 검증. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"k8s-manifest-validator\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"helm-validator\", \"terraform-reviewer\", \"ci-quality-gate\", \"iac-review-agent\", \"sre-devops-expert\", \"infra-drift-detector\", \"performance-profiler\", \"code-review-all\", \"pipeline-builder\", \"setup-doctor\", \"security-expert\", \"commit-to-issue\", \"compliance-gate\", \"technical-writer\", \"backend-expert\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0159",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: tab-stock-sync. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "tab-stock-sync"
      ],
      "distractor_skills": [
        "tab-fundamental-sync",
        "tab-dualma-refresh",
        "tab-llm-agents",
        "weekly-stock-update",
        "tab-kiwoom",
        "tab-genai-features",
        "tab-strategy-comparison",
        "tab-report-generate",
        "tab-turtle-refresh",
        "tab-market-breadth",
        "tab-dashboard-summary",
        "tab-news-fetch",
        "tab-bollinger-refresh",
        "stock-csv-downloader",
        "tab-technical-analysis",
        "tab-hot-stock-discovery",
        "tab-screening",
        "tab-patterns",
        "tab-market-environment",
        "alphaear-stock",
        "tab-glossary",
        "daily-stock-check",
        "tab-event-detect",
        "tab-analysis-run",
        "trading-sector-analyst",
        "today",
        "alphaear-search",
        "trading-technical-analyst",
        "alphaear-news",
        "trading-us-stock-analysis",
        "trading-backtest-expert",
        "alphaear-deepear-lite",
        "hf-jobs",
        "trading-market-breadth-analyzer",
        "alphaear-predictor",
        "trading-us-market-bubble-detector",
        "trading-scenario-analyzer",
        "trading-breadth-chart-analyst",
        "morning",
        "alphaear-reporter",
        "hf-cli",
        "alphaear-signal-tracker",
        "trading-stanley-druckenmiller-investment",
        "kwp-sync",
        "trading-trade-hypothesis-ideator",
        "hf-trackio",
        "kwp-data-data-context-extractor",
        "hf-datasets",
        "trading-market-environment-analysis",
        "morning-ship"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: tab-stock-sync. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"tab-stock-sync\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"tab-fundamental-sync\", \"tab-dualma-refresh\", \"tab-llm-agents\", \"weekly-stock-update\", \"tab-kiwoom\", \"tab-genai-features\", \"tab-strategy-comparison\", \"tab-report-generate\", \"tab-turtle-refresh\", \"tab-market-breadth\", \"tab-dashboard-summary\", \"tab-news-fetch\", \"tab-bollinger-refresh\", \"stock-csv-downloader\", \"tab-technical-analysis\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0160",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Compensation Benchmarking. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-human-resources-compensation-benchmarking"
      ],
      "distractor_skills": [
        "kwp-human-resources-people-analytics",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-operations-resource-planning",
        "kwp-engineering-system-design",
        "kwp-human-resources-org-planning",
        "kwp-design-ux-writing",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-engineering-testing-strategy",
        "kwp-design-design-handoff",
        "kwp-operations-process-optimization",
        "kwp-operations-compliance-tracking",
        "kwp-operations-change-management",
        "kwp-common-room-account-research",
        "kwp-sales-competitive-intelligence",
        "kwp-design-system",
        "kwp-productivity-task-management",
        "kwp-product-management-competitive-analysis",
        "kwp-sales-daily-briefing",
        "kwp-operations-risk-assessment",
        "kwp-marketing-performance-analytics",
        "kwp-legal-legal-risk-assessment",
        "kwp-marketing-competitive-analysis",
        "kwp-data-data-context-extractor",
        "kwp-design-user-research",
        "kwp-product-management-metrics-tracking",
        "kwp-data-data-validation",
        "kwp-engineering-tech-debt",
        "kwp-engineering-documentation",
        "kwp-common-room-call-prep"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Compensation Benchmarking. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-human-resources-compensation-benchmarking\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-human-resources-people-analytics\", \"kwp-operations-vendor-management\", \"kwp-human-resources-interview-prep\", \"kwp-operations-resource-planning\", \"kwp-engineering-system-design\", \"kwp-human-resources-org-planning\", \"kwp-design-ux-writing\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-engineering-testing-strategy\", \"kwp-design-design-handoff\", \"kwp-operations-process-optimization\", \"kwp-operations-compliance-tracking\", \"kwp-operations-change-management\", \"kwp-common-room-account-research\", \"kwp-sales-competitive-intelligence\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0161",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: approaching context limits. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "ecc-strategic-compact"
      ],
      "distractor_skills": [
        "ecc-autonomous-loops",
        "ecc-search-first",
        "ecc-eval-harness",
        "ecc-verification-loop",
        "ecc-agentic-engineering",
        "ecc-configure",
        "ecc-coding-standards",
        "ecc-continuous-learning",
        "ecc-iterative-retrieval",
        "sp-writing-skills",
        "sp-tdd",
        "ralph-loop",
        "auto-research-distribute",
        "recall",
        "sp-verification",
        "kwp-sales-competitive-intelligence",
        "lead-lifecycle-manager",
        "kwp-legal-meeting-briefing",
        "kwp-data-data-exploration",
        "kwp-customer-support-customer-research",
        "anthropic-doc-coauthoring",
        "kwp-operations-risk-assessment",
        "kwp-engineering-system-design",
        "diagnose",
        "planning-meeting-sentinel",
        "sp-debugging"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: approaching context limits. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"ecc-strategic-compact\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"ecc-autonomous-loops\", \"ecc-search-first\", \"ecc-eval-harness\", \"ecc-verification-loop\", \"ecc-agentic-engineering\", \"ecc-configure\", \"ecc-coding-standards\", \"ecc-continuous-learning\", \"ecc-iterative-retrieval\", \"sp-writing-skills\", \"sp-tdd\", \"ralph-loop\", \"auto-research-distribute\", \"recall\", \"sp-verification\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0162",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: map data flow. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "system-thinker"
      ],
      "distractor_skills": [
        "automation-strategist",
        "pipeline-builder",
        "kwp-engineering-system-design",
        "ai-workflow-integrator",
        "agency-ai-engineer",
        "kwp-design-design-system-management",
        "kwp-design-design-critique",
        "agency-ux-architect",
        "agency-data-consolidation-agent",
        "agency-xr-cockpit-interaction-specialist",
        "kwp-operations-process-optimization",
        "agency-data-engineer",
        "agency-agents-orchestrator",
        "knowledge-to-paper",
        "agency-senior-project-manager",
        "cognee",
        "plans",
        "daily-strategy-post",
        "kwp-human-resources-people-analytics",
        "agency-roster",
        "agency-xr-immersive-developer",
        "ship",
        "ecc-autonomous-loops",
        "sefo-orchestrator",
        "issue-to-code",
        "kwp-operations-change-management",
        "kwp-engineering-code-review",
        "autoskill-merger",
        "kwp-productivity-task-management",
        "agency-visual-storyteller",
        "skill-composer",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-brand-voice-guideline-generation",
        "kwp-cowork-create-cowork-plugin",
        "kwp-data-data-visualization",
        "agency-mobile-app-builder",
        "kwp-finance-close-management",
        "pr-to-issue-linker",
        "deep-research-pipeline",
        "ux-expert",
        "kwp-engineering-tech-debt",
        "kwp-engineering-documentation",
        "agency-macos-spatial-metal-engineer",
        "agency-lsp-index-engineer",
        "agency-devops-automator"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: map data flow. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"system-thinker\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"automation-strategist\", \"pipeline-builder\", \"kwp-engineering-system-design\", \"ai-workflow-integrator\", \"agency-ai-engineer\", \"kwp-design-design-system-management\", \"kwp-design-design-critique\", \"agency-ux-architect\", \"agency-data-consolidation-agent\", \"agency-xr-cockpit-interaction-specialist\", \"kwp-operations-process-optimization\", \"agency-data-engineer\", \"agency-agents-orchestrator\", \"knowledge-to-paper\", \"agency-senior-project-manager\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0163",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: setting up real-time dashboards on HF Spaces. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "hf-trackio"
      ],
      "distractor_skills": [
        "hf-evaluation",
        "hf-endpoints",
        "hf-datasets",
        "hf-cli",
        "hf-jobs",
        "hf-dataset-viewer",
        "hf-models",
        "hf-collections",
        "hf-model-trainer",
        "hf-papers",
        "hf-spaces",
        "hf-buckets",
        "hf-hub",
        "hf-discussions",
        "hf-model-arena",
        "kwp-product-management-metrics-tracking"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: setting up real-time dashboards on HF Spaces. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"hf-trackio\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"hf-evaluation\", \"hf-endpoints\", \"hf-datasets\", \"hf-cli\", \"hf-jobs\", \"hf-dataset-viewer\", \"hf-models\", \"hf-collections\", \"hf-model-trainer\", \"hf-papers\", \"hf-spaces\", \"hf-buckets\", \"hf-hub\", \"hf-discussions\", \"hf-model-arena\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0164",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 논문 설명. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "alphaxiv-paper-lookup"
      ],
      "distractor_skills": [
        "related-papers-scout",
        "paper-review",
        "helm-validator",
        "auto-research",
        "nlm-arxiv-slides",
        "commit-to-issue",
        "auto-research-distribute",
        "ci-quality-gate",
        "ppt-template-engine",
        "meeting-digest",
        "hf-papers",
        "scrapling",
        "prd-research-factory"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 논문 설명. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"alphaxiv-paper-lookup\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"related-papers-scout\", \"paper-review\", \"helm-validator\", \"auto-research\", \"nlm-arxiv-slides\", \"commit-to-issue\", \"auto-research-distribute\", \"ci-quality-gate\", \"ppt-template-engine\", \"meeting-digest\", \"hf-papers\", \"scrapling\", \"prd-research-factory\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0165",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: detecting conflicts with installed skills. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "skill-seekers"
      ],
      "distractor_skills": [
        "skill-optimizer",
        "skill-guide-generator",
        "anthropic-skill-creator",
        "skill-composer",
        "prompt-transformer",
        "kwp-sync",
        "ecc-configure",
        "docs-tutor-setup",
        "gws-recipe-doc-from-template",
        "ecc-continuous-learning",
        "autoskill-extractor",
        "autoskill-merger",
        "sp-writing-skills",
        "md-enhance-publish",
        "autoskill-judge",
        "agency-roster",
        "commit-to-issue",
        "anthropic-algorithmic-art",
        "kwp-design-design-handoff",
        "gws-recipe-audit-sharing",
        "anthropic-pptx",
        "anthropic-docx",
        "kwp-engineering-documentation",
        "kwp-cowork-create-plugin",
        "nlm-video",
        "nlm-slides"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: detecting conflicts with installed skills. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"skill-seekers\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"skill-optimizer\", \"skill-guide-generator\", \"anthropic-skill-creator\", \"skill-composer\", \"prompt-transformer\", \"kwp-sync\", \"ecc-configure\", \"docs-tutor-setup\", \"gws-recipe-doc-from-template\", \"ecc-continuous-learning\", \"autoskill-extractor\", \"autoskill-merger\", \"sp-writing-skills\", \"md-enhance-publish\", \"autoskill-judge\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0166",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: executing implementation plans with independent tasks in the current session. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "sp-subagent-dev"
      ],
      "distractor_skills": [
        "sp-executing-plans",
        "sp-git-worktrees",
        "sp-tdd",
        "sp-brainstorming",
        "sp-request-review",
        "sp-writing-plans",
        "sp-parallel-agents",
        "sp-writing-skills",
        "sp-finish-branch",
        "kwp-human-resources-interview-prep",
        "sp-debugging",
        "kwp-design-design-handoff",
        "sp-core",
        "sp-receive-review",
        "kwp-operations-vendor-management",
        "anthropic-claude-api",
        "kwp-slack-slack-search",
        "kwp-finance-close-management",
        "kwp-design-system",
        "kwp-product-management-competitive-analysis",
        "kwp-common-room-account-research",
        "kwp-data-data-context-extractor",
        "kwp-human-resources-people-analytics",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-org-planning",
        "kwp-data-data-validation",
        "kwp-cowork-create-plugin",
        "sp-verification",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-finance-variance-analysis",
        "kwp-finance-financial-statements",
        "kwp-product-management-feature-spec",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-finance-audit-support",
        "kwp-data-interactive-dashboard-builder",
        "kwp-customer-support-escalation",
        "ship",
        "kwp-apollo-enrich-lead",
        "kwp-legal-meeting-briefing",
        "kwp-sales-daily-briefing",
        "kwp-sales-account-research",
        "kwp-marketing-performance-analytics",
        "kwp-marketing-campaign-planning",
        "kwp-legal-legal-risk-assessment",
        "kwp-finance-reconciliation",
        "kwp-finance-journal-entry-prep",
        "kwp-cowork-plugin-customizer",
        "kwp-apollo-prospect"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: executing implementation plans with independent tasks in the current session. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"sp-subagent-dev\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"sp-executing-plans\", \"sp-git-worktrees\", \"sp-tdd\", \"sp-brainstorming\", \"sp-request-review\", \"sp-writing-plans\", \"sp-parallel-agents\", \"sp-writing-skills\", \"sp-finish-branch\", \"kwp-human-resources-interview-prep\", \"sp-debugging\", \"kwp-design-design-handoff\", \"sp-core\", \"sp-receive-review\", \"kwp-operations-vendor-management\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0167",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Call Prep. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-common-room-call-prep"
      ],
      "distractor_skills": [
        "kwp-common-room-account-research",
        "kwp-common-room-weekly-prep-brief",
        "kwp-common-room-compose-outreach",
        "kwp-common-room-prospect",
        "kwp-common-room-contact-research",
        "kwp-sales-call-prep",
        "kwp-legal-meeting-briefing",
        "kwp-human-resources-employee-handbook",
        "kwp-data-data-context-extractor",
        "kwp-sales-draft-outreach",
        "kwp-slack-slack-messaging",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-apollo-enrich-lead",
        "kwp-slack-slack-search",
        "kwp-design-system",
        "kwp-apollo-prospect",
        "kwp-sales-daily-briefing",
        "kwp-sales-account-research",
        "kwp-finance-journal-entry-prep",
        "kwp-operations-resource-planning",
        "kwp-customer-support-response-drafting",
        "kwp-sales-competitive-intelligence",
        "kwp-human-resources-org-planning",
        "kwp-data-data-validation",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-customer-support-escalation",
        "kwp-customer-support-knowledge-management",
        "kwp-finance-close-management",
        "kwp-design-design-handoff"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Call Prep. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-common-room-call-prep\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-common-room-account-research\", \"kwp-common-room-weekly-prep-brief\", \"kwp-common-room-compose-outreach\", \"kwp-common-room-prospect\", \"kwp-common-room-contact-research\", \"kwp-sales-call-prep\", \"kwp-legal-meeting-briefing\", \"kwp-human-resources-employee-handbook\", \"kwp-data-data-context-extractor\", \"kwp-sales-draft-outreach\", \"kwp-slack-slack-messaging\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-operations-vendor-management\", \"kwp-human-resources-interview-prep\", \"kwp-apollo-enrich-lead\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0168",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: authenticate with Google. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "gws-workspace"
      ],
      "distractor_skills": [
        "gws-workflows",
        "gws-drive",
        "gws-gmail",
        "gws-chat",
        "gws-recipe-doc-from-template",
        "gws-docs",
        "gws-recipe-save-email-to-doc",
        "gws-sheets",
        "gws-recipe-personalized-emails",
        "gws-recipe-events-from-sheet",
        "gws-calendar",
        "gws-recipe-audit-sharing",
        "hf-hub",
        "google-daily",
        "gws-email-reply",
        "anthropic-claude-api",
        "calendar-daily-briefing",
        "local-dev-runner",
        "paperclip-setup",
        "agency-devops-automator",
        "agency-xr-immersive-developer",
        "ai-chief-of-staff",
        "agency-terminal-integration-specialist",
        "kwp-operations-vendor-management",
        "kwp-engineering-code-review",
        "md-enhance-publish",
        "kwp-cowork-plugin-customizer",
        "agency-xr-interface-architect",
        "docs-tutor-setup",
        "agency-infrastructure-maintainer",
        "agency-ux-architect",
        "agency-mobile-app-builder",
        "agency-data-consolidation-agent"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: authenticate with Google. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"gws-workspace\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"gws-workflows\", \"gws-drive\", \"gws-gmail\", \"gws-chat\", \"gws-recipe-doc-from-template\", \"gws-docs\", \"gws-recipe-save-email-to-doc\", \"gws-sheets\", \"gws-recipe-personalized-emails\", \"gws-recipe-events-from-sheet\", \"gws-calendar\", \"gws-recipe-audit-sharing\", \"hf-hub\", \"google-daily\", \"gws-email-reply\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0169",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: triage messages. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "gws-gmail"
      ],
      "distractor_skills": [
        "gws-chat",
        "gws-calendar",
        "gws-workflows",
        "gws-recipe-personalized-emails",
        "gws-workspace",
        "gws-sheets",
        "gws-drive",
        "gws-docs",
        "gws-email-reply",
        "gws-recipe-save-email-to-doc",
        "gws-recipe-audit-sharing"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: triage messages. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"gws-gmail\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"gws-chat\", \"gws-calendar\", \"gws-workflows\", \"gws-recipe-personalized-emails\", \"gws-workspace\", \"gws-sheets\", \"gws-drive\", \"gws-docs\", \"gws-email-reply\", \"gws-recipe-save-email-to-doc\", \"gws-recipe-audit-sharing\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0170",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Whimsy Injector agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-whimsy-injector"
      ],
      "distractor_skills": [
        "agency-app-store-optimizer",
        "agency-brand-guardian",
        "agency-tiktok-strategist",
        "agency-tool-evaluator",
        "agency-support-responder",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-test-results-analyzer",
        "agency-api-tester",
        "agency-behavioral-nudge-engine",
        "agency-zhihu-strategist",
        "agency-terminal-integration-specialist",
        "agency-workflow-optimizer",
        "agency-xr-interface-architect",
        "agency-twitter-engager",
        "agency-xiaohongshu-specialist",
        "agency-visual-storyteller",
        "agency-macos-spatial-metal-engineer",
        "agency-lsp-index-engineer",
        "agency-reddit-community-builder",
        "agency-instagram-curator",
        "agency-report-distribution-agent",
        "agency-inclusive-visuals-specialist",
        "agency-infrastructure-maintainer",
        "agency-cultural-intelligence-strategist",
        "agency-ux-architect",
        "agency-data-consolidation-agent",
        "agency-content-creator",
        "agency-visionos-spatial-engineer",
        "agency-senior-developer",
        "agency-trend-researcher",
        "agency-experiment-tracker"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Whimsy Injector agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-whimsy-injector\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-app-store-optimizer\", \"agency-brand-guardian\", \"agency-tiktok-strategist\", \"agency-tool-evaluator\", \"agency-support-responder\", \"agency-xr-immersive-developer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-test-results-analyzer\", \"agency-api-tester\", \"agency-behavioral-nudge-engine\", \"agency-zhihu-strategist\", \"agency-terminal-integration-specialist\", \"agency-workflow-optimizer\", \"agency-xr-interface-architect\", \"agency-twitter-engager\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0171",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Vendor Management. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-operations-vendor-management"
      ],
      "distractor_skills": [
        "kwp-human-resources-interview-prep",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-org-planning",
        "kwp-operations-resource-planning",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-human-resources-people-analytics",
        "kwp-design-system",
        "kwp-finance-close-management",
        "kwp-product-management-feature-spec",
        "kwp-engineering-system-design",
        "kwp-sales-competitive-intelligence",
        "kwp-design-design-system-management",
        "kwp-design-design-critique",
        "kwp-slack-slack-search",
        "kwp-design-design-handoff",
        "kwp-engineering-code-review",
        "kwp-sales-call-prep",
        "kwp-product-management-competitive-analysis",
        "kwp-design-accessibility-review",
        "kwp-sales-daily-briefing",
        "kwp-sales-account-research",
        "kwp-operations-risk-assessment",
        "kwp-legal-legal-risk-assessment",
        "kwp-marketing-competitive-analysis",
        "kwp-legal-contract-review",
        "kwp-engineering-incident-response",
        "kwp-sales-draft-outreach",
        "kwp-product-management-metrics-tracking",
        "kwp-data-data-validation",
        "kwp-engineering-testing-strategy",
        "kwp-engineering-tech-debt",
        "kwp-finance-variance-analysis",
        "kwp-finance-financial-statements",
        "kwp-operations-compliance-tracking",
        "kwp-legal-canned-responses",
        "kwp-finance-audit-support",
        "kwp-data-interactive-dashboard-builder",
        "kwp-customer-support-escalation",
        "kwp-apollo-enrich-lead",
        "kwp-legal-meeting-briefing",
        "kwp-legal-compliance",
        "kwp-design-ux-writing",
        "kwp-customer-support-knowledge-management",
        "kwp-finance-reconciliation"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Vendor Management. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-operations-vendor-management\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-human-resources-interview-prep\", \"kwp-operations-process-optimization\", \"kwp-operations-change-management\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-human-resources-org-planning\", \"kwp-operations-resource-planning\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-human-resources-people-analytics\", \"kwp-design-system\", \"kwp-finance-close-management\", \"kwp-product-management-feature-spec\", \"kwp-engineering-system-design\", \"kwp-sales-competitive-intelligence\", \"kwp-design-design-system-management\", \"kwp-design-design-critique\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0172",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 영업 대시보드. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "sales-weekly-dashboard"
      ],
      "distractor_skills": [
        "sales-meeting-pipeline",
        "planning-weekly-pulse",
        "lead-lifecycle-manager",
        "docs-freshness-guardian",
        "proposal-factory",
        "dependency-radar",
        "prd-research-factory",
        "planning-meeting-sentinel",
        "nlm-arxiv-slides",
        "feature-intake-router",
        "incident-lifecycle-manager",
        "hf-trending-intelligence",
        "bespin-news-digest",
        "gws-workflows",
        "md-enhance-publish",
        "daily-stock-check",
        "knowledge-to-paper",
        "kwp-human-resources-people-analytics",
        "morning",
        "meeting-digest",
        "tech-support-brain",
        "pm-data-analytics",
        "stock-csv-downloader",
        "agency-data-consolidation-agent",
        "paper-review",
        "deep-research-pipeline",
        "ci-quality-gate",
        "agency-sales-data-extraction-agent"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 영업 대시보드. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"sales-weekly-dashboard\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"sales-meeting-pipeline\", \"planning-weekly-pulse\", \"lead-lifecycle-manager\", \"docs-freshness-guardian\", \"proposal-factory\", \"dependency-radar\", \"prd-research-factory\", \"planning-meeting-sentinel\", \"nlm-arxiv-slides\", \"feature-intake-router\", \"incident-lifecycle-manager\", \"hf-trending-intelligence\", \"bespin-news-digest\", \"gws-workflows\", \"md-enhance-publish\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0173",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: generate a multi-phase plan. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "full-stack-planner"
      ],
      "distractor_skills": [
        "plans",
        "onboarding-accelerator",
        "daily-am-orchestrator",
        "context-engineer",
        "prd-research-factory",
        "problem-definition",
        "setup-doctor",
        "cursor-automations",
        "pipeline-builder",
        "ppt-template-engine",
        "eod-ship",
        "critical-review",
        "commit-to-issue",
        "code-review-all",
        "technical-writer",
        "jira-linear-adapter",
        "auto-research-distribute",
        "daily-strategy-post",
        "auto-research",
        "scrapling",
        "mission-control copy",
        "helm-validator",
        "proposal-factory",
        "executive-briefing",
        "alphaxiv-paper-lookup",
        "mission-control",
        "simplify",
        "feature-intake-router",
        "cursor-sync",
        "decision-router",
        "ci-quality-gate",
        "related-papers-scout",
        "automation-strategist",
        "calendar-daily-briefing",
        "refactor-simulator",
        "k8s-manifest-validator",
        "performance-profiler",
        "sprint-orchestrator",
        "nlm-strategy-hub",
        "role-ux-designer",
        "domain-commit",
        "ai-workflow-integrator"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: generate a multi-phase plan. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"full-stack-planner\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"plans\", \"onboarding-accelerator\", \"daily-am-orchestrator\", \"context-engineer\", \"prd-research-factory\", \"problem-definition\", \"setup-doctor\", \"cursor-automations\", \"pipeline-builder\", \"ppt-template-engine\", \"eod-ship\", \"critical-review\", \"commit-to-issue\", \"code-review-all\", \"technical-writer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0174",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: UX 관점. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "role-ux-designer"
      ],
      "distractor_skills": [
        "role-finance",
        "role-hr",
        "role-data-scientist",
        "role-pm",
        "role-developer",
        "role-sales",
        "role-security-engineer",
        "role-cso",
        "role-ceo",
        "role-trading-expert",
        "role-cto",
        "role-dispatcher",
        "deep-research-pipeline",
        "kwp-design-design-system-management",
        "ux-expert",
        "kwp-design-user-research",
        "kwp-design-design-critique",
        "kwp-engineering-system-design",
        "paper-review",
        "design-architect",
        "kwp-design-accessibility-review",
        "agency-infrastructure-maintainer",
        "agency-ux-researcher",
        "agency-xr-immersive-developer",
        "agency-brand-guardian",
        "executive-briefing",
        "kwp-operations-compliance-tracking",
        "kwp-engineering-code-review",
        "agency-project-shepherd",
        "kwp-design-ux-writing",
        "frontend-expert",
        "agency-accessibility-auditor",
        "agency-finance-tracker",
        "agency-xr-interface-architect",
        "full-stack-planner",
        "agency-legal-compliance-checker",
        "kwp-engineering-testing-strategy",
        "kwp-engineering-tech-debt"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: UX 관점. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"role-ux-designer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"role-finance\", \"role-hr\", \"role-data-scientist\", \"role-pm\", \"role-developer\", \"role-sales\", \"role-security-engineer\", \"role-cso\", \"role-ceo\", \"role-trading-expert\", \"role-cto\", \"role-dispatcher\", \"deep-research-pipeline\", \"kwp-design-design-system-management\", \"ux-expert\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0175",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: manage cache. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "hf-hub"
      ],
      "distractor_skills": [
        "hf-collections",
        "hf-models",
        "hf-discussions",
        "hf-repos",
        "hf-buckets",
        "hf-spaces",
        "hf-endpoints",
        "hf-cli",
        "hf-papers",
        "gws-workspace",
        "hf-datasets",
        "hf-dataset-viewer",
        "hf-trending-intelligence",
        "hf-model-trainer",
        "hf-evaluation",
        "hf-model-arena",
        "hf-auto-finetune-planner",
        "hf-trackio",
        "hf-jobs",
        "paperclip-setup",
        "gws-gmail",
        "gws-chat",
        "gws-drive",
        "local-dev-runner",
        "gws-docs",
        "skill-optimizer",
        "kwp-operations-vendor-management",
        "gws-sheets",
        "alphaear-stock",
        "gws-recipe-audit-sharing",
        "anthropic-claude-api",
        "paperclip-agents",
        "agency-infrastructure-maintainer",
        "agency-devops-automator",
        "pm-product-strategy",
        "domain-commit",
        "gws-calendar",
        "gws-recipe-doc-from-template",
        "agency-xr-immersive-developer",
        "agency-studio-producer",
        "agency-terminal-integration-specialist",
        "kwp-operations-change-management",
        "kwp-engineering-code-review",
        "agency-project-shepherd",
        "kwp-sync",
        "kwp-design-accessibility-review",
        "sefo-federation",
        "agency-legal-compliance-checker",
        "trading-market-environment-analysis"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: manage cache. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"hf-hub\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"hf-collections\", \"hf-models\", \"hf-discussions\", \"hf-repos\", \"hf-buckets\", \"hf-spaces\", \"hf-endpoints\", \"hf-cli\", \"hf-papers\", \"gws-workspace\", \"hf-datasets\", \"hf-dataset-viewer\", \"hf-trending-intelligence\", \"hf-model-trainer\", \"hf-evaluation\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0176",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: build a plugin. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-cowork-create-plugin"
      ],
      "distractor_skills": [
        "kwp-cowork-create-cowork-plugin",
        "kwp-design-system",
        "kwp-design-design-handoff",
        "kwp-cowork-plugin-customizer",
        "kwp-data-data-context-extractor",
        "kwp-design-user-research",
        "kwp-data-data-visualization",
        "kwp-sales-competitive-intelligence",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-operations-process-optimization",
        "kwp-data-interactive-dashboard-builder",
        "kwp-brand-voice-guideline-generation",
        "kwp-bio-research-scientific-problem-selection",
        "kwp-sync",
        "kwp-human-resources-people-analytics",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-org-planning",
        "kwp-brand-voice-brand-voice-enforcement",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-finance-financial-statements",
        "kwp-product-management-feature-spec",
        "kwp-operations-change-management",
        "kwp-productivity-task-management",
        "kwp-customer-support-knowledge-management",
        "kwp-finance-close-management",
        "kwp-enterprise-search-source-management",
        "kwp-sales-daily-briefing",
        "kwp-product-management-competitive-analysis",
        "kwp-design-design-system-management",
        "kwp-slack-slack-search",
        "kwp-marketing-brand-voice",
        "kwp-legal-nda-triage",
        "kwp-engineering-documentation",
        "kwp-common-room-account-research",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-data-data-validation",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-finance-variance-analysis",
        "kwp-finance-audit-support",
        "kwp-enterprise-search-knowledge-synthesis",
        "kwp-customer-support-escalation",
        "kwp-legal-meeting-briefing",
        "kwp-legal-compliance",
        "kwp-design-accessibility-review",
        "kwp-data-data-exploration",
        "kwp-cowork-cowork-plugin-customizer",
        "kwp-apollo-enrich-lead",
        "kwp-sales-create-an-asset",
        "kwp-sales-account-research"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: build a plugin. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-cowork-create-plugin\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-cowork-create-cowork-plugin\", \"kwp-design-system\", \"kwp-design-design-handoff\", \"kwp-cowork-plugin-customizer\", \"kwp-data-data-context-extractor\", \"kwp-design-user-research\", \"kwp-data-data-visualization\", \"kwp-sales-competitive-intelligence\", \"kwp-operations-vendor-management\", \"kwp-human-resources-interview-prep\", \"kwp-operations-process-optimization\", \"kwp-data-interactive-dashboard-builder\", \"kwp-brand-voice-guideline-generation\", \"kwp-bio-research-scientific-problem-selection\", \"kwp-sync\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0177",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: RSS event scan. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "tab-event-detect"
      ],
      "distractor_skills": [
        "tab-analysis-run",
        "tab-patterns",
        "tab-fundamental-sync",
        "tab-genai-features",
        "tab-screening",
        "tab-turtle-refresh",
        "tab-strategy-comparison",
        "tab-dualma-refresh",
        "tab-technical-analysis",
        "tab-llm-agents",
        "tab-hot-stock-discovery",
        "tab-market-breadth",
        "tab-dashboard-summary",
        "tab-news-fetch",
        "tab-report-generate",
        "tab-stock-sync",
        "tab-bollinger-refresh",
        "tab-glossary",
        "tab-market-environment",
        "tab-kiwoom"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: RSS event scan. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"tab-event-detect\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"tab-analysis-run\", \"tab-patterns\", \"tab-fundamental-sync\", \"tab-genai-features\", \"tab-screening\", \"tab-turtle-refresh\", \"tab-strategy-comparison\", \"tab-dualma-refresh\", \"tab-technical-analysis\", \"tab-llm-agents\", \"tab-hot-stock-discovery\", \"tab-market-breadth\", \"tab-dashboard-summary\", \"tab-news-fetch\", \"tab-report-generate\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0178",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Skill Composer. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "skill-composer"
      ],
      "distractor_skills": [
        "skill-guide-generator",
        "skill-optimizer",
        "autoskill-extractor",
        "skill-seekers",
        "autoskill-judge",
        "agency-roster",
        "kwp-sync",
        "workflow-miner",
        "anthropic-skill-creator",
        "autoskill-merger",
        "ai-workflow-integrator",
        "skill-autoimprove",
        "autoskill-evolve",
        "sp-writing-skills",
        "kwp-cowork-create-cowork-plugin",
        "release-commander",
        "pipeline-builder",
        "kwp-cowork-create-plugin",
        "setup-doctor",
        "agency-workflow-optimizer",
        "prompt-transformer",
        "gws-recipe-save-email-to-doc",
        "kwp-brand-voice-guideline-generation",
        "pm-data-analytics",
        "knowledge-to-paper",
        "hf-auto-finetune-planner",
        "plans",
        "nlm-video",
        "nlm-slides",
        "agency-image-prompt-engineer",
        "onboarding-accelerator"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Skill Composer. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"skill-composer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"skill-guide-generator\", \"skill-optimizer\", \"autoskill-extractor\", \"skill-seekers\", \"autoskill-judge\", \"agency-roster\", \"kwp-sync\", \"workflow-miner\", \"anthropic-skill-creator\", \"autoskill-merger\", \"ai-workflow-integrator\", \"skill-autoimprove\", \"autoskill-evolve\", \"sp-writing-skills\", \"kwp-cowork-create-cowork-plugin\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0179",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: any request to simulate public opinion dynamics. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "mirofish-opinion-sim"
      ],
      "distractor_skills": [
        "mirofish-financial-sim",
        "mirofish",
        "mirofish-graph-explorer",
        "twitter-timeline-to-slack",
        "agency-social-media-strategist",
        "trading-market-news-analyst",
        "slack-agent",
        "cursor-automations",
        "trading-sector-analyst",
        "paperclip-setup",
        "standup-digest",
        "alphaear-deepear-lite",
        "alphaear-sentiment",
        "daily-am-orchestrator",
        "agency-wechat-official-account-manager",
        "daiso-mcp",
        "notebooklm-research",
        "daily-strategy-post",
        "agency-lsp-index-engineer",
        "agency-devops-automator",
        "trading-market-breadth-analyzer",
        "helm-validator",
        "skill-optimizer",
        "sefo-orchestrator",
        "terraform-reviewer",
        "paperclip-agents",
        "dependency-auditor",
        "paper-review",
        "k8s-manifest-validator",
        "diagnose",
        "auto-research-distribute",
        "agency-trend-researcher",
        "agency-behavioral-nudge-engine",
        "trading-options-theta",
        "hf-trending-intelligence",
        "email-research-dispatcher",
        "deep-review",
        "intent-alignment-tracker",
        "agency-rapid-prototyper",
        "role-dispatcher",
        "agency-twitter-engager",
        "agency-developer-advocate",
        "paperclip-tasks",
        "domain-commit",
        "autoskill-extractor",
        "alphaear-search",
        "alphaear-news",
        "agency-tool-evaluator",
        "gws-chat"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: any request to simulate public opinion dynamics. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"mirofish-opinion-sim\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"mirofish-financial-sim\", \"mirofish\", \"mirofish-graph-explorer\", \"twitter-timeline-to-slack\", \"agency-social-media-strategist\", \"trading-market-news-analyst\", \"slack-agent\", \"cursor-automations\", \"trading-sector-analyst\", \"paperclip-setup\", \"standup-digest\", \"alphaear-deepear-lite\", \"alphaear-sentiment\", \"daily-am-orchestrator\", \"agency-wechat-official-account-manager\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0180",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Screen Implementation Orchestrator. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "implement-screen"
      ],
      "distractor_skills": [
        "design-review",
        "figma-to-tds",
        "fe-pipeline",
        "screen-description",
        "fsd-development",
        "swagger-api-doc-generator",
        "github-workflow-automation",
        "overlay-layout-patterns",
        "notion-docs-sync",
        "technical-writer",
        "sre-devops-expert",
        "k8s-manifest-validator",
        "performance-profiler",
        "auto-research-distribute"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Screen Implementation Orchestrator. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"implement-screen\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"design-review\", \"figma-to-tds\", \"fe-pipeline\", \"screen-description\", \"fsd-development\", \"swagger-api-doc-generator\", \"github-workflow-automation\", \"overlay-layout-patterns\", \"notion-docs-sync\", \"technical-writer\", \"sre-devops-expert\", \"k8s-manifest-validator\", \"performance-profiler\", \"auto-research-distribute\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0181",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: recent price changes. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "alphaear-stock"
      ],
      "distractor_skills": [
        "stock-csv-downloader",
        "alphaear-search",
        "alphaear-deepear-lite",
        "weekly-stock-update",
        "daily-stock-check",
        "alphaear-news",
        "tab-stock-sync",
        "alphaear-predictor",
        "alphaear-sentiment",
        "trading-sector-analyst",
        "alphaear-reporter",
        "alphaear-signal-tracker",
        "trading-technical-analyst",
        "trading-us-stock-analysis",
        "trading-market-environment-analysis",
        "trading-stanley-druckenmiller-investment",
        "kwp-sync",
        "agency-xr-interface-architect",
        "today",
        "agency-ux-architect",
        "kwp-engineering-tech-debt",
        "trading-breadth-chart-analyst",
        "gws-recipe-doc-from-template",
        "gws-drive",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "hf-models",
        "agency-test-results-analyzer",
        "kwp-data-data-context-extractor",
        "agency-terminal-integration-specialist",
        "agency-brand-guardian"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: recent price changes. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-stock\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"stock-csv-downloader\", \"alphaear-search\", \"alphaear-deepear-lite\", \"weekly-stock-update\", \"daily-stock-check\", \"alphaear-news\", \"tab-stock-sync\", \"alphaear-predictor\", \"alphaear-sentiment\", \"trading-sector-analyst\", \"alphaear-reporter\", \"alphaear-signal-tracker\", \"trading-technical-analyst\", \"trading-us-stock-analysis\", \"trading-market-environment-analysis\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0182",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: UI 변경 요청 시. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "fe-pipeline"
      ],
      "distractor_skills": [
        "implement-screen",
        "design-review",
        "swagger-api-doc-generator",
        "figma-to-tds",
        "screen-description",
        "github-workflow-automation",
        "fsd-development",
        "technical-writer",
        "overlay-layout-patterns",
        "sre-devops-expert",
        "performance-profiler",
        "notion-docs-sync",
        "k8s-manifest-validator",
        "auto-research-distribute",
        "ci-quality-gate",
        "security-expert",
        "commit-to-issue",
        "feature-intake-router",
        "proposal-factory",
        "prd-research-factory",
        "alphaxiv-paper-lookup",
        "auto-research",
        "calendar-daily-briefing",
        "helm-validator",
        "anthropic-claude-api",
        "code-review-all",
        "simplify",
        "related-papers-scout",
        "problem-definition",
        "domain-commit"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: UI 변경 요청 시. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"fe-pipeline\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"implement-screen\", \"design-review\", \"swagger-api-doc-generator\", \"figma-to-tds\", \"screen-description\", \"github-workflow-automation\", \"fsd-development\", \"technical-writer\", \"overlay-layout-patterns\", \"sre-devops-expert\", \"performance-profiler\", \"notion-docs-sync\", \"k8s-manifest-validator\", \"auto-research-distribute\", \"ci-quality-gate\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0183",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: autonomous loop. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "pm-sentinel"
      ],
      "distractor_skills": [
        "ralph-loop",
        "pm-toolkit",
        "pm-product-strategy",
        "release-commander",
        "pm-data-analytics",
        "pm-marketing-growth",
        "mission-control copy",
        "pm-market-research",
        "pm-execution",
        "pm-go-to-market",
        "plans",
        "sefo-orchestrator",
        "mission-control",
        "pm-product-discovery",
        "ai-workflow-integrator",
        "agency-xr-immersive-developer",
        "agency-report-distribution-agent",
        "agency-cultural-intelligence-strategist",
        "refactor-simulator",
        "full-stack-planner",
        "skill-autoimprove",
        "agency-autonomous-optimization-architect",
        "test-suite",
        "intent-alignment-tracker",
        "agency-reality-checker",
        "agency-rapid-prototyper",
        "agency-data-consolidation-agent"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: autonomous loop. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"pm-sentinel\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"ralph-loop\", \"pm-toolkit\", \"pm-product-strategy\", \"release-commander\", \"pm-data-analytics\", \"pm-marketing-growth\", \"mission-control copy\", \"pm-market-research\", \"pm-execution\", \"pm-go-to-market\", \"plans\", \"sefo-orchestrator\", \"mission-control\", \"pm-product-discovery\", \"ai-workflow-integrator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0184",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: runbook generation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "docs-freshness-guardian"
      ],
      "distractor_skills": [
        "lead-lifecycle-manager",
        "sales-weekly-dashboard",
        "dependency-radar",
        "docs-tutor-setup",
        "planning-weekly-pulse",
        "feature-intake-router",
        "incident-lifecycle-manager",
        "planning-meeting-sentinel",
        "ui-suite",
        "docs-tutor",
        "sales-meeting-pipeline",
        "prd-research-factory",
        "proposal-factory",
        "ship",
        "i18n-sync",
        "tech-support-brain",
        "anthropic-doc-coauthoring",
        "skill-guide-generator",
        "pr-review-captain",
        "kwp-engineering-documentation",
        "deep-review",
        "pm-toolkit",
        "skill-optimizer",
        "gws-recipe-doc-from-template",
        "frontend-expert",
        "ecc-verification-loop",
        "role-developer",
        "kwp-product-management-feature-spec",
        "test-suite",
        "release-commander",
        "compliance-gate",
        "qa-test-expert",
        "agency-agents-orchestrator",
        "sp-debugging",
        "diagnose",
        "agency-data-consolidation-agent",
        "auto-research",
        "sp-receive-review"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: runbook generation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"docs-freshness-guardian\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"lead-lifecycle-manager\", \"sales-weekly-dashboard\", \"dependency-radar\", \"docs-tutor-setup\", \"planning-weekly-pulse\", \"feature-intake-router\", \"incident-lifecycle-manager\", \"planning-meeting-sentinel\", \"ui-suite\", \"docs-tutor\", \"sales-meeting-pipeline\", \"prd-research-factory\", \"proposal-factory\", \"ship\", \"i18n-sync\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0185",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: tab-bollinger-refresh. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "tab-bollinger-refresh"
      ],
      "distractor_skills": [
        "tab-turtle-refresh",
        "tab-dualma-refresh",
        "tab-strategy-comparison",
        "tab-market-breadth",
        "tab-technical-analysis",
        "tab-llm-agents",
        "tab-report-generate",
        "tab-market-environment",
        "tab-fundamental-sync",
        "tab-hot-stock-discovery",
        "tab-screening",
        "tab-news-fetch",
        "tab-stock-sync",
        "tab-genai-features",
        "tab-patterns",
        "tab-glossary",
        "tab-dashboard-summary",
        "tab-event-detect",
        "tab-analysis-run",
        "tab-kiwoom",
        "daily-stock-check",
        "trading-backtest-expert",
        "stock-csv-downloader",
        "trading-breadth-chart-analyst",
        "alphaear-predictor",
        "trading-trade-hypothesis-ideator",
        "trading-market-environment-analysis",
        "trading-sector-analyst",
        "trading-market-breadth-analyzer",
        "trading-us-stock-analysis",
        "trading-theme-detector",
        "today",
        "trading-technical-analyst"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: tab-bollinger-refresh. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"tab-bollinger-refresh\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"tab-turtle-refresh\", \"tab-dualma-refresh\", \"tab-strategy-comparison\", \"tab-market-breadth\", \"tab-technical-analysis\", \"tab-llm-agents\", \"tab-report-generate\", \"tab-market-environment\", \"tab-fundamental-sync\", \"tab-hot-stock-discovery\", \"tab-screening\", \"tab-news-fetch\", \"tab-stock-sync\", \"tab-genai-features\", \"tab-patterns\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0186",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: risk communications. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-product-management-stakeholder-comms"
      ],
      "distractor_skills": [
        "kwp-customer-support-escalation",
        "kwp-design-design-handoff",
        "kwp-marketing-campaign-planning",
        "kwp-customer-support-response-drafting",
        "kwp-product-management-metrics-tracking",
        "kwp-product-management-feature-spec",
        "kwp-finance-close-management",
        "kwp-slack-slack-search",
        "kwp-design-system",
        "kwp-product-management-competitive-analysis",
        "kwp-sales-create-an-asset",
        "kwp-human-resources-people-analytics",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-data-data-validation",
        "kwp-data-data-context-extractor",
        "kwp-finance-financial-statements",
        "kwp-common-room-compose-outreach",
        "kwp-finance-audit-support",
        "kwp-data-interactive-dashboard-builder",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-marketing-content-creation",
        "kwp-legal-compliance",
        "kwp-customer-support-customer-research",
        "kwp-product-management-user-research-synthesis",
        "kwp-product-management-roadmap-management",
        "kwp-marketing-performance-analytics",
        "kwp-legal-legal-risk-assessment",
        "kwp-finance-journal-entry-prep",
        "kwp-finance-reconciliation",
        "kwp-cowork-plugin-customizer",
        "kwp-common-room-weekly-prep-brief",
        "kwp-apollo-prospect",
        "kwp-marketing-competitive-analysis",
        "kwp-marketing-brand-voice",
        "kwp-legal-nda-triage",
        "kwp-data-sql-queries",
        "kwp-common-room-contact-research",
        "kwp-common-room-account-research",
        "kwp-sales-draft-outreach",
        "kwp-human-resources-org-planning",
        "kwp-data-statistical-analysis",
        "kwp-human-resources-compensation-benchmarking"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: risk communications. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-product-management-stakeholder-comms\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-customer-support-escalation\", \"kwp-design-design-handoff\", \"kwp-marketing-campaign-planning\", \"kwp-customer-support-response-drafting\", \"kwp-product-management-metrics-tracking\", \"kwp-product-management-feature-spec\", \"kwp-finance-close-management\", \"kwp-slack-slack-search\", \"kwp-design-system\", \"kwp-product-management-competitive-analysis\", \"kwp-sales-create-an-asset\", \"kwp-human-resources-people-analytics\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-data-data-validation\", \"kwp-data-data-context-extractor\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0187",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 런북 생성. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "infra-runbook-generator"
      ],
      "distractor_skills": [
        "infra-drift-detector",
        "sre-devops-expert",
        "helm-validator",
        "docs-tutor-setup",
        "kwp-engineering-documentation",
        "k8s-manifest-validator",
        "terraform-reviewer",
        "incident-to-improvement",
        "iac-review-agent",
        "role-cto",
        "agency-devops-automator",
        "kwp-engineering-incident-response",
        "service-health-doctor",
        "local-dev-runner",
        "skill-guide-generator",
        "prompt-transformer",
        "i18n-sync",
        "role-developer",
        "pr-review-captain",
        "gws-recipe-doc-from-template"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 런북 생성. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"infra-runbook-generator\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"infra-drift-detector\", \"sre-devops-expert\", \"helm-validator\", \"docs-tutor-setup\", \"kwp-engineering-documentation\", \"k8s-manifest-validator\", \"terraform-reviewer\", \"incident-to-improvement\", \"iac-review-agent\", \"role-cto\", \"agency-devops-automator\", \"kwp-engineering-incident-response\", \"service-health-doctor\", \"local-dev-runner\", \"skill-guide-generator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0188",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: fetch stock CSVs. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "stock-csv-downloader"
      ],
      "distractor_skills": [
        "weekly-stock-update",
        "daily-stock-check",
        "alphaear-stock",
        "tab-stock-sync",
        "trading-sector-analyst",
        "alphaear-search",
        "alphaear-news",
        "alphaear-deepear-lite",
        "trading-market-environment-analysis",
        "trading-technical-analyst",
        "tab-kiwoom"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: fetch stock CSVs. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"stock-csv-downloader\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"weekly-stock-update\", \"daily-stock-check\", \"alphaear-stock\", \"tab-stock-sync\", \"trading-sector-analyst\", \"alphaear-search\", \"alphaear-news\", \"alphaear-deepear-lite\", \"trading-market-environment-analysis\", \"trading-technical-analyst\", \"tab-kiwoom\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0189",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: before we solve this. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "problem-definition"
      ],
      "distractor_skills": [
        "refactor-simulator",
        "codebase-archaeologist",
        "code-review-all",
        "simplify",
        "plans",
        "commit-to-issue",
        "technical-writer",
        "diagnose",
        "critical-review",
        "performance-profiler",
        "calendar-daily-briefing",
        "alphaxiv-paper-lookup",
        "sre-devops-expert",
        "auto-research",
        "full-stack-planner",
        "ci-quality-gate",
        "onboarding-accelerator",
        "helm-validator",
        "security-expert",
        "executive-briefing",
        "related-papers-scout",
        "kwp-engineering-code-review",
        "ship",
        "k8s-manifest-validator",
        "daily-stock-check"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: before we solve this. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"problem-definition\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"refactor-simulator\", \"codebase-archaeologist\", \"code-review-all\", \"simplify\", \"plans\", \"commit-to-issue\", \"technical-writer\", \"diagnose\", \"critical-review\", \"performance-profiler\", \"calendar-daily-briefing\", \"alphaxiv-paper-lookup\", \"sre-devops-expert\", \"auto-research\", \"full-stack-planner\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0190",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 추세 분석. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "trading-technical-analyst"
      ],
      "distractor_skills": [
        "trading-us-stock-analysis",
        "trading-scenario-analyzer",
        "trading-sector-analyst",
        "trading-breadth-chart-analyst",
        "trading-backtest-expert",
        "trading-stanley-druckenmiller-investment",
        "trading-us-market-bubble-detector",
        "trading-market-environment-analysis",
        "trading-market-breadth-analyzer",
        "trading-finviz-screener",
        "stock-csv-downloader",
        "daily-stock-check",
        "trading-market-top-detector",
        "trading-theme-detector",
        "alphaear-deepear-lite"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 추세 분석. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"trading-technical-analyst\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"trading-us-stock-analysis\", \"trading-scenario-analyzer\", \"trading-sector-analyst\", \"trading-breadth-chart-analyst\", \"trading-backtest-expert\", \"trading-stanley-druckenmiller-investment\", \"trading-us-market-bubble-detector\", \"trading-market-environment-analysis\", \"trading-market-breadth-analyzer\", \"trading-finviz-screener\", \"stock-csv-downloader\", \"daily-stock-check\", \"trading-market-top-detector\", \"trading-theme-detector\", \"alphaear-deepear-lite\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0191",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: brand colors. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "anthropic-brand-guidelines"
      ],
      "distractor_skills": [
        "anthropic-theme-factory",
        "anthropic-slack-gif-creator",
        "anthropic-canvas-design",
        "anthropic-claude-api",
        "anthropic-algorithmic-art",
        "anthropic-mcp-builder",
        "anthropic-web-artifacts-builder",
        "anthropic-webapp-testing",
        "anthropic-skill-creator",
        "kwp-marketing-brand-voice",
        "anthropic-frontend-design",
        "anthropic-internal-comms",
        "anthropic-docx",
        "anthropic-doc-coauthoring",
        "anthropic-pdf",
        "anthropic-pptx",
        "kwp-design-system",
        "anthropic-xlsx",
        "agency-ui-designer",
        "pm-marketing-growth",
        "kwp-brand-voice-brand-voice-enforcement",
        "kwp-brand-voice-guideline-generation",
        "sp-tdd",
        "sp-git-worktrees",
        "alphaear-logic-visualizer",
        "sp-debugging",
        "agency-accessibility-auditor",
        "kwp-sync",
        "kwp-data-data-context-extractor",
        "agency-visual-storyteller",
        "sp-core",
        "kwp-sales-account-research",
        "design-architect",
        "kwp-design-design-handoff",
        "kwp-design-user-research"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: brand colors. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-brand-guidelines\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"anthropic-theme-factory\", \"anthropic-slack-gif-creator\", \"anthropic-canvas-design\", \"anthropic-claude-api\", \"anthropic-algorithmic-art\", \"anthropic-mcp-builder\", \"anthropic-web-artifacts-builder\", \"anthropic-webapp-testing\", \"anthropic-skill-creator\", \"kwp-marketing-brand-voice\", \"anthropic-frontend-design\", \"anthropic-internal-comms\", \"anthropic-docx\", \"anthropic-doc-coauthoring\", \"anthropic-pdf\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0192",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: references agency-support-responder. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-support-responder"
      ],
      "distractor_skills": [
        "agency-app-store-optimizer",
        "agency-terminal-integration-specialist",
        "agency-brand-guardian",
        "agency-tiktok-strategist",
        "agency-tool-evaluator",
        "agency-whimsy-injector",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-instagram-curator",
        "agency-xr-interface-architect",
        "agency-macos-spatial-metal-engineer",
        "agency-test-results-analyzer",
        "agency-inclusive-visuals-specialist",
        "agency-api-tester"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: references agency-support-responder. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-support-responder\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-app-store-optimizer\", \"agency-terminal-integration-specialist\", \"agency-brand-guardian\", \"agency-tiktok-strategist\", \"agency-tool-evaluator\", \"agency-whimsy-injector\", \"agency-xr-immersive-developer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-instagram-curator\", \"agency-xr-interface-architect\", \"agency-macos-spatial-metal-engineer\", \"agency-test-results-analyzer\", \"agency-inclusive-visuals-specialist\", \"agency-api-tester\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0193",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: reviewing data processing agreements. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-legal-compliance"
      ],
      "distractor_skills": [
        "kwp-legal-canned-responses",
        "kwp-legal-legal-risk-assessment",
        "kwp-legal-contract-review",
        "kwp-marketing-brand-voice",
        "kwp-customer-support-response-drafting",
        "kwp-data-data-validation",
        "kwp-legal-nda-triage",
        "kwp-operations-vendor-management",
        "kwp-slack-slack-search",
        "kwp-legal-meeting-briefing",
        "kwp-data-data-exploration",
        "kwp-design-system",
        "kwp-product-management-competitive-analysis"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: reviewing data processing agreements. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-legal-compliance\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-legal-canned-responses\", \"kwp-legal-legal-risk-assessment\", \"kwp-legal-contract-review\", \"kwp-marketing-brand-voice\", \"kwp-customer-support-response-drafting\", \"kwp-data-data-validation\", \"kwp-legal-nda-triage\", \"kwp-operations-vendor-management\", \"kwp-slack-slack-search\", \"kwp-legal-meeting-briefing\", \"kwp-data-data-exploration\", \"kwp-design-system\", \"kwp-product-management-competitive-analysis\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0194",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 보안 관점. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "role-security-engineer"
      ],
      "distractor_skills": [
        "role-finance",
        "role-data-scientist",
        "role-trading-expert",
        "role-ux-designer",
        "role-cto",
        "role-hr",
        "role-developer",
        "role-pm",
        "role-cso",
        "role-sales",
        "role-ceo",
        "security-expert",
        "role-dispatcher",
        "compliance-governance",
        "deep-research-pipeline",
        "kwp-operations-risk-assessment",
        "semantic-guard",
        "dependency-auditor",
        "agency-security-engineer",
        "executive-briefing",
        "kwp-operations-compliance-tracking",
        "kwp-design-accessibility-review",
        "agency-legal-compliance-checker",
        "paper-review",
        "sefo-governance",
        "kwp-engineering-code-review",
        "agency-infrastructure-maintainer",
        "gws-recipe-audit-sharing",
        "incident-to-improvement",
        "kwp-engineering-incident-response",
        "kwp-design-design-system-management",
        "ux-expert",
        "kwp-engineering-tech-debt",
        "infra-runbook-generator",
        "kwp-human-resources-people-analytics"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 보안 관점. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"role-security-engineer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"role-finance\", \"role-data-scientist\", \"role-trading-expert\", \"role-ux-designer\", \"role-cto\", \"role-hr\", \"role-developer\", \"role-pm\", \"role-cso\", \"role-sales\", \"role-ceo\", \"security-expert\", \"role-dispatcher\", \"compliance-governance\", \"deep-research-pipeline\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0195",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Tool Evaluator agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-tool-evaluator"
      ],
      "distractor_skills": [
        "agency-app-store-optimizer",
        "agency-workflow-optimizer",
        "agency-xr-immersive-developer",
        "agency-terminal-integration-specialist",
        "agency-test-results-analyzer",
        "agency-api-tester",
        "agency-infrastructure-maintainer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-tiktok-strategist",
        "agency-instagram-curator",
        "agency-behavioral-nudge-engine",
        "agency-brand-guardian",
        "agency-xr-interface-architect",
        "agency-studio-operations",
        "agency-reddit-community-builder",
        "agency-inclusive-visuals-specialist",
        "agency-cultural-intelligence-strategist",
        "agency-ux-architect",
        "agency-rapid-prototyper",
        "agency-visionos-spatial-engineer",
        "agency-senior-developer",
        "agency-performance-benchmarker",
        "agency-xiaohongshu-specialist",
        "agency-legal-compliance-checker",
        "agency-senior-project-manager",
        "agency-macos-spatial-metal-engineer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Tool Evaluator agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-tool-evaluator\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-app-store-optimizer\", \"agency-workflow-optimizer\", \"agency-xr-immersive-developer\", \"agency-terminal-integration-specialist\", \"agency-test-results-analyzer\", \"agency-api-tester\", \"agency-infrastructure-maintainer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-tiktok-strategist\", \"agency-instagram-curator\", \"agency-behavioral-nudge-engine\", \"agency-brand-guardian\", \"agency-xr-interface-architect\", \"agency-studio-operations\", \"agency-reddit-community-builder\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0196",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: references agency-senior-project-manager. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-senior-project-manager"
      ],
      "distractor_skills": [
        "agency-xr-interface-architect",
        "agency-app-store-optimizer",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-terminal-integration-specialist",
        "agency-brand-guardian",
        "agency-visionos-spatial-engineer",
        "agency-senior-developer",
        "agency-tool-evaluator",
        "agency-report-distribution-agent",
        "agency-test-results-analyzer",
        "agency-api-tester",
        "agency-behavioral-nudge-engine",
        "agency-agents-orchestrator",
        "agency-workflow-optimizer",
        "agency-studio-operations",
        "agency-ux-architect",
        "agency-rapid-prototyper",
        "agency-mobile-app-builder",
        "agency-data-consolidation-agent",
        "agency-tiktok-strategist",
        "agency-project-shepherd",
        "agency-macos-spatial-metal-engineer",
        "agency-lsp-index-engineer",
        "agency-experiment-tracker",
        "agency-evidence-collector",
        "agency-reddit-community-builder",
        "agency-instagram-curator",
        "agency-sales-data-extraction-agent",
        "agency-inclusive-visuals-specialist",
        "agency-infrastructure-maintainer",
        "agency-cultural-intelligence-strategist",
        "agency-xiaohongshu-specialist",
        "agency-reality-checker"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: references agency-senior-project-manager. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-senior-project-manager\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-xr-interface-architect\", \"agency-app-store-optimizer\", \"agency-xr-immersive-developer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-terminal-integration-specialist\", \"agency-brand-guardian\", \"agency-visionos-spatial-engineer\", \"agency-senior-developer\", \"agency-tool-evaluator\", \"agency-report-distribution-agent\", \"agency-test-results-analyzer\", \"agency-api-tester\", \"agency-behavioral-nudge-engine\", \"agency-agents-orchestrator\", \"agency-workflow-optimizer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0197",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: compliance-gate. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "compliance-gate"
      ],
      "distractor_skills": [
        "compliance-governance",
        "ci-quality-gate",
        "release-commander",
        "iac-review-agent",
        "ecc-verification-loop",
        "semantic-guard",
        "k8s-manifest-validator",
        "security-expert",
        "helm-validator",
        "terraform-reviewer",
        "dependency-auditor",
        "intent-alignment-tracker",
        "code-review-all",
        "infra-drift-detector",
        "deep-review",
        "ui-suite",
        "agency-data-consolidation-agent",
        "test-suite",
        "workflow-sequential",
        "auto-research",
        "trading-data-quality-checker",
        "ship",
        "kwp-engineering-code-review",
        "commit-to-issue",
        "ai-quality-evaluator"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: compliance-gate. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"compliance-gate\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"compliance-governance\", \"ci-quality-gate\", \"release-commander\", \"iac-review-agent\", \"ecc-verification-loop\", \"semantic-guard\", \"k8s-manifest-validator\", \"security-expert\", \"helm-validator\", \"terraform-reviewer\", \"dependency-auditor\", \"intent-alignment-tracker\", \"code-review-all\", \"infra-drift-detector\", \"deep-review\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0198",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: reviewing AI-generated code. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "ecc-agentic-engineering"
      ],
      "distractor_skills": [
        "ecc-search-first",
        "ecc-autonomous-loops",
        "ecc-coding-standards",
        "ecc-strategic-compact",
        "ecc-eval-harness",
        "ecc-configure",
        "ecc-continuous-learning",
        "ecc-verification-loop",
        "sefo-orchestrator",
        "agency-visionos-spatial-engineer",
        "agency-security-engineer",
        "sp-finish-branch",
        "kwp-finance-close-management",
        "ecc-iterative-retrieval",
        "agency-ux-architect",
        "agency-lsp-index-engineer",
        "sp-subagent-dev",
        "plans",
        "agency-xr-cockpit-interaction-specialist",
        "sp-tdd",
        "paperclip-setup",
        "kwp-productivity-task-management",
        "sp-request-review",
        "kwp-product-management-roadmap-management",
        "kwp-operations-resource-planning",
        "hf-jobs",
        "agency-rapid-prototyper",
        "sp-receive-review",
        "paperclip-tasks",
        "agency-senior-project-manager",
        "agency-macos-spatial-metal-engineer",
        "sp-brainstorming",
        "kwp-cowork-plugin-customizer",
        "agency-xr-interface-architect",
        "agency-devops-automator",
        "agency-ai-engineer",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-marketing-brand-voice"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: reviewing AI-generated code. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"ecc-agentic-engineering\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"ecc-search-first\", \"ecc-autonomous-loops\", \"ecc-coding-standards\", \"ecc-strategic-compact\", \"ecc-eval-harness\", \"ecc-configure\", \"ecc-continuous-learning\", \"ecc-verification-loop\", \"sefo-orchestrator\", \"agency-visionos-spatial-engineer\", \"agency-security-engineer\", \"sp-finish-branch\", \"kwp-finance-close-management\", \"ecc-iterative-retrieval\", \"agency-ux-architect\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0199",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Xiaohongshu Specialist agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-xiaohongshu-specialist"
      ],
      "distractor_skills": [
        "agency-reddit-community-builder",
        "agency-instagram-curator",
        "agency-tiktok-strategist",
        "agency-twitter-engager",
        "agency-zhihu-strategist",
        "agency-app-store-optimizer",
        "agency-wechat-official-account-manager",
        "agency-growth-hacker",
        "agency-tool-evaluator",
        "agency-visual-storyteller",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-test-results-analyzer",
        "agency-api-tester",
        "agency-terminal-integration-specialist",
        "agency-brand-guardian",
        "agency-xr-interface-architect",
        "agency-workflow-optimizer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Xiaohongshu Specialist agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-xiaohongshu-specialist\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-reddit-community-builder\", \"agency-instagram-curator\", \"agency-tiktok-strategist\", \"agency-twitter-engager\", \"agency-zhihu-strategist\", \"agency-app-store-optimizer\", \"agency-wechat-official-account-manager\", \"agency-growth-hacker\", \"agency-tool-evaluator\", \"agency-visual-storyteller\", \"agency-xr-immersive-developer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-test-results-analyzer\", \"agency-api-tester\", \"agency-terminal-integration-specialist\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0200",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: references agency-data-consolidation-agent. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-data-consolidation-agent"
      ],
      "distractor_skills": [
        "agency-sales-data-extraction-agent",
        "agency-xr-immersive-developer",
        "agency-report-distribution-agent",
        "agency-xr-interface-architect",
        "agency-xr-cockpit-interaction-specialist",
        "agency-behavioral-nudge-engine",
        "agency-agents-orchestrator",
        "agency-terminal-integration-specialist",
        "agency-brand-guardian",
        "agency-ux-architect",
        "agency-mobile-app-builder",
        "agency-visionos-spatial-engineer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: references agency-data-consolidation-agent. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-data-consolidation-agent\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-sales-data-extraction-agent\", \"agency-xr-immersive-developer\", \"agency-report-distribution-agent\", \"agency-xr-interface-architect\", \"agency-xr-cockpit-interaction-specialist\", \"agency-behavioral-nudge-engine\", \"agency-agents-orchestrator\", \"agency-terminal-integration-specialist\", \"agency-brand-guardian\", \"agency-ux-architect\", \"agency-mobile-app-builder\", \"agency-visionos-spatial-engineer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0201",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: generate daily report. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "tab-analysis-run"
      ],
      "distractor_skills": [
        "tab-patterns",
        "tab-genai-features",
        "tab-event-detect",
        "tab-report-generate",
        "tab-screening",
        "tab-llm-agents",
        "tab-strategy-comparison",
        "tab-market-breadth",
        "tab-turtle-refresh",
        "tab-dualma-refresh",
        "tab-kiwoom",
        "tab-glossary",
        "tab-dashboard-summary",
        "tab-news-fetch",
        "tab-fundamental-sync",
        "tab-market-environment",
        "tab-stock-sync",
        "tab-bollinger-refresh",
        "tab-technical-analysis",
        "tab-hot-stock-discovery",
        "alphaear-reporter",
        "hf-jobs",
        "ai-quality-evaluator",
        "trading-sector-analyst",
        "kwp-finance-financial-statements",
        "trading-us-stock-analysis",
        "weekly-status-report",
        "trading-breadth-chart-analyst",
        "morning-ship",
        "trading-theme-detector"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: generate daily report. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"tab-analysis-run\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"tab-patterns\", \"tab-genai-features\", \"tab-event-detect\", \"tab-report-generate\", \"tab-screening\", \"tab-llm-agents\", \"tab-strategy-comparison\", \"tab-market-breadth\", \"tab-turtle-refresh\", \"tab-dualma-refresh\", \"tab-kiwoom\", \"tab-glossary\", \"tab-dashboard-summary\", \"tab-news-fetch\", \"tab-fundamental-sync\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0202",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: more recent 3. Present both perspectives with context 4. Recommend how to resolve the discrepancy 5. If going to a customer: use the most conservative/cautious answer until resolved ### Synthesis Structure ``` **Direct Answer:** [Bottom-line answer — lead with this] **Confidence:** [High / Medium / Low] **Supporting Evidence:** - [Source 1]: [What it says] - [Source 2]: [What it says — corroborates. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-customer-support-customer-research"
      ],
      "distractor_skills": [
        "kwp-slack-slack-search",
        "kwp-sales-call-prep",
        "kwp-legal-meeting-briefing",
        "kwp-common-room-account-research",
        "kwp-marketing-competitive-analysis",
        "kwp-customer-support-response-drafting",
        "kwp-legal-contract-review",
        "kwp-human-resources-interview-prep",
        "kwp-customer-support-escalation",
        "kwp-enterprise-search-source-management",
        "kwp-customer-support-knowledge-management",
        "kwp-design-system",
        "kwp-marketing-content-creation",
        "kwp-sales-account-research",
        "kwp-product-management-user-research-synthesis",
        "kwp-marketing-campaign-planning",
        "kwp-common-room-contact-research",
        "kwp-marketing-brand-voice",
        "kwp-data-data-context-extractor",
        "kwp-sales-draft-outreach",
        "kwp-product-management-metrics-tracking",
        "kwp-data-data-validation",
        "kwp-operations-vendor-management",
        "kwp-legal-canned-responses",
        "kwp-finance-audit-support",
        "kwp-enterprise-search-knowledge-synthesis",
        "kwp-data-interactive-dashboard-builder",
        "kwp-finance-close-management",
        "kwp-design-design-handoff",
        "kwp-sales-daily-briefing",
        "kwp-marketing-performance-analytics",
        "kwp-finance-journal-entry-prep",
        "kwp-product-management-competitive-analysis",
        "kwp-human-resources-employee-handbook",
        "kwp-finance-reconciliation",
        "kwp-cowork-plugin-customizer",
        "kwp-apollo-prospect",
        "kwp-operations-resource-planning",
        "kwp-data-data-visualization",
        "kwp-human-resources-people-analytics",
        "kwp-data-sql-queries",
        "kwp-sales-competitive-intelligence",
        "kwp-product-management-stakeholder-comms",
        "kwp-human-resources-recruiting-pipeline"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: more recent 3. Present both perspectives with context 4. Recommend how to resolve the discrepancy 5. If going to a customer: use the most conservative/cautious answer until resolved ### Synthesis Structure ``` **Direct Answer:** [Bottom-line answer — lead with this] **Confidence:** [High / Medium / Low] **Supporting Evidence:** - [Source 1]: [What it says] - [Source 2]: [What it says — corroborates. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-customer-support-customer-research\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-slack-slack-search\", \"kwp-sales-call-prep\", \"kwp-legal-meeting-briefing\", \"kwp-common-room-account-research\", \"kwp-marketing-competitive-analysis\", \"kwp-customer-support-response-drafting\", \"kwp-legal-contract-review\", \"kwp-human-resources-interview-prep\", \"kwp-customer-support-escalation\", \"kwp-enterprise-search-source-management\", \"kwp-customer-support-knowledge-management\", \"kwp-design-system\", \"kwp-marketing-content-creation\", \"kwp-sales-account-research\", \"kwp-product-management-user-research-synthesis\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0203",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: responding to data subject access. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-legal-compliance"
      ],
      "distractor_skills": [
        "kwp-legal-canned-responses",
        "kwp-legal-legal-risk-assessment",
        "kwp-legal-contract-review",
        "kwp-marketing-brand-voice",
        "kwp-customer-support-response-drafting",
        "kwp-data-data-validation",
        "kwp-legal-nda-triage",
        "kwp-operations-vendor-management",
        "kwp-slack-slack-search",
        "kwp-legal-meeting-briefing",
        "kwp-data-data-exploration",
        "kwp-design-system",
        "kwp-product-management-competitive-analysis",
        "kwp-finance-reconciliation",
        "kwp-human-resources-people-analytics",
        "kwp-data-data-context-extractor",
        "kwp-data-statistical-analysis",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-human-resources-interview-prep",
        "kwp-product-management-feature-spec",
        "kwp-operations-compliance-tracking",
        "kwp-finance-audit-support",
        "kwp-data-interactive-dashboard-builder",
        "kwp-customer-support-escalation",
        "kwp-finance-close-management",
        "kwp-product-management-user-research-synthesis",
        "kwp-marketing-performance-analytics",
        "kwp-customer-support-ticket-triage",
        "kwp-cowork-plugin-customizer",
        "kwp-data-data-visualization",
        "kwp-marketing-competitive-analysis",
        "kwp-common-room-contact-research",
        "kwp-common-room-account-research",
        "kwp-product-management-stakeholder-comms",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-org-planning"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: responding to data subject access. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-legal-compliance\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-legal-canned-responses\", \"kwp-legal-legal-risk-assessment\", \"kwp-legal-contract-review\", \"kwp-marketing-brand-voice\", \"kwp-customer-support-response-drafting\", \"kwp-data-data-validation\", \"kwp-legal-nda-triage\", \"kwp-operations-vendor-management\", \"kwp-slack-slack-search\", \"kwp-legal-meeting-briefing\", \"kwp-data-data-exploration\", \"kwp-design-system\", \"kwp-product-management-competitive-analysis\", \"kwp-finance-reconciliation\", \"kwp-human-resources-people-analytics\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0204",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: re-evaluate. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "workflow-eval-opt"
      ],
      "distractor_skills": [
        "workflow-sequential",
        "workflow-parallel",
        "workflow-miner",
        "skill-optimizer",
        "kwp-operations-process-optimization",
        "ralph-loop",
        "kwp-design-accessibility-review",
        "kwp-design-user-research",
        "sentence-polisher",
        "agency-test-results-analyzer",
        "role-data-scientist",
        "sefo-orchestrator",
        "issue-to-code",
        "prompt-transformer",
        "paperclip-control",
        "qa-test-expert",
        "gws-calendar",
        "kwp-operations-vendor-management",
        "intent-alignment-tracker",
        "ecc-eval-harness",
        "email-auto-reply",
        "paperclip-tasks",
        "skill-autoimprove",
        "autoskill-judge",
        "hf-evaluation",
        "role-developer",
        "e2e-testing",
        "evals-skills",
        "sefo-federation",
        "transcribee",
        "paperclip-setup",
        "compliance-gate",
        "paperclip-agents",
        "kwp-engineering-code-review",
        "gws-workspace",
        "autoskill-merger",
        "trading-trade-hypothesis-ideator",
        "kwp-productivity-task-management",
        "kwp-design-ux-writing",
        "trading-finviz-screener",
        "ai-quality-evaluator",
        "kwp-operations-resource-planning",
        "kwp-design-design-critique",
        "office-template-enforcer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: re-evaluate. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"workflow-eval-opt\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"workflow-sequential\", \"workflow-parallel\", \"workflow-miner\", \"skill-optimizer\", \"kwp-operations-process-optimization\", \"ralph-loop\", \"kwp-design-accessibility-review\", \"kwp-design-user-research\", \"sentence-polisher\", \"agency-test-results-analyzer\", \"role-data-scientist\", \"sefo-orchestrator\", \"issue-to-code\", \"prompt-transformer\", \"paperclip-control\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0205",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Memory Management. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-productivity-memory-management"
      ],
      "distractor_skills": [
        "kwp-design-system",
        "kwp-slack-slack-search",
        "kwp-data-data-context-extractor",
        "kwp-customer-support-knowledge-management",
        "kwp-apollo-enrich-lead",
        "kwp-cowork-plugin-customizer",
        "kwp-apollo-prospect",
        "kwp-slack-slack-messaging",
        "kwp-human-resources-people-analytics",
        "kwp-customer-support-response-drafting",
        "kwp-apollo-sequence-load",
        "kwp-operations-change-management",
        "kwp-legal-canned-responses",
        "kwp-finance-audit-support",
        "kwp-data-interactive-dashboard-builder",
        "kwp-customer-support-escalation",
        "kwp-cowork-cowork-plugin-customizer",
        "kwp-legal-compliance",
        "kwp-enterprise-search-search-strategy",
        "kwp-customer-support-customer-research",
        "kwp-product-management-roadmap-management",
        "kwp-operations-vendor-management",
        "kwp-marketing-performance-analytics",
        "kwp-human-resources-interview-prep",
        "kwp-data-data-visualization",
        "kwp-finance-close-management",
        "kwp-enterprise-search-source-management",
        "kwp-legal-nda-triage",
        "kwp-product-management-competitive-analysis",
        "kwp-human-resources-employee-handbook",
        "kwp-finance-reconciliation",
        "kwp-customer-support-ticket-triage",
        "kwp-marketing-competitive-analysis",
        "kwp-marketing-brand-voice",
        "kwp-sales-draft-outreach",
        "kwp-sales-competitive-intelligence",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-org-planning",
        "kwp-data-statistical-analysis",
        "kwp-data-data-validation",
        "kwp-finance-variance-analysis",
        "kwp-finance-financial-statements",
        "kwp-common-room-prospect",
        "kwp-common-room-contact-research",
        "kwp-common-room-account-research",
        "kwp-product-management-feature-spec",
        "kwp-operations-process-optimization"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Memory Management. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-productivity-memory-management\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-design-system\", \"kwp-slack-slack-search\", \"kwp-data-data-context-extractor\", \"kwp-customer-support-knowledge-management\", \"kwp-apollo-enrich-lead\", \"kwp-cowork-plugin-customizer\", \"kwp-apollo-prospect\", \"kwp-slack-slack-messaging\", \"kwp-human-resources-people-analytics\", \"kwp-customer-support-response-drafting\", \"kwp-apollo-sequence-load\", \"kwp-operations-change-management\", \"kwp-legal-canned-responses\", \"kwp-finance-audit-support\", \"kwp-data-interactive-dashboard-builder\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0206",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: any request to simulate public opinion dynamics. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "mirofish-opinion-sim"
      ],
      "distractor_skills": [
        "mirofish-financial-sim",
        "mirofish",
        "mirofish-graph-explorer",
        "twitter-timeline-to-slack",
        "agency-social-media-strategist",
        "trading-market-news-analyst",
        "slack-agent",
        "cursor-automations",
        "trading-sector-analyst",
        "paperclip-setup",
        "standup-digest",
        "alphaear-deepear-lite",
        "alphaear-sentiment",
        "daily-am-orchestrator",
        "agency-wechat-official-account-manager",
        "daiso-mcp",
        "notebooklm-research",
        "daily-strategy-post",
        "agency-lsp-index-engineer",
        "agency-devops-automator",
        "trading-market-breadth-analyzer",
        "helm-validator",
        "skill-optimizer",
        "sefo-orchestrator",
        "terraform-reviewer",
        "paperclip-agents",
        "dependency-auditor",
        "paper-review",
        "k8s-manifest-validator",
        "diagnose",
        "auto-research-distribute",
        "agency-trend-researcher",
        "agency-behavioral-nudge-engine",
        "trading-options-theta",
        "hf-trending-intelligence",
        "email-research-dispatcher",
        "deep-review",
        "intent-alignment-tracker",
        "agency-rapid-prototyper",
        "role-dispatcher",
        "agency-twitter-engager",
        "agency-developer-advocate",
        "paperclip-tasks",
        "domain-commit",
        "autoskill-extractor",
        "alphaear-search",
        "alphaear-news",
        "agency-tool-evaluator",
        "gws-chat",
        "nlm-deep-learn"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: any request to simulate public opinion dynamics. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"mirofish-opinion-sim\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"mirofish-financial-sim\", \"mirofish\", \"mirofish-graph-explorer\", \"twitter-timeline-to-slack\", \"agency-social-media-strategist\", \"trading-market-news-analyst\", \"slack-agent\", \"cursor-automations\", \"trading-sector-analyst\", \"paperclip-setup\", \"standup-digest\", \"alphaear-deepear-lite\", \"alphaear-sentiment\", \"daily-am-orchestrator\", \"agency-wechat-official-account-manager\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0207",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: GSE/GSM/SRR accessions. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-bio-research-nextflow-development"
      ],
      "distractor_skills": [
        "kwp-data-data-context-extractor",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-human-resources-people-analytics",
        "kwp-data-statistical-analysis",
        "kwp-data-data-exploration",
        "kwp-data-data-validation",
        "kwp-finance-variance-analysis",
        "kwp-bio-research-scvi-tools",
        "kwp-product-management-user-research-synthesis",
        "kwp-marketing-performance-analytics",
        "kwp-finance-close-management",
        "kwp-design-design-handoff",
        "kwp-product-management-competitive-analysis",
        "kwp-finance-reconciliation",
        "kwp-marketing-competitive-analysis",
        "kwp-design-system",
        "kwp-common-room-contact-research",
        "kwp-common-room-account-research",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-finance-financial-statements",
        "kwp-slack-slack-search",
        "kwp-legal-canned-responses",
        "kwp-data-interactive-dashboard-builder",
        "kwp-legal-compliance",
        "kwp-customer-support-knowledge-management",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-legal-contract-review",
        "kwp-data-data-visualization",
        "kwp-legal-nda-triage",
        "kwp-data-sql-queries",
        "kwp-cowork-plugin-customizer",
        "kwp-cowork-create-plugin",
        "kwp-bio-research-instrument-data-to-allotrope",
        "kwp-marketing-brand-voice",
        "kwp-customer-support-response-drafting",
        "kwp-product-management-stakeholder-comms",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-org-planning",
        "kwp-product-management-feature-spec",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: GSE/GSM/SRR accessions. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-bio-research-nextflow-development\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-data-data-context-extractor\", \"kwp-bio-research-single-cell-rna-qc\", \"kwp-human-resources-people-analytics\", \"kwp-data-statistical-analysis\", \"kwp-data-data-exploration\", \"kwp-data-data-validation\", \"kwp-finance-variance-analysis\", \"kwp-bio-research-scvi-tools\", \"kwp-product-management-user-research-synthesis\", \"kwp-marketing-performance-analytics\", \"kwp-finance-close-management\", \"kwp-design-design-handoff\", \"kwp-product-management-competitive-analysis\", \"kwp-finance-reconciliation\", \"kwp-marketing-competitive-analysis\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0208",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: evaluate RAG. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "evals-skills"
      ],
      "distractor_skills": [
        "e2e-testing",
        "skill-optimizer",
        "ai-quality-evaluator",
        "pm-data-analytics",
        "compliance-governance",
        "backend-expert",
        "ux-expert",
        "semantic-guard",
        "agency-test-results-analyzer",
        "performance-profiler",
        "frontend-expert",
        "qa-test-expert",
        "i18n-sync",
        "intent-alignment-tracker",
        "ai-workflow-integrator",
        "kwp-engineering-tech-debt",
        "agency-tool-evaluator",
        "agency-xr-immersive-developer",
        "ci-quality-gate",
        "agency-inclusive-visuals-specialist",
        "test-suite",
        "deep-review",
        "kwp-engineering-code-review",
        "agency-ai-engineer",
        "kwp-design-ux-writing",
        "ui-suite",
        "agency-xr-interface-architect",
        "agency-ux-researcher",
        "agency-data-consolidation-agent",
        "kwp-human-resources-people-analytics",
        "diagnose",
        "agency-xr-cockpit-interaction-specialist"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: evaluate RAG. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"evals-skills\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"e2e-testing\", \"skill-optimizer\", \"ai-quality-evaluator\", \"pm-data-analytics\", \"compliance-governance\", \"backend-expert\", \"ux-expert\", \"semantic-guard\", \"agency-test-results-analyzer\", \"performance-profiler\", \"frontend-expert\", \"qa-test-expert\", \"i18n-sync\", \"intent-alignment-tracker\", \"ai-workflow-integrator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0209",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Org Planning. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-human-resources-org-planning"
      ],
      "distractor_skills": [
        "kwp-human-resources-interview-prep",
        "kwp-operations-resource-planning",
        "kwp-operations-change-management",
        "kwp-operations-vendor-management",
        "kwp-human-resources-people-analytics",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-sales-account-research",
        "kwp-design-system",
        "kwp-finance-close-management",
        "kwp-design-design-handoff",
        "kwp-operations-process-optimization",
        "kwp-marketing-campaign-planning",
        "kwp-product-management-metrics-tracking",
        "kwp-engineering-testing-strategy",
        "kwp-slack-slack-search",
        "kwp-sales-call-prep",
        "kwp-product-management-feature-spec",
        "kwp-customer-support-escalation",
        "kwp-apollo-enrich-lead",
        "kwp-product-management-competitive-analysis",
        "kwp-human-resources-employee-handbook",
        "kwp-common-room-account-research",
        "kwp-sales-daily-briefing"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Org Planning. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-human-resources-org-planning\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-human-resources-interview-prep\", \"kwp-operations-resource-planning\", \"kwp-operations-change-management\", \"kwp-operations-vendor-management\", \"kwp-human-resources-people-analytics\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-sales-account-research\", \"kwp-design-system\", \"kwp-finance-close-management\", \"kwp-design-design-handoff\", \"kwp-operations-process-optimization\", \"kwp-marketing-campaign-planning\", \"kwp-product-management-metrics-tracking\", \"kwp-engineering-testing-strategy\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0210",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Brand Voice Enforcement. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-brand-voice-brand-voice-enforcement"
      ],
      "distractor_skills": [
        "kwp-brand-voice-discover-brand",
        "kwp-brand-voice-guideline-generation",
        "kwp-marketing-brand-voice",
        "kwp-common-room-compose-outreach",
        "kwp-cowork-create-plugin",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-design-ux-writing",
        "kwp-design-system",
        "kwp-sales-draft-outreach",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-marketing-content-creation",
        "kwp-data-data-context-extractor",
        "kwp-productivity-task-management",
        "kwp-customer-support-knowledge-management",
        "kwp-design-design-handoff",
        "kwp-legal-contract-review",
        "kwp-bio-research-scientific-problem-selection",
        "kwp-common-room-weekly-prep-brief",
        "kwp-cowork-create-cowork-plugin",
        "kwp-engineering-documentation",
        "kwp-human-resources-people-analytics",
        "kwp-customer-support-response-drafting",
        "kwp-sales-call-prep",
        "kwp-human-resources-org-planning",
        "kwp-common-room-contact-research",
        "kwp-slack-slack-search",
        "kwp-product-management-feature-spec",
        "kwp-operations-process-optimization"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Brand Voice Enforcement. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-brand-voice-brand-voice-enforcement\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-brand-voice-discover-brand\", \"kwp-brand-voice-guideline-generation\", \"kwp-marketing-brand-voice\", \"kwp-common-room-compose-outreach\", \"kwp-cowork-create-plugin\", \"kwp-operations-vendor-management\", \"kwp-human-resources-interview-prep\", \"kwp-design-ux-writing\", \"kwp-design-system\", \"kwp-sales-draft-outreach\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-marketing-content-creation\", \"kwp-data-data-context-extractor\", \"kwp-productivity-task-management\", \"kwp-customer-support-knowledge-management\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0211",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 런북 생성. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "infra-runbook-generator"
      ],
      "distractor_skills": [
        "infra-drift-detector",
        "sre-devops-expert",
        "helm-validator",
        "docs-tutor-setup",
        "kwp-engineering-documentation",
        "k8s-manifest-validator",
        "terraform-reviewer",
        "incident-to-improvement",
        "iac-review-agent",
        "role-cto",
        "agency-devops-automator",
        "kwp-engineering-incident-response",
        "service-health-doctor",
        "local-dev-runner",
        "skill-guide-generator",
        "prompt-transformer",
        "i18n-sync",
        "role-developer",
        "pr-review-captain",
        "gws-recipe-doc-from-template",
        "ppt-template-engine",
        "onboarding-accelerator",
        "technical-writer",
        "role-security-engineer",
        "office-template-enforcer",
        "weekly-status-report"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 런북 생성. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"infra-runbook-generator\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"infra-drift-detector\", \"sre-devops-expert\", \"helm-validator\", \"docs-tutor-setup\", \"kwp-engineering-documentation\", \"k8s-manifest-validator\", \"terraform-reviewer\", \"incident-to-improvement\", \"iac-review-agent\", \"role-cto\", \"agency-devops-automator\", \"kwp-engineering-incident-response\", \"service-health-doctor\", \"local-dev-runner\", \"skill-guide-generator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0212",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: auto-reply to emails. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "email-auto-reply"
      ],
      "distractor_skills": [
        "gws-email-reply",
        "proactive-meeting-scheduler",
        "email-research-dispatcher",
        "feedback-meeting-scheduler",
        "gws-gmail",
        "knowledge-daily-aggregator",
        "gws-calendar",
        "smart-meeting-scheduler",
        "calendar-daily-briefing",
        "issue-to-code"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: auto-reply to emails. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"email-auto-reply\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"gws-email-reply\", \"proactive-meeting-scheduler\", \"email-research-dispatcher\", \"feedback-meeting-scheduler\", \"gws-gmail\", \"knowledge-daily-aggregator\", \"gws-calendar\", \"smart-meeting-scheduler\", \"calendar-daily-briefing\", \"issue-to-code\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0213",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: classifying control deficiencies. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-finance-audit-support"
      ],
      "distractor_skills": [
        "kwp-data-data-validation",
        "kwp-human-resources-interview-prep",
        "kwp-finance-variance-analysis",
        "kwp-finance-financial-statements",
        "kwp-finance-close-management",
        "kwp-legal-meeting-briefing",
        "kwp-product-management-competitive-analysis",
        "kwp-finance-journal-entry-prep",
        "kwp-operations-vendor-management",
        "kwp-data-interactive-dashboard-builder",
        "kwp-customer-support-escalation",
        "kwp-slack-slack-search",
        "kwp-design-design-handoff",
        "kwp-design-system",
        "kwp-finance-reconciliation",
        "kwp-legal-legal-risk-assessment",
        "kwp-marketing-competitive-analysis",
        "kwp-marketing-brand-voice"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: classifying control deficiencies. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-finance-audit-support\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-data-data-validation\", \"kwp-human-resources-interview-prep\", \"kwp-finance-variance-analysis\", \"kwp-finance-financial-statements\", \"kwp-finance-close-management\", \"kwp-legal-meeting-briefing\", \"kwp-product-management-competitive-analysis\", \"kwp-finance-journal-entry-prep\", \"kwp-operations-vendor-management\", \"kwp-data-interactive-dashboard-builder\", \"kwp-customer-support-escalation\", \"kwp-slack-slack-search\", \"kwp-design-design-handoff\", \"kwp-design-system\", \"kwp-finance-reconciliation\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0214",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: recovering from context window limits. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "ecc-autonomous-loops"
      ],
      "distractor_skills": [
        "ecc-strategic-compact",
        "ralph-loop",
        "ecc-agentic-engineering",
        "ecc-continuous-learning",
        "ecc-eval-harness",
        "ecc-search-first",
        "ecc-iterative-retrieval",
        "ai-workflow-integrator",
        "ecc-configure",
        "ecc-coding-standards",
        "ecc-verification-loop",
        "agency-xr-cockpit-interaction-specialist",
        "deep-research-pipeline",
        "ui-suite",
        "pipeline-builder",
        "mission-control copy",
        "planning-weekly-pulse",
        "mission-control",
        "agency-devops-automator",
        "tab-llm-agents",
        "pm-data-analytics",
        "alphaear-search",
        "agency-lsp-index-engineer",
        "hf-auto-finetune-planner",
        "kwp-design-system",
        "automation-strategist",
        "kwp-product-management-metrics-tracking",
        "agency-xr-immersive-developer",
        "skill-optimizer",
        "workflow-miner",
        "sefo-orchestrator",
        "gws-workflows",
        "agency-reddit-community-builder",
        "agency-instagram-curator",
        "agency-data-engineer",
        "system-thinker",
        "kwp-customer-support-knowledge-management",
        "kwp-customer-support-customer-research",
        "sp-core",
        "kwp-human-resources-interview-prep"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: recovering from context window limits. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"ecc-autonomous-loops\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"ecc-strategic-compact\", \"ralph-loop\", \"ecc-agentic-engineering\", \"ecc-continuous-learning\", \"ecc-eval-harness\", \"ecc-search-first\", \"ecc-iterative-retrieval\", \"ai-workflow-integrator\", \"ecc-configure\", \"ecc-coding-standards\", \"ecc-verification-loop\", \"agency-xr-cockpit-interaction-specialist\", \"deep-research-pipeline\", \"ui-suite\", \"pipeline-builder\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0215",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: references agency-analytics-reporter. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-analytics-reporter"
      ],
      "distractor_skills": [
        "agency-data-analytics-reporter",
        "agency-ux-researcher",
        "agency-finance-tracker",
        "agency-data-consolidation-agent",
        "agency-feedback-synthesizer",
        "agency-brand-guardian",
        "agency-xr-immersive-developer",
        "agency-growth-hacker",
        "agency-trend-researcher",
        "agency-experiment-tracker",
        "agency-data-engineer",
        "agency-tool-evaluator",
        "agency-performance-benchmarker",
        "agency-app-store-optimizer",
        "agency-xr-interface-architect",
        "agency-devops-automator",
        "agency-sprint-prioritizer",
        "agency-test-results-analyzer",
        "agency-sales-data-extraction-agent",
        "agency-legal-compliance-checker",
        "agency-workflow-optimizer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-frontend-developer",
        "agency-whimsy-injector",
        "agency-support-responder",
        "agency-ux-architect",
        "agency-terminal-integration-specialist",
        "agency-lsp-index-engineer",
        "agency-visual-storyteller",
        "agency-visionos-spatial-engineer",
        "agency-ai-engineer",
        "agency-infrastructure-maintainer",
        "agency-executive-summary-generator",
        "agency-inclusive-visuals-specialist",
        "agency-api-tester",
        "agency-security-engineer",
        "agency-xiaohongshu-specialist",
        "agency-social-media-strategist",
        "agency-content-creator",
        "agency-ui-designer",
        "agency-behavioral-nudge-engine"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: references agency-analytics-reporter. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-analytics-reporter\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-data-analytics-reporter\", \"agency-ux-researcher\", \"agency-finance-tracker\", \"agency-data-consolidation-agent\", \"agency-feedback-synthesizer\", \"agency-brand-guardian\", \"agency-xr-immersive-developer\", \"agency-growth-hacker\", \"agency-trend-researcher\", \"agency-experiment-tracker\", \"agency-data-engineer\", \"agency-tool-evaluator\", \"agency-performance-benchmarker\", \"agency-app-store-optimizer\", \"agency-xr-interface-architect\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0216",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: switching task phases. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "ecc-strategic-compact"
      ],
      "distractor_skills": [
        "ecc-autonomous-loops",
        "ecc-search-first",
        "ecc-eval-harness",
        "ecc-verification-loop",
        "ecc-agentic-engineering",
        "ecc-configure",
        "ecc-coding-standards",
        "ecc-continuous-learning",
        "ecc-iterative-retrieval",
        "sp-writing-skills"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: switching task phases. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"ecc-strategic-compact\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"ecc-autonomous-loops\", \"ecc-search-first\", \"ecc-eval-harness\", \"ecc-verification-loop\", \"ecc-agentic-engineering\", \"ecc-configure\", \"ecc-coding-standards\", \"ecc-continuous-learning\", \"ecc-iterative-retrieval\", \"sp-writing-skills\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0217",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: check health. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "service-health-doctor"
      ],
      "distractor_skills": [
        "local-dev-runner",
        "backend-expert",
        "agency-devops-automator",
        "sre-devops-expert",
        "kwp-engineering-code-review",
        "performance-profiler",
        "ci-quality-gate",
        "infra-drift-detector",
        "deep-review",
        "db-expert",
        "agency-backend-architect",
        "agency-performance-benchmarker",
        "commit-to-issue",
        "kwp-design-accessibility-review",
        "ui-suite",
        "kwp-engineering-system-design",
        "i18n-sync",
        "dependency-auditor",
        "kwp-engineering-incident-response",
        "kwp-legal-nda-triage",
        "infra-runbook-generator",
        "agency-brand-guardian",
        "skill-optimizer",
        "gws-workspace",
        "kwp-sync",
        "alphaear-sentiment",
        "frontend-expert",
        "qa-test-expert"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: check health. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"service-health-doctor\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"local-dev-runner\", \"backend-expert\", \"agency-devops-automator\", \"sre-devops-expert\", \"kwp-engineering-code-review\", \"performance-profiler\", \"ci-quality-gate\", \"infra-drift-detector\", \"deep-review\", \"db-expert\", \"agency-backend-architect\", \"agency-performance-benchmarker\", \"commit-to-issue\", \"kwp-design-accessibility-review\", \"ui-suite\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0218",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 글 다듬어줘. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "sentence-polisher"
      ],
      "distractor_skills": [
        "kwp-brand-voice-brand-voice-enforcement",
        "pm-toolkit",
        "role-data-scientist",
        "gws-email-reply",
        "mission-control",
        "qa-test-expert",
        "pm-marketing-growth",
        "release-commander",
        "design-architect",
        "gws-calendar"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 글 다듬어줘. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"sentence-polisher\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-brand-voice-brand-voice-enforcement\", \"pm-toolkit\", \"role-data-scientist\", \"gws-email-reply\", \"mission-control\", \"qa-test-expert\", \"pm-marketing-growth\", \"release-commander\", \"design-architect\", \"gws-calendar\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0219",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: querying existing datasets with SQL. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "hf-datasets"
      ],
      "distractor_skills": [
        "hf-cli",
        "hf-dataset-viewer",
        "hf-jobs",
        "hf-collections",
        "hf-models",
        "hf-discussions",
        "hf-model-trainer",
        "hf-evaluation",
        "hf-trackio",
        "hf-repos",
        "hf-buckets",
        "hf-hub",
        "hf-endpoints",
        "hf-auto-finetune-planner",
        "hf-model-arena",
        "hf-papers",
        "hf-spaces",
        "hf-trending-intelligence",
        "stock-csv-downloader",
        "alphaear-news",
        "daily-stock-check",
        "weekly-stock-update"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: querying existing datasets with SQL. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"hf-datasets\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"hf-cli\", \"hf-dataset-viewer\", \"hf-jobs\", \"hf-collections\", \"hf-models\", \"hf-discussions\", \"hf-model-trainer\", \"hf-evaluation\", \"hf-trackio\", \"hf-repos\", \"hf-buckets\", \"hf-hub\", \"hf-endpoints\", \"hf-auto-finetune-planner\", \"hf-model-arena\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0220",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: tab-market-breadth. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "tab-market-breadth"
      ],
      "distractor_skills": [
        "tab-market-environment",
        "tab-turtle-refresh",
        "tab-dualma-refresh",
        "tab-bollinger-refresh",
        "tab-news-fetch",
        "tab-llm-agents",
        "tab-dashboard-summary",
        "tab-strategy-comparison",
        "tab-report-generate",
        "tab-patterns",
        "tab-technical-analysis",
        "tab-fundamental-sync",
        "tab-glossary",
        "tab-stock-sync",
        "tab-hot-stock-discovery",
        "tab-analysis-run",
        "tab-kiwoom",
        "tab-screening",
        "tab-genai-features",
        "tab-event-detect",
        "daily-stock-check",
        "trading-market-environment-analysis",
        "stock-csv-downloader",
        "alphaear-news",
        "trading-breadth-chart-analyst",
        "trading-sector-analyst",
        "trading-market-breadth-analyzer",
        "trading-technical-analyst",
        "weekly-stock-update",
        "kwp-data-statistical-analysis",
        "trading-uptrend-analyzer",
        "trading-theme-detector",
        "today",
        "hf-trackio",
        "hf-jobs",
        "notebooklm",
        "alphaear-search",
        "morning-ship",
        "trading-market-top-detector",
        "alphaear-predictor",
        "kwp-human-resources-people-analytics",
        "trading-us-stock-analysis",
        "morning",
        "kwp-data-data-validation",
        "gws-drive"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: tab-market-breadth. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"tab-market-breadth\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"tab-market-environment\", \"tab-turtle-refresh\", \"tab-dualma-refresh\", \"tab-bollinger-refresh\", \"tab-news-fetch\", \"tab-llm-agents\", \"tab-dashboard-summary\", \"tab-strategy-comparison\", \"tab-report-generate\", \"tab-patterns\", \"tab-technical-analysis\", \"tab-fundamental-sync\", \"tab-glossary\", \"tab-stock-sync\", \"tab-hot-stock-discovery\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0221",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: get Space info. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "hf-spaces"
      ],
      "distractor_skills": [
        "hf-models",
        "hf-endpoints",
        "hf-hub",
        "hf-collections",
        "hf-discussions",
        "hf-buckets",
        "hf-papers",
        "hf-repos",
        "hf-jobs",
        "hf-cli",
        "hf-model-trainer",
        "hf-trackio",
        "hf-dataset-viewer",
        "anthropic-claude-api",
        "gws-chat",
        "hf-model-arena",
        "hf-auto-finetune-planner",
        "gws-drive",
        "hf-evaluation",
        "gws-docs",
        "gws-sheets",
        "hf-datasets",
        "sre-devops-expert",
        "agency-brand-guardian",
        "hf-trending-intelligence"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: get Space info. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"hf-spaces\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"hf-models\", \"hf-endpoints\", \"hf-hub\", \"hf-collections\", \"hf-discussions\", \"hf-buckets\", \"hf-papers\", \"hf-repos\", \"hf-jobs\", \"hf-cli\", \"hf-model-trainer\", \"hf-trackio\", \"hf-dataset-viewer\", \"anthropic-claude-api\", \"gws-chat\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0222",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: fill a PPT template. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "ppt-template-engine"
      ],
      "distractor_skills": [
        "docx-template-engine",
        "office-template-enforcer",
        "alphaxiv-paper-lookup",
        "jira-linear-adapter",
        "auto-research",
        "anthropic-pptx",
        "full-stack-planner",
        "commit-to-issue",
        "calendar-daily-briefing",
        "related-papers-scout",
        "feedback-meeting-scheduler",
        "gws-recipe-doc-from-template",
        "auto-research-distribute",
        "onboarding-accelerator",
        "sprint-orchestrator",
        "helm-validator",
        "nlm-slides",
        "daily-am-orchestrator",
        "technical-writer",
        "scrapling",
        "refactor-simulator",
        "plans",
        "daily-strategy-post",
        "critical-review",
        "prd-research-factory",
        "executive-briefing",
        "domain-commit",
        "eod-ship",
        "security-expert",
        "code-review-all",
        "cursor-automations",
        "simplify",
        "context-engineer",
        "daily-stock-check"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: fill a PPT template. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"ppt-template-engine\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"docx-template-engine\", \"office-template-enforcer\", \"alphaxiv-paper-lookup\", \"jira-linear-adapter\", \"auto-research\", \"anthropic-pptx\", \"full-stack-planner\", \"commit-to-issue\", \"calendar-daily-briefing\", \"related-papers-scout\", \"feedback-meeting-scheduler\", \"gws-recipe-doc-from-template\", \"auto-research-distribute\", \"onboarding-accelerator\", \"sprint-orchestrator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0223",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: run CI locally. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "ci-quality-gate"
      ],
      "distractor_skills": [
        "k8s-manifest-validator",
        "helm-validator",
        "setup-doctor",
        "security-expert",
        "code-review-all",
        "dependency-auditor",
        "compliance-gate",
        "sre-devops-expert",
        "performance-profiler",
        "qa-test-expert",
        "domain-commit",
        "e2e-testing",
        "ecc-verification-loop",
        "ux-expert",
        "technical-writer",
        "commit-to-issue",
        "daily-am-orchestrator",
        "simplify",
        "daily-stock-check",
        "frontend-expert",
        "calendar-daily-briefing",
        "semantic-guard",
        "terraform-reviewer",
        "alphaxiv-paper-lookup",
        "auto-research-distribute",
        "deep-review"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: run CI locally. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"ci-quality-gate\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"k8s-manifest-validator\", \"helm-validator\", \"setup-doctor\", \"security-expert\", \"code-review-all\", \"dependency-auditor\", \"compliance-gate\", \"sre-devops-expert\", \"performance-profiler\", \"qa-test-expert\", \"domain-commit\", \"e2e-testing\", \"ecc-verification-loop\", \"ux-expert\", \"technical-writer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0224",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 일일 지식 종합. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "knowledge-daily-aggregator"
      ],
      "distractor_skills": [
        "knowledge-to-paper",
        "email-auto-reply",
        "ai-chief-of-staff",
        "daily-strategy-post",
        "daily-am-orchestrator",
        "cognee",
        "weekly-status-report",
        "eod-ship",
        "ship",
        "morning",
        "morning-ship",
        "mirofish-graph-explorer",
        "gws-workflows",
        "standup-digest",
        "context-engineer",
        "agency-data-consolidation-agent",
        "release-commander",
        "proactive-meeting-scheduler",
        "portfolio-report-generator",
        "trading-breadth-chart-analyst",
        "setup-doctor",
        "sprint-orchestrator",
        "github-sprint-digest",
        "paper-archive",
        "calendar-daily-briefing",
        "autoskill-merger",
        "critical-review",
        "smart-meeting-scheduler",
        "qa-test-expert",
        "decision-router",
        "sales-meeting-pipeline",
        "stock-csv-downloader",
        "skill-guide-generator",
        "iac-review-agent",
        "ai-workflow-integrator",
        "release-ship",
        "trading-edge-candidate-agent",
        "pipeline-builder",
        "recall",
        "bespin-news-digest",
        "sefo-orchestrator",
        "nlm-arxiv-slides",
        "feedback-meeting-scheduler",
        "dependency-auditor",
        "kwp-sync",
        "weekly-stock-update",
        "hf-model-arena",
        "hf-auto-finetune-planner",
        "unified-intel-intake",
        "diagnose"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 일일 지식 종합. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"knowledge-daily-aggregator\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"knowledge-to-paper\", \"email-auto-reply\", \"ai-chief-of-staff\", \"daily-strategy-post\", \"daily-am-orchestrator\", \"cognee\", \"weekly-status-report\", \"eod-ship\", \"ship\", \"morning\", \"morning-ship\", \"mirofish-graph-explorer\", \"gws-workflows\", \"standup-digest\", \"context-engineer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0225",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Knowledge Synthesis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-enterprise-search-knowledge-synthesis"
      ],
      "distractor_skills": [
        "kwp-enterprise-search-source-management",
        "kwp-enterprise-search-search-strategy",
        "kwp-slack-slack-search",
        "kwp-human-resources-interview-prep",
        "kwp-design-design-handoff",
        "kwp-human-resources-people-analytics",
        "kwp-data-data-context-extractor",
        "kwp-finance-variance-analysis",
        "kwp-operations-vendor-management",
        "kwp-finance-close-management",
        "kwp-apollo-enrich-lead",
        "kwp-customer-support-customer-research",
        "kwp-sales-create-an-asset",
        "kwp-sales-account-research",
        "kwp-product-management-competitive-analysis",
        "kwp-marketing-performance-analytics",
        "kwp-design-system",
        "kwp-apollo-prospect",
        "kwp-common-room-account-research",
        "kwp-sales-draft-outreach",
        "kwp-sales-competitive-intelligence",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-org-planning",
        "kwp-data-statistical-analysis",
        "kwp-data-data-validation",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-finance-financial-statements",
        "kwp-product-management-feature-spec",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-finance-audit-support",
        "kwp-data-interactive-dashboard-builder",
        "kwp-customer-support-escalation",
        "kwp-legal-meeting-briefing",
        "kwp-customer-support-knowledge-management",
        "kwp-sales-daily-briefing",
        "kwp-product-management-user-research-synthesis",
        "kwp-marketing-campaign-planning",
        "kwp-legal-legal-risk-assessment",
        "kwp-finance-journal-entry-prep",
        "kwp-finance-reconciliation",
        "kwp-cowork-plugin-customizer",
        "kwp-operations-resource-planning",
        "kwp-data-data-visualization",
        "kwp-marketing-competitive-analysis",
        "kwp-marketing-brand-voice",
        "kwp-customer-support-response-drafting",
        "kwp-legal-nda-triage"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Knowledge Synthesis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-enterprise-search-knowledge-synthesis\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-enterprise-search-source-management\", \"kwp-enterprise-search-search-strategy\", \"kwp-slack-slack-search\", \"kwp-human-resources-interview-prep\", \"kwp-design-design-handoff\", \"kwp-human-resources-people-analytics\", \"kwp-data-data-context-extractor\", \"kwp-finance-variance-analysis\", \"kwp-operations-vendor-management\", \"kwp-finance-close-management\", \"kwp-apollo-enrich-lead\", \"kwp-customer-support-customer-research\", \"kwp-sales-create-an-asset\", \"kwp-sales-account-research\", \"kwp-product-management-competitive-analysis\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0226",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: choosing model tiers. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "ecc-agentic-engineering"
      ],
      "distractor_skills": [
        "ecc-search-first",
        "ecc-autonomous-loops",
        "ecc-coding-standards",
        "ecc-strategic-compact",
        "ecc-eval-harness",
        "ecc-configure",
        "ecc-continuous-learning",
        "ecc-verification-loop",
        "sefo-orchestrator",
        "agency-visionos-spatial-engineer",
        "agency-security-engineer",
        "sp-finish-branch",
        "kwp-finance-close-management",
        "ecc-iterative-retrieval",
        "agency-ux-architect",
        "agency-lsp-index-engineer",
        "sp-subagent-dev",
        "plans",
        "agency-xr-cockpit-interaction-specialist",
        "sp-tdd",
        "paperclip-setup",
        "kwp-productivity-task-management",
        "sp-request-review",
        "kwp-product-management-roadmap-management",
        "kwp-operations-resource-planning",
        "hf-jobs",
        "agency-rapid-prototyper",
        "sp-receive-review",
        "paperclip-tasks",
        "agency-senior-project-manager",
        "agency-macos-spatial-metal-engineer",
        "sp-brainstorming",
        "kwp-cowork-plugin-customizer",
        "agency-xr-interface-architect",
        "agency-devops-automator",
        "agency-ai-engineer",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-marketing-brand-voice",
        "anthropic-claude-api",
        "agency-workflow-optimizer",
        "sp-verification",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-org-planning",
        "kwp-data-data-validation"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: choosing model tiers. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"ecc-agentic-engineering\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"ecc-search-first\", \"ecc-autonomous-loops\", \"ecc-coding-standards\", \"ecc-strategic-compact\", \"ecc-eval-harness\", \"ecc-configure\", \"ecc-continuous-learning\", \"ecc-verification-loop\", \"sefo-orchestrator\", \"agency-visionos-spatial-engineer\", \"agency-security-engineer\", \"sp-finish-branch\", \"kwp-finance-close-management\", \"ecc-iterative-retrieval\", \"agency-ux-architect\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0227",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: create runbook. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "infra-runbook-generator"
      ],
      "distractor_skills": [
        "infra-drift-detector",
        "sre-devops-expert",
        "helm-validator",
        "docs-tutor-setup",
        "kwp-engineering-documentation",
        "k8s-manifest-validator",
        "terraform-reviewer",
        "incident-to-improvement",
        "iac-review-agent",
        "role-cto",
        "agency-devops-automator",
        "kwp-engineering-incident-response",
        "service-health-doctor",
        "local-dev-runner",
        "skill-guide-generator",
        "prompt-transformer",
        "i18n-sync",
        "role-developer",
        "pr-review-captain",
        "gws-recipe-doc-from-template",
        "ppt-template-engine",
        "onboarding-accelerator"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: create runbook. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"infra-runbook-generator\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"infra-drift-detector\", \"sre-devops-expert\", \"helm-validator\", \"docs-tutor-setup\", \"kwp-engineering-documentation\", \"k8s-manifest-validator\", \"terraform-reviewer\", \"incident-to-improvement\", \"iac-review-agent\", \"role-cto\", \"agency-devops-automator\", \"kwp-engineering-incident-response\", \"service-health-doctor\", \"local-dev-runner\", \"skill-guide-generator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0228",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: produce one. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "anthropic-pdf"
      ],
      "distractor_skills": [
        "anthropic-pptx",
        "anthropic-docx",
        "anthropic-xlsx",
        "anthropic-canvas-design",
        "anthropic-doc-coauthoring",
        "anthropic-claude-api",
        "gws-docs",
        "anthropic-skill-creator",
        "gws-sheets",
        "anthropic-slack-gif-creator",
        "skill-optimizer",
        "anthropic-theme-factory",
        "anthropic-algorithmic-art",
        "anthropic-brand-guidelines",
        "cognee",
        "anthropic-mcp-builder",
        "gws-recipe-events-from-sheet",
        "anthropic-frontend-design"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: produce one. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-pdf\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"anthropic-pptx\", \"anthropic-docx\", \"anthropic-xlsx\", \"anthropic-canvas-design\", \"anthropic-doc-coauthoring\", \"anthropic-claude-api\", \"gws-docs\", \"anthropic-skill-creator\", \"gws-sheets\", \"anthropic-slack-gif-creator\", \"skill-optimizer\", \"anthropic-theme-factory\", \"anthropic-algorithmic-art\", \"anthropic-brand-guidelines\", \"cognee\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0229",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: references agency-ui-designer. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-ui-designer"
      ],
      "distractor_skills": [
        "agency-brand-guardian",
        "agency-ux-researcher",
        "agency-xr-interface-architect",
        "agency-devops-automator",
        "agency-security-engineer",
        "agency-frontend-developer",
        "agency-backend-architect",
        "agency-visionos-spatial-engineer",
        "agency-accessibility-auditor",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-whimsy-injector",
        "agency-tiktok-strategist",
        "agency-technical-writer",
        "agency-ux-architect",
        "agency-performance-benchmarker",
        "agency-app-store-optimizer",
        "agency-visual-storyteller",
        "agency-api-tester",
        "agency-content-creator",
        "agency-behavioral-nudge-engine",
        "agency-wechat-official-account-manager",
        "agency-trend-researcher",
        "agency-experiment-tracker",
        "agency-agentic-identity-trust-architect",
        "agency-report-distribution-agent",
        "agency-rapid-prototyper",
        "agency-data-consolidation-agent",
        "agency-terminal-integration-specialist",
        "agency-reddit-community-builder",
        "agency-instagram-curator",
        "agency-data-analytics-reporter",
        "agency-analytics-reporter",
        "agency-tool-evaluator",
        "agency-macos-spatial-metal-engineer",
        "agency-lsp-index-engineer",
        "agency-twitter-engager",
        "agency-ai-engineer",
        "agency-sprint-prioritizer",
        "agency-finance-tracker",
        "agency-autonomous-optimization-architect",
        "agency-test-results-analyzer",
        "agency-inclusive-visuals-specialist",
        "agency-image-prompt-engineer",
        "agency-xiaohongshu-specialist",
        "agency-social-media-strategist",
        "agency-legal-compliance-checker",
        "agency-growth-hacker",
        "agency-evidence-collector",
        "agency-workflow-optimizer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: references agency-ui-designer. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-ui-designer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-brand-guardian\", \"agency-ux-researcher\", \"agency-xr-interface-architect\", \"agency-devops-automator\", \"agency-security-engineer\", \"agency-frontend-developer\", \"agency-backend-architect\", \"agency-visionos-spatial-engineer\", \"agency-accessibility-auditor\", \"agency-xr-immersive-developer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-whimsy-injector\", \"agency-tiktok-strategist\", \"agency-technical-writer\", \"agency-ux-architect\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0230",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: enhance and publish markdown. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "md-enhance-publish"
      ],
      "distractor_skills": [
        "md-to-slack-canvas",
        "md-to-notion",
        "deep-research-pipeline",
        "notion-meeting-sync",
        "auto-research-distribute",
        "daily-stock-check",
        "gws-workspace",
        "kwp-sync",
        "github-sprint-digest",
        "bespin-news-digest",
        "knowledge-to-paper",
        "anthropic-claude-api",
        "docs-tutor-setup",
        "twitter-timeline-to-slack",
        "skill-seekers",
        "anthropic-doc-coauthoring",
        "kwp-human-resources-people-analytics",
        "anthropic-canvas-design",
        "gws-recipe-doc-from-template",
        "gws-docs",
        "skill-optimizer",
        "cognee",
        "autoskill-merger",
        "commit-to-issue",
        "nlm-slides",
        "ui-suite",
        "gws-workflows",
        "gws-chat",
        "qa-test-expert",
        "weekly-status-report",
        "prompt-transformer",
        "nlm-arxiv-slides",
        "stock-csv-downloader",
        "meeting-digest",
        "agency-devops-automator",
        "executive-briefing",
        "docx-template-engine",
        "test-suite"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: enhance and publish markdown. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"md-enhance-publish\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"md-to-slack-canvas\", \"md-to-notion\", \"deep-research-pipeline\", \"notion-meeting-sync\", \"auto-research-distribute\", \"daily-stock-check\", \"gws-workspace\", \"kwp-sync\", \"github-sprint-digest\", \"bespin-news-digest\", \"knowledge-to-paper\", \"anthropic-claude-api\", \"docs-tutor-setup\", \"twitter-timeline-to-slack\", \"skill-seekers\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0231",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: battlecard. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "pm-go-to-market"
      ],
      "distractor_skills": [
        "pm-marketing-growth",
        "pm-market-research",
        "pm-product-strategy",
        "pm-toolkit",
        "pm-product-discovery",
        "pm-execution",
        "pm-data-analytics",
        "role-cso",
        "kwp-marketing-competitive-analysis",
        "kwp-marketing-campaign-planning",
        "agency-trend-researcher",
        "nlm-launch-kit",
        "kwp-human-resources-interview-prep",
        "agency-tiktok-strategist",
        "role-sales",
        "agency-xiaohongshu-specialist",
        "gws-recipe-doc-from-template",
        "e2e-testing",
        "agency-reddit-community-builder",
        "qa-test-expert",
        "kwp-product-management-competitive-analysis",
        "pm-sentinel",
        "kwp-engineering-testing-strategy",
        "anthropic-internal-comms",
        "trading-breadth-chart-analyst",
        "gws-recipe-events-from-sheet",
        "gws-drive",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-report-distribution-agent",
        "agency-twitter-engager",
        "gws-docs",
        "agency-app-store-optimizer",
        "mission-control copy",
        "role-ceo",
        "agency-instagram-curator",
        "kwp-sync",
        "agency-security-engineer",
        "agency-legal-compliance-checker",
        "agency-growth-hacker",
        "gws-recipe-save-email-to-doc",
        "gws-chat",
        "agency-executive-summary-generator",
        "trading-market-environment-analysis",
        "agency-ux-researcher",
        "agency-rapid-prototyper",
        "agency-data-consolidation-agent",
        "gws-calendar",
        "agency-wechat-official-account-manager"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: battlecard. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"pm-go-to-market\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"pm-marketing-growth\", \"pm-market-research\", \"pm-product-strategy\", \"pm-toolkit\", \"pm-product-discovery\", \"pm-execution\", \"pm-data-analytics\", \"role-cso\", \"kwp-marketing-competitive-analysis\", \"kwp-marketing-campaign-planning\", \"agency-trend-researcher\", \"nlm-launch-kit\", \"kwp-human-resources-interview-prep\", \"agency-tiktok-strategist\", \"role-sales\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0232",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: make me a GIF of X doing Y for Slack. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "anthropic-slack-gif-creator"
      ],
      "distractor_skills": [
        "anthropic-algorithmic-art",
        "anthropic-canvas-design",
        "anthropic-brand-guidelines",
        "anthropic-web-artifacts-builder",
        "anthropic-skill-creator",
        "anthropic-theme-factory",
        "anthropic-claude-api",
        "anthropic-mcp-builder",
        "anthropic-doc-coauthoring",
        "anthropic-pdf",
        "anthropic-webapp-testing",
        "anthropic-frontend-design",
        "anthropic-xlsx",
        "anthropic-internal-comms",
        "anthropic-docx",
        "gws-chat",
        "anthropic-pptx",
        "slack-agent",
        "kwp-data-data-visualization",
        "kwp-cowork-create-plugin",
        "kwp-design-system",
        "pm-toolkit",
        "ecc-configure",
        "md-to-slack-canvas",
        "gws-recipe-events-from-sheet",
        "gws-recipe-doc-from-template",
        "x-to-slack",
        "kwp-slack-slack-search",
        "alphaear-logic-visualizer",
        "commit-to-issue"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: make me a GIF of X doing Y for Slack. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-slack-gif-creator\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"anthropic-algorithmic-art\", \"anthropic-canvas-design\", \"anthropic-brand-guidelines\", \"anthropic-web-artifacts-builder\", \"anthropic-skill-creator\", \"anthropic-theme-factory\", \"anthropic-claude-api\", \"anthropic-mcp-builder\", \"anthropic-doc-coauthoring\", \"anthropic-pdf\", \"anthropic-webapp-testing\", \"anthropic-frontend-design\", \"anthropic-xlsx\", \"anthropic-internal-comms\", \"anthropic-docx\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0233",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: news-informed forecast adjustments. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "alphaear-predictor"
      ],
      "distractor_skills": [
        "alphaear-deepear-lite",
        "alphaear-sentiment",
        "alphaear-news",
        "alphaear-search",
        "alphaear-stock",
        "alphaear-signal-tracker",
        "alphaear-reporter",
        "tab-strategy-comparison",
        "tab-report-generate",
        "tab-kiwoom",
        "trading-market-environment-analysis",
        "stock-csv-downloader",
        "trading-breadth-chart-analyst"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: news-informed forecast adjustments. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-predictor\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"alphaear-deepear-lite\", \"alphaear-sentiment\", \"alphaear-news\", \"alphaear-search\", \"alphaear-stock\", \"alphaear-signal-tracker\", \"alphaear-reporter\", \"tab-strategy-comparison\", \"tab-report-generate\", \"tab-kiwoom\", \"trading-market-environment-analysis\", \"stock-csv-downloader\", \"trading-breadth-chart-analyst\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0234",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Content Creator agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-content-creator"
      ],
      "distractor_skills": [
        "agency-social-media-strategist",
        "agency-brand-guardian",
        "agency-instagram-curator",
        "agency-xr-interface-architect",
        "agency-growth-hacker",
        "agency-tiktok-strategist",
        "agency-xr-immersive-developer",
        "agency-wechat-official-account-manager",
        "agency-api-tester",
        "agency-workflow-optimizer",
        "agency-developer-advocate",
        "agency-xiaohongshu-specialist",
        "agency-app-store-optimizer",
        "agency-visual-storyteller",
        "agency-reddit-community-builder",
        "agency-xr-cockpit-interaction-specialist",
        "agency-zhihu-strategist",
        "agency-terminal-integration-specialist",
        "agency-cultural-intelligence-strategist",
        "agency-mobile-app-builder",
        "agency-legal-compliance-checker"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Content Creator agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-content-creator\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-social-media-strategist\", \"agency-brand-guardian\", \"agency-instagram-curator\", \"agency-xr-interface-architect\", \"agency-growth-hacker\", \"agency-tiktok-strategist\", \"agency-xr-immersive-developer\", \"agency-wechat-official-account-manager\", \"agency-api-tester\", \"agency-workflow-optimizer\", \"agency-developer-advocate\", \"agency-xiaohongshu-specialist\", \"agency-app-store-optimizer\", \"agency-visual-storyteller\", \"agency-reddit-community-builder\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0235",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: run E2E tests. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "e2e-testing"
      ],
      "distractor_skills": [
        "qa-test-expert",
        "e2e-overhaul",
        "kwp-engineering-testing-strategy",
        "anthropic-webapp-testing",
        "frontend-expert",
        "agency-xr-immersive-developer",
        "ci-quality-gate",
        "agency-test-results-analyzer",
        "evals-skills",
        "agency-performance-benchmarker",
        "pm-data-analytics",
        "deep-review",
        "pm-marketing-growth",
        "agency-ux-researcher",
        "pm-execution",
        "agency-xr-cockpit-interaction-specialist",
        "agency-frontend-developer",
        "ux-expert",
        "agency-tool-evaluator",
        "gws-docs",
        "agency-brand-guardian",
        "local-dev-runner",
        "agency-api-tester",
        "kwp-human-resources-interview-prep",
        "technical-writer",
        "agency-app-store-optimizer",
        "gws-sheets"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: run E2E tests. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"e2e-testing\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"qa-test-expert\", \"e2e-overhaul\", \"kwp-engineering-testing-strategy\", \"anthropic-webapp-testing\", \"frontend-expert\", \"agency-xr-immersive-developer\", \"ci-quality-gate\", \"agency-test-results-analyzer\", \"evals-skills\", \"agency-performance-benchmarker\", \"pm-data-analytics\", \"deep-review\", \"pm-marketing-growth\", \"agency-ux-researcher\", \"pm-execution\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0236",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 프레젠테이션 템플릿 생성. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "ppt-template-engine"
      ],
      "distractor_skills": [
        "docx-template-engine",
        "office-template-enforcer",
        "alphaxiv-paper-lookup",
        "jira-linear-adapter",
        "auto-research",
        "anthropic-pptx",
        "full-stack-planner",
        "commit-to-issue",
        "calendar-daily-briefing",
        "related-papers-scout",
        "feedback-meeting-scheduler",
        "gws-recipe-doc-from-template",
        "auto-research-distribute",
        "onboarding-accelerator",
        "sprint-orchestrator",
        "helm-validator",
        "nlm-slides",
        "daily-am-orchestrator",
        "technical-writer",
        "scrapling",
        "refactor-simulator",
        "plans",
        "daily-strategy-post",
        "critical-review",
        "prd-research-factory",
        "executive-briefing",
        "domain-commit",
        "eod-ship",
        "security-expert",
        "code-review-all",
        "cursor-automations",
        "simplify",
        "context-engineer",
        "daily-stock-check",
        "cursor-sync",
        "proposal-factory",
        "setup-doctor",
        "pipeline-builder"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 프레젠테이션 템플릿 생성. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"ppt-template-engine\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"docx-template-engine\", \"office-template-enforcer\", \"alphaxiv-paper-lookup\", \"jira-linear-adapter\", \"auto-research\", \"anthropic-pptx\", \"full-stack-planner\", \"commit-to-issue\", \"calendar-daily-briefing\", \"related-papers-scout\", \"feedback-meeting-scheduler\", \"gws-recipe-doc-from-template\", \"auto-research-distribute\", \"onboarding-accelerator\", \"sprint-orchestrator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0237",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Analytics Reporter agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-analytics-reporter"
      ],
      "distractor_skills": [
        "agency-data-analytics-reporter",
        "agency-ux-researcher",
        "agency-finance-tracker",
        "agency-data-consolidation-agent",
        "agency-feedback-synthesizer",
        "agency-brand-guardian",
        "agency-xr-immersive-developer",
        "agency-growth-hacker",
        "agency-trend-researcher",
        "agency-experiment-tracker",
        "agency-data-engineer",
        "agency-tool-evaluator",
        "agency-performance-benchmarker",
        "agency-app-store-optimizer",
        "agency-xr-interface-architect",
        "agency-devops-automator",
        "agency-sprint-prioritizer",
        "agency-test-results-analyzer",
        "agency-sales-data-extraction-agent",
        "agency-legal-compliance-checker",
        "agency-workflow-optimizer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-frontend-developer",
        "agency-whimsy-injector"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Analytics Reporter agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-analytics-reporter\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-data-analytics-reporter\", \"agency-ux-researcher\", \"agency-finance-tracker\", \"agency-data-consolidation-agent\", \"agency-feedback-synthesizer\", \"agency-brand-guardian\", \"agency-xr-immersive-developer\", \"agency-growth-hacker\", \"agency-trend-researcher\", \"agency-experiment-tracker\", \"agency-data-engineer\", \"agency-tool-evaluator\", \"agency-performance-benchmarker\", \"agency-app-store-optimizer\", \"agency-xr-interface-architect\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0238",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: multi-day price predictions. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "alphaear-predictor"
      ],
      "distractor_skills": [
        "alphaear-deepear-lite",
        "alphaear-sentiment",
        "alphaear-news",
        "alphaear-search",
        "alphaear-stock",
        "alphaear-signal-tracker",
        "alphaear-reporter",
        "tab-strategy-comparison",
        "tab-report-generate",
        "tab-kiwoom"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: multi-day price predictions. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-predictor\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"alphaear-deepear-lite\", \"alphaear-sentiment\", \"alphaear-news\", \"alphaear-search\", \"alphaear-stock\", \"alphaear-signal-tracker\", \"alphaear-reporter\", \"tab-strategy-comparison\", \"tab-report-generate\", \"tab-kiwoom\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0239",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 일정 요약. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "calendar-daily-briefing"
      ],
      "distractor_skills": [
        "gws-workflows",
        "feedback-meeting-scheduler",
        "gws-calendar",
        "daily-am-orchestrator",
        "kwp-sales-daily-briefing",
        "gws-gmail",
        "smart-meeting-scheduler",
        "gws-recipe-events-from-sheet",
        "gws-chat",
        "gws-workspace",
        "sprint-orchestrator",
        "setup-doctor",
        "ai-chief-of-staff",
        "proactive-meeting-scheduler",
        "jira-linear-adapter",
        "gws-drive",
        "simplify",
        "daily-stock-check",
        "ci-quality-gate",
        "commit-to-issue",
        "gws-recipe-save-email-to-doc",
        "technical-writer",
        "morning",
        "critical-review",
        "gws-recipe-personalized-emails",
        "problem-definition",
        "morning-ship",
        "domain-commit",
        "standup-digest",
        "eod-ship",
        "gws-recipe-doc-from-template"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 일정 요약. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"calendar-daily-briefing\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"gws-workflows\", \"feedback-meeting-scheduler\", \"gws-calendar\", \"daily-am-orchestrator\", \"kwp-sales-daily-briefing\", \"gws-gmail\", \"smart-meeting-scheduler\", \"gws-recipe-events-from-sheet\", \"gws-chat\", \"gws-workspace\", \"sprint-orchestrator\", \"setup-doctor\", \"ai-chief-of-staff\", \"proactive-meeting-scheduler\", \"jira-linear-adapter\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0240",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: the role-dispatcher invokes this skill after collecting role analyses. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "executive-briefing"
      ],
      "distractor_skills": [
        "daily-strategy-post",
        "role-ceo",
        "critical-review",
        "daily-stock-check",
        "domain-commit",
        "eod-ship",
        "commit-to-issue",
        "calendar-daily-briefing",
        "alphaxiv-paper-lookup",
        "technical-writer",
        "problem-definition",
        "daily-am-orchestrator",
        "auto-research-distribute",
        "pipeline-builder",
        "codebase-archaeologist",
        "portfolio-report-generator",
        "ci-quality-gate",
        "deep-research-pipeline",
        "security-expert",
        "role-dispatcher",
        "setup-doctor",
        "simplify",
        "agency-feedback-synthesizer",
        "agency-xr-interface-architect",
        "role-trading-expert",
        "role-pm",
        "full-stack-planner",
        "sprint-orchestrator",
        "role-sales",
        "role-finance",
        "agency-brand-guardian",
        "kwp-brand-voice-guideline-generation",
        "code-review-all",
        "role-developer",
        "sre-devops-expert",
        "feedback-meeting-scheduler",
        "automation-strategist",
        "context-engineer",
        "role-cso",
        "k8s-manifest-validator",
        "refactor-simulator"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: the role-dispatcher invokes this skill after collecting role analyses. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"executive-briefing\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"daily-strategy-post\", \"role-ceo\", \"critical-review\", \"daily-stock-check\", \"domain-commit\", \"eod-ship\", \"commit-to-issue\", \"calendar-daily-briefing\", \"alphaxiv-paper-lookup\", \"technical-writer\", \"problem-definition\", \"daily-am-orchestrator\", \"auto-research-distribute\", \"pipeline-builder\", \"codebase-archaeologist\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0241",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: make a new plugin. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-cowork-create-plugin"
      ],
      "distractor_skills": [
        "kwp-cowork-create-cowork-plugin",
        "kwp-design-system",
        "kwp-design-design-handoff",
        "kwp-cowork-plugin-customizer",
        "kwp-data-data-context-extractor",
        "kwp-design-user-research",
        "kwp-data-data-visualization",
        "kwp-sales-competitive-intelligence",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-operations-process-optimization",
        "kwp-data-interactive-dashboard-builder"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: make a new plugin. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-cowork-create-plugin\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-cowork-create-cowork-plugin\", \"kwp-design-system\", \"kwp-design-design-handoff\", \"kwp-cowork-plugin-customizer\", \"kwp-data-data-context-extractor\", \"kwp-design-user-research\", \"kwp-data-data-visualization\", \"kwp-sales-competitive-intelligence\", \"kwp-operations-vendor-management\", \"kwp-human-resources-interview-prep\", \"kwp-operations-process-optimization\", \"kwp-data-interactive-dashboard-builder\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0242",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Security Engineer agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-security-engineer"
      ],
      "distractor_skills": [
        "agency-devops-automator",
        "agency-backend-architect",
        "agency-frontend-developer",
        "agency-visionos-spatial-engineer",
        "agency-xr-immersive-developer",
        "agency-terminal-integration-specialist",
        "agency-brand-guardian",
        "agency-mobile-app-builder",
        "agency-xr-interface-architect",
        "agency-app-store-optimizer",
        "agency-tool-evaluator",
        "agency-macos-spatial-metal-engineer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-ai-engineer",
        "agency-ux-researcher",
        "agency-ux-architect",
        "agency-trend-researcher",
        "agency-experiment-tracker",
        "agency-performance-benchmarker"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Security Engineer agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-security-engineer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-devops-automator\", \"agency-backend-architect\", \"agency-frontend-developer\", \"agency-visionos-spatial-engineer\", \"agency-xr-immersive-developer\", \"agency-terminal-integration-specialist\", \"agency-brand-guardian\", \"agency-mobile-app-builder\", \"agency-xr-interface-architect\", \"agency-app-store-optimizer\", \"agency-tool-evaluator\", \"agency-macos-spatial-metal-engineer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-ai-engineer\", \"agency-ux-researcher\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0243",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Design Review. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "design-review"
      ],
      "distractor_skills": [
        "implement-screen",
        "figma-to-tds",
        "screen-description",
        "fe-pipeline",
        "github-workflow-automation",
        "fsd-development",
        "swagger-api-doc-generator",
        "overlay-layout-patterns",
        "notion-docs-sync",
        "design-architect",
        "sp-executing-plans",
        "anthropic-web-artifacts-builder",
        "kwp-operations-vendor-management",
        "anthropic-claude-api",
        "local-dev-setup",
        "sp-subagent-dev",
        "agency-accessibility-auditor",
        "sp-debugging",
        "kwp-engineering-system-design",
        "anthropic-webapp-testing",
        "kwp-design-design-system-management",
        "kwp-design-design-critique",
        "agency-mobile-app-builder",
        "sp-request-review",
        "sp-finish-branch",
        "kwp-design-design-handoff",
        "kwp-legal-nda-triage",
        "anthropic-frontend-design",
        "agency-devops-automator",
        "agency-backend-architect",
        "anthropic-theme-factory",
        "kwp-sales-competitive-intelligence",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-org-planning",
        "agency-xr-immersive-developer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Design Review. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"design-review\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"implement-screen\", \"figma-to-tds\", \"screen-description\", \"fe-pipeline\", \"github-workflow-automation\", \"fsd-development\", \"swagger-api-doc-generator\", \"overlay-layout-patterns\", \"notion-docs-sync\", \"design-architect\", \"sp-executing-plans\", \"anthropic-web-artifacts-builder\", \"kwp-operations-vendor-management\", \"anthropic-claude-api\", \"local-dev-setup\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0244",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Recruiting Pipeline. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-human-resources-recruiting-pipeline"
      ],
      "distractor_skills": [
        "kwp-human-resources-interview-prep",
        "kwp-operations-vendor-management",
        "kwp-operations-process-optimization",
        "kwp-human-resources-people-analytics",
        "kwp-human-resources-org-planning",
        "kwp-finance-close-management",
        "kwp-operations-change-management",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-operations-resource-planning",
        "kwp-sales-competitive-intelligence",
        "kwp-product-management-metrics-tracking",
        "kwp-slack-slack-search",
        "kwp-sales-call-prep",
        "kwp-product-management-feature-spec",
        "kwp-design-design-handoff",
        "kwp-design-system",
        "kwp-legal-meeting-briefing",
        "kwp-product-management-competitive-analysis",
        "kwp-human-resources-employee-handbook",
        "kwp-apollo-prospect",
        "kwp-sales-daily-briefing",
        "kwp-sales-account-research",
        "kwp-sales-draft-outreach",
        "kwp-data-data-validation",
        "kwp-finance-variance-analysis",
        "kwp-finance-financial-statements",
        "kwp-operations-compliance-tracking",
        "kwp-finance-audit-support",
        "kwp-data-interactive-dashboard-builder",
        "kwp-customer-support-escalation",
        "kwp-apollo-enrich-lead",
        "kwp-customer-support-knowledge-management",
        "kwp-finance-reconciliation",
        "kwp-cowork-plugin-customizer",
        "kwp-product-management-user-research-synthesis",
        "kwp-marketing-performance-analytics",
        "kwp-marketing-campaign-planning",
        "kwp-legal-legal-risk-assessment",
        "kwp-finance-journal-entry-prep",
        "kwp-engineering-system-design"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Recruiting Pipeline. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-human-resources-recruiting-pipeline\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-human-resources-interview-prep\", \"kwp-operations-vendor-management\", \"kwp-operations-process-optimization\", \"kwp-human-resources-people-analytics\", \"kwp-human-resources-org-planning\", \"kwp-finance-close-management\", \"kwp-operations-change-management\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-operations-resource-planning\", \"kwp-sales-competitive-intelligence\", \"kwp-product-management-metrics-tracking\", \"kwp-slack-slack-search\", \"kwp-sales-call-prep\", \"kwp-product-management-feature-spec\", \"kwp-design-design-handoff\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0245",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Web Artifacts Builder. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "anthropic-web-artifacts-builder"
      ],
      "distractor_skills": [
        "anthropic-frontend-design",
        "anthropic-theme-factory",
        "anthropic-claude-api",
        "anthropic-webapp-testing",
        "anthropic-algorithmic-art",
        "anthropic-canvas-design",
        "anthropic-slack-gif-creator",
        "frontend-expert",
        "ux-expert",
        "anthropic-xlsx",
        "anthropic-brand-guidelines",
        "kwp-design-design-system-management",
        "anthropic-internal-comms",
        "anthropic-mcp-builder",
        "anthropic-doc-coauthoring",
        "backend-expert",
        "anthropic-skill-creator",
        "sp-executing-plans",
        "anthropic-pptx",
        "anthropic-docx",
        "agency-xr-immersive-developer",
        "agency-frontend-developer",
        "ecc-coding-standards",
        "ci-quality-gate",
        "sp-debugging",
        "design-review"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Web Artifacts Builder. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-web-artifacts-builder\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"anthropic-frontend-design\", \"anthropic-theme-factory\", \"anthropic-claude-api\", \"anthropic-webapp-testing\", \"anthropic-algorithmic-art\", \"anthropic-canvas-design\", \"anthropic-slack-gif-creator\", \"frontend-expert\", \"ux-expert\", \"anthropic-xlsx\", \"anthropic-brand-guidelines\", \"kwp-design-design-system-management\", \"anthropic-internal-comms\", \"anthropic-mcp-builder\", \"anthropic-doc-coauthoring\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0246",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: implementation is complete. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "sp-finish-branch"
      ],
      "distractor_skills": [
        "sp-verification",
        "sp-executing-plans",
        "sp-receive-review",
        "sp-subagent-dev",
        "sp-tdd",
        "sp-debugging",
        "sp-request-review",
        "sp-core",
        "sp-brainstorming",
        "sp-writing-skills"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: implementation is complete. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"sp-finish-branch\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"sp-verification\", \"sp-executing-plans\", \"sp-receive-review\", \"sp-subagent-dev\", \"sp-tdd\", \"sp-debugging\", \"sp-request-review\", \"sp-core\", \"sp-brainstorming\", \"sp-writing-skills\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0247",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 병렬 실행. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "workflow-parallel"
      ],
      "distractor_skills": [
        "workflow-sequential",
        "workflow-eval-opt",
        "workflow-miner",
        "mission-control",
        "deep-review",
        "mission-control copy",
        "ship",
        "sp-subagent-dev",
        "code-review-all",
        "test-suite",
        "autoskill-merger",
        "paperclip-agents",
        "compliance-gate",
        "sp-parallel-agents",
        "paperclip-tasks",
        "deep-research-pipeline",
        "auto-research",
        "ui-suite",
        "paperclip-setup",
        "kwp-productivity-task-management",
        "paperclip-control",
        "kwp-operations-vendor-management",
        "autoskill-pipeline",
        "auto-research-distribute",
        "simplify"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 병렬 실행. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"workflow-parallel\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"workflow-sequential\", \"workflow-eval-opt\", \"workflow-miner\", \"mission-control\", \"deep-review\", \"mission-control copy\", \"ship\", \"sp-subagent-dev\", \"code-review-all\", \"test-suite\", \"autoskill-merger\", \"paperclip-agents\", \"compliance-gate\", \"sp-parallel-agents\", \"paperclip-tasks\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0248",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: threat analysis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "role-security-engineer"
      ],
      "distractor_skills": [
        "role-finance",
        "role-data-scientist",
        "role-trading-expert",
        "role-ux-designer",
        "role-cto",
        "role-hr",
        "role-developer",
        "role-pm",
        "role-cso",
        "role-sales",
        "role-ceo",
        "security-expert",
        "role-dispatcher",
        "compliance-governance",
        "deep-research-pipeline",
        "kwp-operations-risk-assessment",
        "semantic-guard",
        "dependency-auditor",
        "agency-security-engineer",
        "executive-briefing",
        "kwp-operations-compliance-tracking",
        "kwp-design-accessibility-review",
        "agency-legal-compliance-checker",
        "paper-review",
        "sefo-governance",
        "kwp-engineering-code-review",
        "agency-infrastructure-maintainer",
        "gws-recipe-audit-sharing",
        "incident-to-improvement",
        "kwp-engineering-incident-response",
        "kwp-design-design-system-management",
        "ux-expert",
        "kwp-engineering-tech-debt",
        "infra-runbook-generator",
        "kwp-human-resources-people-analytics",
        "agency-xr-immersive-developer",
        "agency-workflow-optimizer",
        "mission-control"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: threat analysis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"role-security-engineer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"role-finance\", \"role-data-scientist\", \"role-trading-expert\", \"role-ux-designer\", \"role-cto\", \"role-hr\", \"role-developer\", \"role-pm\", \"role-cso\", \"role-sales\", \"role-ceo\", \"security-expert\", \"role-dispatcher\", \"compliance-governance\", \"deep-research-pipeline\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0249",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: test generation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "qa-test-expert"
      ],
      "distractor_skills": [
        "kwp-engineering-testing-strategy",
        "e2e-testing",
        "frontend-expert",
        "test-suite",
        "ci-quality-gate",
        "sre-devops-expert",
        "dependency-auditor",
        "anthropic-webapp-testing",
        "agency-test-results-analyzer",
        "kwp-human-resources-org-planning",
        "skill-optimizer",
        "db-expert"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: test generation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"qa-test-expert\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-engineering-testing-strategy\", \"e2e-testing\", \"frontend-expert\", \"test-suite\", \"ci-quality-gate\", \"sre-devops-expert\", \"dependency-auditor\", \"anthropic-webapp-testing\", \"agency-test-results-analyzer\", \"kwp-human-resources-org-planning\", \"skill-optimizer\", \"db-expert\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0250",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: tailor plugin. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-cowork-cowork-plugin-customizer"
      ],
      "distractor_skills": [
        "kwp-cowork-plugin-customizer",
        "kwp-cowork-create-cowork-plugin",
        "kwp-enterprise-search-source-management",
        "kwp-cowork-create-plugin",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-data-data-context-extractor",
        "kwp-customer-support-knowledge-management",
        "kwp-customer-support-customer-research",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-human-resources-people-analytics",
        "kwp-design-system",
        "kwp-human-resources-org-planning",
        "kwp-slack-slack-search",
        "kwp-productivity-memory-management",
        "kwp-product-management-feature-spec"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: tailor plugin. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-cowork-cowork-plugin-customizer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-cowork-plugin-customizer\", \"kwp-cowork-create-cowork-plugin\", \"kwp-enterprise-search-source-management\", \"kwp-cowork-create-plugin\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-data-data-context-extractor\", \"kwp-customer-support-knowledge-management\", \"kwp-customer-support-customer-research\", \"kwp-operations-vendor-management\", \"kwp-human-resources-interview-prep\", \"kwp-human-resources-people-analytics\", \"kwp-design-system\", \"kwp-human-resources-org-planning\", \"kwp-slack-slack-search\", \"kwp-productivity-memory-management\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0251",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: slide masters. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "office-template-enforcer"
      ],
      "distractor_skills": [
        "ppt-template-engine",
        "docx-template-engine",
        "gws-recipe-doc-from-template",
        "nlm-slides",
        "anthropic-pptx",
        "anthropic-pdf",
        "gws-docs",
        "kwp-sync",
        "anthropic-docx",
        "anthropic-doc-coauthoring",
        "md-enhance-publish",
        "knowledge-to-paper",
        "infra-runbook-generator",
        "skill-guide-generator",
        "executive-briefing",
        "kwp-cowork-create-cowork-plugin",
        "gws-recipe-save-email-to-doc",
        "paper-review",
        "autoskill-extractor",
        "anthropic-canvas-design",
        "release-ship",
        "autoskill-judge",
        "notebooklm-studio"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: slide masters. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"office-template-enforcer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"ppt-template-engine\", \"docx-template-engine\", \"gws-recipe-doc-from-template\", \"nlm-slides\", \"anthropic-pptx\", \"anthropic-pdf\", \"gws-docs\", \"kwp-sync\", \"anthropic-docx\", \"anthropic-doc-coauthoring\", \"md-enhance-publish\", \"knowledge-to-paper\", \"infra-runbook-generator\", \"skill-guide-generator\", \"executive-briefing\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0252",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: check model info. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "hf-models"
      ],
      "distractor_skills": [
        "hf-discussions",
        "hf-collections",
        "hf-cli",
        "hf-repos",
        "hf-hub",
        "hf-buckets",
        "hf-spaces",
        "hf-endpoints",
        "hf-papers",
        "hf-dataset-viewer",
        "hf-model-arena",
        "hf-auto-finetune-planner",
        "hf-datasets",
        "hf-model-trainer",
        "hf-evaluation",
        "gws-drive",
        "hf-trackio",
        "hf-jobs",
        "hf-trending-intelligence",
        "gws-docs",
        "gws-calendar",
        "gws-chat",
        "gws-sheets",
        "gws-gmail",
        "gws-recipe-audit-sharing",
        "gws-workflows",
        "daiso-mcp",
        "public-apis",
        "gws-workspace"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: check model info. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"hf-models\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"hf-discussions\", \"hf-collections\", \"hf-cli\", \"hf-repos\", \"hf-hub\", \"hf-buckets\", \"hf-spaces\", \"hf-endpoints\", \"hf-papers\", \"hf-dataset-viewer\", \"hf-model-arena\", \"hf-auto-finetune-planner\", \"hf-datasets\", \"hf-model-trainer\", \"hf-evaluation\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0253",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: updating signal confidence and intensity. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "alphaear-signal-tracker"
      ],
      "distractor_skills": [
        "alphaear-news",
        "alphaear-sentiment",
        "alphaear-deepear-lite",
        "alphaear-search",
        "alphaear-predictor",
        "alphaear-reporter",
        "alphaear-logic-visualizer",
        "trading-market-environment-analysis",
        "tab-report-generate",
        "alphaear-stock",
        "stock-csv-downloader",
        "trading-breadth-chart-analyst",
        "tab-news-fetch",
        "tab-market-environment",
        "hf-evaluation",
        "tab-kiwoom",
        "daily-stock-check",
        "trading-sector-analyst",
        "trading-scenario-analyzer",
        "trading-technical-analyst",
        "trading-stanley-druckenmiller-investment",
        "tab-llm-agents",
        "trading-market-top-detector",
        "trading-us-stock-analysis",
        "hf-model-trainer",
        "trading-backtest-expert",
        "tab-strategy-comparison",
        "weekly-stock-update",
        "hf-trackio",
        "tab-stock-sync",
        "role-trading-expert",
        "hf-jobs",
        "ecc-eval-harness",
        "trading-edge-candidate-agent",
        "tab-dualma-refresh",
        "tab-bollinger-refresh",
        "tab-analysis-run",
        "trading-market-breadth-analyzer",
        "trading-uptrend-analyzer",
        "trading-theme-detector",
        "today",
        "pm-market-research",
        "ecc-autonomous-loops",
        "ecc-continuous-learning",
        "kwp-finance-financial-statements",
        "mirofish-financial-sim",
        "tab-screening",
        "tab-market-breadth",
        "tab-dashboard-summary",
        "kwp-finance-audit-support"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: updating signal confidence and intensity. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-signal-tracker\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"alphaear-news\", \"alphaear-sentiment\", \"alphaear-deepear-lite\", \"alphaear-search\", \"alphaear-predictor\", \"alphaear-reporter\", \"alphaear-logic-visualizer\", \"trading-market-environment-analysis\", \"tab-report-generate\", \"alphaear-stock\", \"stock-csv-downloader\", \"trading-breadth-chart-analyst\", \"tab-news-fetch\", \"tab-market-environment\", \"hf-evaluation\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0254",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Feedback Synthesizer agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-feedback-synthesizer"
      ],
      "distractor_skills": [
        "agency-brand-guardian",
        "agency-xr-immersive-developer",
        "agency-test-results-analyzer",
        "agency-xr-interface-architect",
        "agency-app-store-optimizer",
        "agency-trend-researcher",
        "agency-xr-cockpit-interaction-specialist",
        "agency-finance-tracker",
        "agency-terminal-integration-specialist",
        "agency-workflow-optimizer",
        "agency-data-consolidation-agent",
        "agency-visionos-spatial-engineer",
        "agency-tool-evaluator",
        "agency-report-distribution-agent",
        "agency-inclusive-visuals-specialist",
        "agency-api-tester",
        "agency-behavioral-nudge-engine",
        "agency-legal-compliance-checker",
        "agency-growth-hacker",
        "agency-content-creator",
        "agency-ux-architect",
        "agency-rapid-prototyper",
        "agency-mobile-app-builder",
        "agency-project-shepherd",
        "agency-tiktok-strategist",
        "agency-senior-developer",
        "agency-support-responder",
        "agency-senior-project-manager",
        "agency-macos-spatial-metal-engineer",
        "agency-lsp-index-engineer",
        "agency-reddit-community-builder",
        "agency-instagram-curator",
        "agency-evidence-collector",
        "agency-sales-data-extraction-agent",
        "agency-infrastructure-maintainer",
        "agency-agents-orchestrator",
        "agency-developer-advocate",
        "agency-ai-engineer",
        "agency-cultural-intelligence-strategist",
        "agency-xiaohongshu-specialist",
        "agency-social-media-strategist",
        "agency-ux-researcher",
        "agency-reality-checker",
        "agency-whimsy-injector",
        "agency-wechat-official-account-manager"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Feedback Synthesizer agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-feedback-synthesizer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-brand-guardian\", \"agency-xr-immersive-developer\", \"agency-test-results-analyzer\", \"agency-xr-interface-architect\", \"agency-app-store-optimizer\", \"agency-trend-researcher\", \"agency-xr-cockpit-interaction-specialist\", \"agency-finance-tracker\", \"agency-terminal-integration-specialist\", \"agency-workflow-optimizer\", \"agency-data-consolidation-agent\", \"agency-visionos-spatial-engineer\", \"agency-tool-evaluator\", \"agency-report-distribution-agent\", \"agency-inclusive-visuals-specialist\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0255",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 리서치 후 종합 분석. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "deep-research-pipeline"
      ],
      "distractor_skills": [
        "paper-review",
        "role-ceo",
        "deep-review",
        "role-cso",
        "critical-review",
        "role-sales",
        "role-developer",
        "role-ux-designer",
        "role-dispatcher",
        "email-research-dispatcher",
        "role-data-scientist",
        "role-cto",
        "role-finance",
        "ui-suite",
        "hf-trending-intelligence",
        "role-pm",
        "auto-research",
        "meeting-digest",
        "role-hr",
        "role-trading-expert",
        "ship",
        "test-suite",
        "md-enhance-publish",
        "role-security-engineer",
        "autoskill-pipeline",
        "kwp-human-resources-people-analytics",
        "weekly-status-report",
        "twitter-timeline-to-slack",
        "nlm-arxiv-slides",
        "daily-strategy-post",
        "executive-briefing",
        "alphaxiv-paper-lookup",
        "hf-papers",
        "docs-tutor-setup",
        "alphaear-search",
        "agency-xr-immersive-developer",
        "portfolio-report-generator",
        "release-ship",
        "incident-to-improvement",
        "notebooklm-research"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 리서치 후 종합 분석. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"deep-research-pipeline\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"paper-review\", \"role-ceo\", \"deep-review\", \"role-cso\", \"critical-review\", \"role-sales\", \"role-developer\", \"role-ux-designer\", \"role-dispatcher\", \"email-research-dispatcher\", \"role-data-scientist\", \"role-cto\", \"role-finance\", \"ui-suite\", \"hf-trending-intelligence\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0256",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: adding a dependency. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "ecc-search-first"
      ],
      "distractor_skills": [
        "sp-tdd",
        "ecc-configure",
        "ecc-coding-standards",
        "ecc-strategic-compact",
        "ecc-agentic-engineering",
        "ecc-continuous-learning",
        "ecc-autonomous-loops",
        "sp-git-worktrees",
        "ecc-iterative-retrieval",
        "ecc-verification-loop",
        "sp-writing-skills",
        "ecc-eval-harness",
        "kwp-product-management-feature-spec",
        "sp-core",
        "hf-papers",
        "issue-to-code",
        "recall",
        "sp-receive-review",
        "kwp-sales-account-research",
        "alphaear-search",
        "sp-verification",
        "skill-optimizer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: adding a dependency. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"ecc-search-first\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"sp-tdd\", \"ecc-configure\", \"ecc-coding-standards\", \"ecc-strategic-compact\", \"ecc-agentic-engineering\", \"ecc-continuous-learning\", \"ecc-autonomous-loops\", \"sp-git-worktrees\", \"ecc-iterative-retrieval\", \"ecc-verification-loop\", \"sp-writing-skills\", \"ecc-eval-harness\", \"kwp-product-management-feature-spec\", \"sp-core\", \"hf-papers\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0257",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: pattern analysis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "tab-patterns"
      ],
      "distractor_skills": [
        "tab-analysis-run",
        "tab-glossary",
        "tab-market-breadth",
        "tab-news-fetch",
        "tab-event-detect",
        "tab-report-generate",
        "tab-strategy-comparison",
        "tab-market-environment",
        "tab-llm-agents",
        "tab-turtle-refresh",
        "tab-fundamental-sync",
        "tab-genai-features",
        "tab-dualma-refresh"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: pattern analysis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"tab-patterns\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"tab-analysis-run\", \"tab-glossary\", \"tab-market-breadth\", \"tab-news-fetch\", \"tab-event-detect\", \"tab-report-generate\", \"tab-strategy-comparison\", \"tab-market-environment\", \"tab-llm-agents\", \"tab-turtle-refresh\", \"tab-fundamental-sync\", \"tab-genai-features\", \"tab-dualma-refresh\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0258",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 캘린더 브리핑. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "calendar-daily-briefing"
      ],
      "distractor_skills": [
        "gws-workflows",
        "feedback-meeting-scheduler",
        "gws-calendar",
        "daily-am-orchestrator",
        "kwp-sales-daily-briefing",
        "gws-gmail",
        "smart-meeting-scheduler",
        "gws-recipe-events-from-sheet",
        "gws-chat",
        "gws-workspace",
        "sprint-orchestrator",
        "setup-doctor",
        "ai-chief-of-staff",
        "proactive-meeting-scheduler",
        "jira-linear-adapter",
        "gws-drive",
        "simplify"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 캘린더 브리핑. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"calendar-daily-briefing\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"gws-workflows\", \"feedback-meeting-scheduler\", \"gws-calendar\", \"daily-am-orchestrator\", \"kwp-sales-daily-briefing\", \"gws-gmail\", \"smart-meeting-scheduler\", \"gws-recipe-events-from-sheet\", \"gws-chat\", \"gws-workspace\", \"sprint-orchestrator\", \"setup-doctor\", \"ai-chief-of-staff\", \"proactive-meeting-scheduler\", \"jira-linear-adapter\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0259",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: quality audit. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "mission-control"
      ],
      "distractor_skills": [
        "mission-control copy",
        "release-commander",
        "agency-agents-orchestrator",
        "workflow-parallel",
        "sefo-orchestrator",
        "agency-data-consolidation-agent",
        "ai-workflow-integrator",
        "agency-test-results-analyzer",
        "full-stack-planner",
        "workflow-sequential",
        "deep-review",
        "agency-workflow-optimizer",
        "commit-to-issue",
        "agency-behavioral-nudge-engine",
        "agency-macos-spatial-metal-engineer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-support-responder",
        "agency-agentic-identity-trust-architect",
        "release-ship",
        "agency-api-tester",
        "iac-review-agent",
        "agency-terminal-integration-specialist",
        "skill-optimizer",
        "agency-cultural-intelligence-strategist",
        "agency-app-store-optimizer",
        "sp-core",
        "ui-suite",
        "kwp-productivity-task-management",
        "gws-workflows",
        "agency-instagram-curator",
        "agency-feedback-synthesizer",
        "agency-autonomous-optimization-architect",
        "ai-chief-of-staff",
        "kwp-cowork-plugin-customizer",
        "agency-xr-interface-architect",
        "agency-devops-automator",
        "agency-ux-architect"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: quality audit. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"mission-control\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"mission-control copy\", \"release-commander\", \"agency-agents-orchestrator\", \"workflow-parallel\", \"sefo-orchestrator\", \"agency-data-consolidation-agent\", \"ai-workflow-integrator\", \"agency-test-results-analyzer\", \"full-stack-planner\", \"workflow-sequential\", \"deep-review\", \"agency-workflow-optimizer\", \"commit-to-issue\", \"agency-behavioral-nudge-engine\", \"agency-macos-spatial-metal-engineer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0260",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: references agency-autonomous-optimization-architect. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-autonomous-optimization-architect"
      ],
      "distractor_skills": [
        "agency-xr-interface-architect",
        "agency-terminal-integration-specialist",
        "agency-infrastructure-maintainer",
        "agency-app-store-optimizer",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-report-distribution-agent",
        "agency-behavioral-nudge-engine",
        "agency-agents-orchestrator",
        "agency-brand-guardian",
        "agency-ux-architect",
        "agency-data-consolidation-agent",
        "agency-visionos-spatial-engineer",
        "agency-tool-evaluator",
        "agency-macos-spatial-metal-engineer",
        "agency-api-tester",
        "agency-finance-tracker",
        "agency-cultural-intelligence-strategist",
        "agency-rapid-prototyper",
        "agency-mobile-app-builder"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: references agency-autonomous-optimization-architect. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-autonomous-optimization-architect\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-xr-interface-architect\", \"agency-terminal-integration-specialist\", \"agency-infrastructure-maintainer\", \"agency-app-store-optimizer\", \"agency-xr-immersive-developer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-report-distribution-agent\", \"agency-behavioral-nudge-engine\", \"agency-agents-orchestrator\", \"agency-brand-guardian\", \"agency-ux-architect\", \"agency-data-consolidation-agent\", \"agency-visionos-spatial-engineer\", \"agency-tool-evaluator\", \"agency-macos-spatial-metal-engineer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0261",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: choosing between Now/Next/Later. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-product-management-roadmap-management"
      ],
      "distractor_skills": [
        "kwp-product-management-metrics-tracking",
        "kwp-design-system",
        "kwp-data-data-visualization",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-interview-prep",
        "kwp-finance-close-management",
        "kwp-slack-slack-search",
        "kwp-product-management-competitive-analysis",
        "kwp-product-management-user-research-synthesis",
        "kwp-marketing-campaign-planning",
        "kwp-data-data-validation",
        "kwp-finance-financial-statements",
        "kwp-product-management-feature-spec",
        "kwp-operations-vendor-management",
        "kwp-operations-change-management",
        "kwp-data-interactive-dashboard-builder",
        "kwp-customer-support-escalation"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: choosing between Now/Next/Later. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-product-management-roadmap-management\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-product-management-metrics-tracking\", \"kwp-design-system\", \"kwp-data-data-visualization\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-interview-prep\", \"kwp-finance-close-management\", \"kwp-slack-slack-search\", \"kwp-product-management-competitive-analysis\", \"kwp-product-management-user-research-synthesis\", \"kwp-marketing-campaign-planning\", \"kwp-data-data-validation\", \"kwp-finance-financial-statements\", \"kwp-product-management-feature-spec\", \"kwp-operations-vendor-management\", \"kwp-operations-change-management\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0262",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: sign a skill. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "sefo-governance"
      ],
      "distractor_skills": [
        "sefo-federation",
        "sefo-orchestrator",
        "dependency-auditor",
        "semantic-guard",
        "release-commander",
        "kwp-engineering-code-review",
        "compliance-governance",
        "i18n-sync",
        "paperclip-tasks",
        "skill-guide-generator",
        "helm-validator",
        "paperclip-control",
        "paperclip-agents",
        "agency-infrastructure-maintainer",
        "kwp-design-design-system-management",
        "intent-alignment-tracker",
        "agency-xr-immersive-developer",
        "pm-toolkit",
        "paperclip-setup",
        "skill-optimizer",
        "security-expert",
        "autoskill-merger",
        "kwp-sync",
        "kwp-design-accessibility-review",
        "agency-security-engineer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: sign a skill. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"sefo-governance\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"sefo-federation\", \"sefo-orchestrator\", \"dependency-auditor\", \"semantic-guard\", \"release-commander\", \"kwp-engineering-code-review\", \"compliance-governance\", \"i18n-sync\", \"paperclip-tasks\", \"skill-guide-generator\", \"helm-validator\", \"paperclip-control\", \"paperclip-agents\", \"agency-infrastructure-maintainer\", \"kwp-design-design-system-management\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0263",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: when first-draft quality consistently falls short. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "workflow-eval-opt"
      ],
      "distractor_skills": [
        "workflow-sequential",
        "workflow-parallel",
        "workflow-miner",
        "skill-optimizer",
        "kwp-operations-process-optimization",
        "ralph-loop",
        "kwp-design-accessibility-review",
        "kwp-design-user-research",
        "sentence-polisher",
        "agency-test-results-analyzer",
        "role-data-scientist",
        "sefo-orchestrator"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: when first-draft quality consistently falls short. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"workflow-eval-opt\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"workflow-sequential\", \"workflow-parallel\", \"workflow-miner\", \"skill-optimizer\", \"kwp-operations-process-optimization\", \"ralph-loop\", \"kwp-design-accessibility-review\", \"kwp-design-user-research\", \"sentence-polisher\", \"agency-test-results-analyzer\", \"role-data-scientist\", \"sefo-orchestrator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0264",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Screen Description Generator. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "screen-description"
      ],
      "distractor_skills": [
        "design-review",
        "implement-screen",
        "swagger-api-doc-generator",
        "overlay-layout-patterns",
        "fsd-development",
        "fe-pipeline",
        "github-workflow-automation",
        "figma-to-tds",
        "notion-docs-sync",
        "local-dev-setup",
        "google-daily",
        "sp-writing-plans",
        "anthropic-claude-api",
        "sp-subagent-dev",
        "sp-parallel-agents",
        "kwp-slack-slack-messaging",
        "sp-writing-skills",
        "sp-tdd",
        "sp-executing-plans",
        "kwp-slack-slack-search",
        "sp-debugging",
        "sp-finish-branch",
        "tab-turtle-refresh",
        "tab-fundamental-sync",
        "sp-request-review",
        "kwp-data-data-context-extractor",
        "agency-devops-automator",
        "tab-event-detect",
        "tab-dualma-refresh",
        "tab-bollinger-refresh",
        "tab-analysis-run",
        "gws-docs",
        "agency-brand-guardian",
        "tab-technical-analysis",
        "tab-llm-agents",
        "tab-hot-stock-discovery",
        "kwp-design-system",
        "auto-research-distribute",
        "agency-agents-orchestrator",
        "tab-screening"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Screen Description Generator. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"screen-description\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"design-review\", \"implement-screen\", \"swagger-api-doc-generator\", \"overlay-layout-patterns\", \"fsd-development\", \"fe-pipeline\", \"github-workflow-automation\", \"figma-to-tds\", \"notion-docs-sync\", \"local-dev-setup\", \"google-daily\", \"sp-writing-plans\", \"anthropic-claude-api\", \"sp-subagent-dev\", \"sp-parallel-agents\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0265",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: exploring public financial datasets. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "hf-dataset-viewer"
      ],
      "distractor_skills": [
        "hf-cli",
        "hf-datasets",
        "hf-models",
        "hf-buckets",
        "hf-endpoints",
        "hf-evaluation",
        "hf-collections",
        "hf-jobs",
        "hf-repos",
        "hf-discussions",
        "hf-trackio",
        "hf-hub",
        "hf-auto-finetune-planner",
        "hf-model-trainer",
        "hf-spaces"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: exploring public financial datasets. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"hf-dataset-viewer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"hf-cli\", \"hf-datasets\", \"hf-models\", \"hf-buckets\", \"hf-endpoints\", \"hf-evaluation\", \"hf-collections\", \"hf-jobs\", \"hf-repos\", \"hf-discussions\", \"hf-trackio\", \"hf-hub\", \"hf-auto-finetune-planner\", \"hf-model-trainer\", \"hf-spaces\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0266",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: AlphaEar Sentiment. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "alphaear-sentiment"
      ],
      "distractor_skills": [
        "alphaear-deepear-lite",
        "alphaear-news",
        "alphaear-search",
        "alphaear-predictor",
        "trading-market-environment-analysis",
        "alphaear-signal-tracker",
        "alphaear-reporter",
        "alphaear-stock",
        "trading-breadth-chart-analyst",
        "trading-market-news-analyst",
        "hf-model-trainer",
        "alphaear-logic-visualizer",
        "daily-stock-check",
        "stock-csv-downloader",
        "role-trading-expert",
        "trading-market-top-detector",
        "trading-market-breadth-analyzer",
        "mirofish-financial-sim",
        "trading-uptrend-analyzer",
        "trading-theme-detector",
        "trading-us-stock-analysis",
        "hf-evaluation",
        "hf-cli",
        "trading-trade-hypothesis-ideator",
        "tab-news-fetch",
        "mirofish",
        "trading-sector-analyst",
        "weekly-stock-update",
        "tab-report-generate",
        "trading-stanley-druckenmiller-investment",
        "trading-technical-analyst",
        "tab-kiwoom",
        "qa-test-expert",
        "tab-llm-agents",
        "kwp-design-user-research",
        "hf-jobs",
        "kwp-human-resources-people-analytics",
        "service-health-doctor",
        "e2e-testing",
        "pm-toolkit",
        "trading-edge-candidate-agent"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: AlphaEar Sentiment. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-sentiment\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"alphaear-deepear-lite\", \"alphaear-news\", \"alphaear-search\", \"alphaear-predictor\", \"trading-market-environment-analysis\", \"alphaear-signal-tracker\", \"alphaear-reporter\", \"alphaear-stock\", \"trading-breadth-chart-analyst\", \"trading-market-news-analyst\", \"hf-model-trainer\", \"alphaear-logic-visualizer\", \"daily-stock-check\", \"stock-csv-downloader\", \"role-trading-expert\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0267",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: automate this process. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "pipeline-builder"
      ],
      "distractor_skills": [
        "automation-strategist",
        "ai-workflow-integrator",
        "sre-devops-expert",
        "onboarding-accelerator",
        "critical-review",
        "context-engineer",
        "cursor-automations",
        "k8s-manifest-validator",
        "performance-profiler",
        "setup-doctor",
        "helm-validator",
        "commit-to-issue",
        "plans",
        "system-thinker",
        "refactor-simulator",
        "simplify",
        "auto-research-distribute",
        "daily-strategy-post",
        "auto-research",
        "full-stack-planner",
        "ci-quality-gate",
        "executive-briefing",
        "calendar-daily-briefing",
        "code-review-all",
        "feedback-meeting-scheduler",
        "daily-stock-check",
        "codebase-archaeologist",
        "sprint-orchestrator",
        "eod-ship",
        "alphaxiv-paper-lookup",
        "problem-definition",
        "technical-writer",
        "daily-am-orchestrator",
        "agency-ai-engineer",
        "scrapling",
        "agency-devops-automator",
        "ai-quality-evaluator",
        "agency-data-engineer",
        "security-expert",
        "related-papers-scout",
        "jira-linear-adapter",
        "agency-workflow-optimizer",
        "cognee",
        "domain-commit",
        "issue-to-code",
        "skill-composer",
        "agency-agents-orchestrator",
        "weekly-status-report",
        "kwp-engineering-incident-response"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: automate this process. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"pipeline-builder\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"automation-strategist\", \"ai-workflow-integrator\", \"sre-devops-expert\", \"onboarding-accelerator\", \"critical-review\", \"context-engineer\", \"cursor-automations\", \"k8s-manifest-validator\", \"performance-profiler\", \"setup-doctor\", \"helm-validator\", \"commit-to-issue\", \"plans\", \"system-thinker\", \"refactor-simulator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0268",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: credit spread setup. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "trading-options-theta"
      ],
      "distractor_skills": [
        "trading-options-strategy-advisor",
        "trading-sector-analyst",
        "trading-position-sizer",
        "trading-market-environment-analysis",
        "trading-market-top-detector",
        "trading-breadth-chart-analyst",
        "trading-theme-detector",
        "trading-stanley-druckenmiller-investment",
        "trading-trade-hypothesis-ideator",
        "trading-market-breadth-analyzer",
        "trading-edge-candidate-agent",
        "trading-uptrend-analyzer",
        "trading-us-stock-analysis",
        "trading-technical-analyst",
        "trading-market-news-analyst",
        "trading-backtest-expert",
        "trading-us-market-bubble-detector",
        "trading-data-quality-checker",
        "trading-finviz-screener",
        "mirofish-financial-sim",
        "trading-scenario-analyzer",
        "stock-csv-downloader",
        "daily-stock-check",
        "role-trading-expert",
        "alphaear-sentiment",
        "weekly-stock-update",
        "mirofish",
        "today",
        "alphaear-stock",
        "alphaear-deepear-lite",
        "tab-strategy-comparison",
        "daily-am-orchestrator",
        "tab-kiwoom",
        "morning-ship",
        "e2e-testing",
        "terraform-reviewer",
        "alphaear-reporter",
        "mirofish-opinion-sim",
        "paperclip-agents",
        "hf-cli",
        "cursor-automations",
        "kwp-productivity-task-management",
        "k8s-manifest-validator",
        "daily-strategy-post",
        "intent-alignment-tracker",
        "ai-quality-evaluator",
        "tab-stock-sync",
        "paperclip-tasks",
        "domain-commit",
        "nlm-deep-learn"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: credit spread setup. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"trading-options-theta\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"trading-options-strategy-advisor\", \"trading-sector-analyst\", \"trading-position-sizer\", \"trading-market-environment-analysis\", \"trading-market-top-detector\", \"trading-breadth-chart-analyst\", \"trading-theme-detector\", \"trading-stanley-druckenmiller-investment\", \"trading-trade-hypothesis-ideator\", \"trading-market-breadth-analyzer\", \"trading-edge-candidate-agent\", \"trading-uptrend-analyzer\", \"trading-us-stock-analysis\", \"trading-technical-analyst\", \"trading-market-news-analyst\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0269",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Lsp Index Engineer agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-lsp-index-engineer"
      ],
      "distractor_skills": [
        "agency-xr-cockpit-interaction-specialist",
        "agency-macos-spatial-metal-engineer",
        "agency-terminal-integration-specialist",
        "agency-xr-interface-architect",
        "agency-ux-architect",
        "agency-visionos-spatial-engineer",
        "agency-app-store-optimizer",
        "agency-xr-immersive-developer",
        "agency-api-tester",
        "agency-behavioral-nudge-engine",
        "agency-brand-guardian",
        "agency-cultural-intelligence-strategist",
        "agency-senior-developer",
        "agency-tiktok-strategist",
        "agency-tool-evaluator",
        "agency-evidence-collector",
        "agency-report-distribution-agent",
        "agency-reddit-community-builder",
        "agency-instagram-curator",
        "agency-test-results-analyzer",
        "agency-inclusive-visuals-specialist",
        "agency-workflow-optimizer",
        "agency-twitter-engager",
        "agency-ai-engineer",
        "agency-rapid-prototyper",
        "agency-mobile-app-builder",
        "agency-data-consolidation-agent",
        "agency-xiaohongshu-specialist",
        "agency-senior-project-manager",
        "agency-whimsy-injector",
        "agency-sales-data-extraction-agent",
        "agency-zhihu-strategist"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Lsp Index Engineer agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-lsp-index-engineer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-xr-cockpit-interaction-specialist\", \"agency-macos-spatial-metal-engineer\", \"agency-terminal-integration-specialist\", \"agency-xr-interface-architect\", \"agency-ux-architect\", \"agency-visionos-spatial-engineer\", \"agency-app-store-optimizer\", \"agency-xr-immersive-developer\", \"agency-api-tester\", \"agency-behavioral-nudge-engine\", \"agency-brand-guardian\", \"agency-cultural-intelligence-strategist\", \"agency-senior-developer\", \"agency-tiktok-strategist\", \"agency-tool-evaluator\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0270",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: competitor analysis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "pm-market-research"
      ],
      "distractor_skills": [
        "pm-marketing-growth",
        "pm-go-to-market",
        "pm-product-discovery",
        "pm-product-strategy",
        "pm-execution",
        "pm-data-analytics",
        "pm-toolkit",
        "role-cso",
        "agency-trend-researcher",
        "kwp-marketing-campaign-planning",
        "kwp-marketing-competitive-analysis",
        "agency-ux-researcher",
        "e2e-testing",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-report-distribution-agent",
        "agency-app-store-optimizer",
        "kwp-product-management-competitive-analysis",
        "agency-xr-interface-architect",
        "trading-market-environment-analysis",
        "agency-data-consolidation-agent",
        "agency-lsp-index-engineer",
        "role-sales",
        "agency-sales-data-extraction-agent",
        "agency-terminal-integration-specialist",
        "agency-brand-guardian",
        "agency-workflow-optimizer",
        "agency-cultural-intelligence-strategist",
        "agency-visionos-spatial-engineer",
        "agency-performance-benchmarker",
        "agency-tiktok-strategist",
        "nlm-strategy-hub",
        "agency-reddit-community-builder",
        "agency-instagram-curator",
        "agency-feedback-synthesizer",
        "agency-data-analytics-reporter",
        "agency-analytics-reporter",
        "mission-control copy",
        "kwp-product-management-user-research-synthesis",
        "agency-behavioral-nudge-engine",
        "alphaear-deepear-lite",
        "agency-sprint-prioritizer",
        "agency-finance-tracker",
        "agency-devops-automator"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: competitor analysis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"pm-market-research\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"pm-marketing-growth\", \"pm-go-to-market\", \"pm-product-discovery\", \"pm-product-strategy\", \"pm-execution\", \"pm-data-analytics\", \"pm-toolkit\", \"role-cso\", \"agency-trend-researcher\", \"kwp-marketing-campaign-planning\", \"kwp-marketing-competitive-analysis\", \"agency-ux-researcher\", \"e2e-testing\", \"agency-xr-immersive-developer\", \"agency-xr-cockpit-interaction-specialist\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0271",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Xiaohongshu Specialist agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-xiaohongshu-specialist"
      ],
      "distractor_skills": [
        "agency-reddit-community-builder",
        "agency-instagram-curator",
        "agency-tiktok-strategist",
        "agency-twitter-engager",
        "agency-zhihu-strategist",
        "agency-app-store-optimizer",
        "agency-wechat-official-account-manager",
        "agency-growth-hacker",
        "agency-tool-evaluator",
        "agency-visual-storyteller",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-test-results-analyzer",
        "agency-api-tester",
        "agency-terminal-integration-specialist",
        "agency-brand-guardian",
        "agency-xr-interface-architect",
        "agency-workflow-optimizer",
        "agency-developer-advocate",
        "agency-senior-developer",
        "agency-legal-compliance-checker",
        "agency-content-creator",
        "agency-lsp-index-engineer",
        "agency-whimsy-injector",
        "agency-experiment-tracker",
        "agency-inclusive-visuals-specialist",
        "agency-behavioral-nudge-engine",
        "agency-infrastructure-maintainer",
        "agency-ux-architect",
        "agency-rapid-prototyper",
        "agency-social-media-strategist",
        "agency-visionos-spatial-engineer",
        "agency-trend-researcher",
        "agency-senior-project-manager",
        "agency-macos-spatial-metal-engineer",
        "agency-evidence-collector",
        "agency-report-distribution-agent",
        "agency-finance-tracker",
        "agency-cultural-intelligence-strategist",
        "agency-mobile-app-builder",
        "agency-data-consolidation-agent"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Xiaohongshu Specialist agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-xiaohongshu-specialist\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-reddit-community-builder\", \"agency-instagram-curator\", \"agency-tiktok-strategist\", \"agency-twitter-engager\", \"agency-zhihu-strategist\", \"agency-app-store-optimizer\", \"agency-wechat-official-account-manager\", \"agency-growth-hacker\", \"agency-tool-evaluator\", \"agency-visual-storyteller\", \"agency-xr-immersive-developer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-test-results-analyzer\", \"agency-api-tester\", \"agency-terminal-integration-specialist\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0272",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: wants automated meeting proposals from sprint activity. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "feedback-meeting-scheduler"
      ],
      "distractor_skills": [
        "proactive-meeting-scheduler",
        "sprint-orchestrator",
        "calendar-daily-briefing",
        "jira-linear-adapter",
        "pr-to-issue-linker",
        "gws-calendar",
        "standup-digest",
        "email-auto-reply",
        "commit-to-issue",
        "critical-review",
        "related-papers-scout",
        "daily-am-orchestrator",
        "plans",
        "auto-research",
        "domain-commit",
        "smart-meeting-scheduler",
        "helm-validator",
        "github-sprint-digest",
        "alphaxiv-paper-lookup",
        "cursor-automations",
        "pipeline-builder",
        "daily-strategy-post",
        "refactor-simulator",
        "ppt-template-engine",
        "security-expert",
        "simplify",
        "technical-writer",
        "automation-strategist",
        "daily-stock-check",
        "context-engineer",
        "performance-profiler",
        "auto-research-distribute",
        "executive-briefing",
        "decision-router",
        "eod-ship",
        "ci-quality-gate",
        "gws-email-reply",
        "gws-recipe-events-from-sheet",
        "setup-doctor",
        "code-review-all",
        "problem-definition",
        "sre-devops-expert",
        "scrapling",
        "gws-workspace",
        "onboarding-accelerator"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: wants automated meeting proposals from sprint activity. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"feedback-meeting-scheduler\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"proactive-meeting-scheduler\", \"sprint-orchestrator\", \"calendar-daily-briefing\", \"jira-linear-adapter\", \"pr-to-issue-linker\", \"gws-calendar\", \"standup-digest\", \"email-auto-reply\", \"commit-to-issue\", \"critical-review\", \"related-papers-scout\", \"daily-am-orchestrator\", \"plans\", \"auto-research\", \"domain-commit\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0273",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: data impact analysis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "role-data-scientist"
      ],
      "distractor_skills": [
        "role-hr",
        "role-developer",
        "role-finance",
        "role-ux-designer",
        "role-sales",
        "role-security-engineer",
        "role-trading-expert",
        "role-pm",
        "role-cso",
        "role-cto",
        "role-ceo",
        "deep-research-pipeline",
        "role-dispatcher",
        "pm-data-analytics",
        "paper-review",
        "executive-briefing",
        "kwp-operations-compliance-tracking",
        "commit-to-issue",
        "qa-test-expert",
        "intent-alignment-tracker",
        "agency-data-consolidation-agent",
        "kwp-human-resources-people-analytics",
        "agency-test-results-analyzer",
        "hf-evaluation",
        "trading-market-news-analyst",
        "agency-data-analytics-reporter",
        "agency-analytics-reporter",
        "kwp-design-accessibility-review",
        "sentence-polisher",
        "kwp-engineering-system-design",
        "agency-agents-orchestrator",
        "meeting-digest"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: data impact analysis. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"role-data-scientist\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"role-hr\", \"role-developer\", \"role-finance\", \"role-ux-designer\", \"role-sales\", \"role-security-engineer\", \"role-trading-expert\", \"role-pm\", \"role-cso\", \"role-cto\", \"role-ceo\", \"deep-research-pipeline\", \"role-dispatcher\", \"pm-data-analytics\", \"paper-review\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0274",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: proposal factory. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "proposal-factory"
      ],
      "distractor_skills": [
        "prd-research-factory",
        "feature-intake-router",
        "commit-to-issue",
        "technical-writer",
        "incident-lifecycle-manager",
        "sales-meeting-pipeline",
        "full-stack-planner",
        "domain-commit",
        "tech-support-brain",
        "ci-quality-gate",
        "sales-weekly-dashboard",
        "code-review-all",
        "plans",
        "auto-research",
        "auto-research-distribute",
        "critical-review",
        "alphaxiv-paper-lookup",
        "security-expert",
        "problem-definition",
        "daily-am-orchestrator",
        "ppt-template-engine",
        "onboarding-accelerator"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: proposal factory. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"proposal-factory\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"prd-research-factory\", \"feature-intake-router\", \"commit-to-issue\", \"technical-writer\", \"incident-lifecycle-manager\", \"sales-meeting-pipeline\", \"full-stack-planner\", \"domain-commit\", \"tech-support-brain\", \"ci-quality-gate\", \"sales-weekly-dashboard\", \"code-review-all\", \"plans\", \"auto-research\", \"auto-research-distribute\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0275",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: auto digest meetings. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "planning-meeting-sentinel"
      ],
      "distractor_skills": [
        "planning-weekly-pulse",
        "lead-lifecycle-manager",
        "sales-meeting-pipeline",
        "docs-freshness-guardian",
        "sales-weekly-dashboard",
        "dependency-radar",
        "feature-intake-router",
        "prd-research-factory",
        "notion-meeting-sync",
        "gws-calendar",
        "gmail-daily-triage",
        "ship",
        "gws-gmail",
        "kwp-operations-change-management",
        "gws-workflows",
        "kwp-legal-meeting-briefing",
        "feedback-meeting-scheduler",
        "tech-support-brain",
        "incident-lifecycle-manager",
        "proactive-meeting-scheduler",
        "gws-recipe-events-from-sheet",
        "meeting-digest",
        "proposal-factory",
        "gws-sheets",
        "md-enhance-publish",
        "anthropic-skill-creator",
        "kwp-finance-close-management",
        "gws-chat",
        "eod-ship",
        "x-to-slack",
        "ai-chief-of-staff",
        "gws-docs",
        "calendar-daily-briefing",
        "ecc-strategic-compact"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: auto digest meetings. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"planning-meeting-sentinel\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"planning-weekly-pulse\", \"lead-lifecycle-manager\", \"sales-meeting-pipeline\", \"docs-freshness-guardian\", \"sales-weekly-dashboard\", \"dependency-radar\", \"feature-intake-router\", \"prd-research-factory\", \"notion-meeting-sync\", \"gws-calendar\", \"gmail-daily-triage\", \"ship\", \"gws-gmail\", \"kwp-operations-change-management\", \"gws-workflows\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0276",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 발표 설계. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "presentation-strategist"
      ],
      "distractor_skills": [
        "qa-test-expert",
        "e2e-testing",
        "mirofish-graph-explorer",
        "kwp-engineering-testing-strategy",
        "nlm-strategy-hub",
        "notebooklm-studio",
        "notion-meeting-sync",
        "ci-quality-gate",
        "kwp-human-resources-org-planning",
        "nlm-arxiv-slides",
        "kwp-brand-voice-brand-voice-enforcement",
        "technical-writer",
        "frontend-expert",
        "agency-instagram-curator",
        "nlm-discovery-lab",
        "kwp-marketing-campaign-planning",
        "pm-data-analytics",
        "kwp-cowork-create-cowork-plugin"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 발표 설계. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"presentation-strategist\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"qa-test-expert\", \"e2e-testing\", \"mirofish-graph-explorer\", \"kwp-engineering-testing-strategy\", \"nlm-strategy-hub\", \"notebooklm-studio\", \"notion-meeting-sync\", \"ci-quality-gate\", \"kwp-human-resources-org-planning\", \"nlm-arxiv-slides\", \"kwp-brand-voice-brand-voice-enforcement\", \"technical-writer\", \"frontend-expert\", \"agency-instagram-curator\", \"nlm-discovery-lab\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0277",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: 터틀 지표 갱신. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "tab-turtle-refresh"
      ],
      "distractor_skills": [
        "tab-dualma-refresh",
        "tab-bollinger-refresh",
        "tab-strategy-comparison",
        "tab-market-breadth",
        "tab-llm-agents",
        "tab-technical-analysis",
        "tab-fundamental-sync",
        "tab-report-generate",
        "tab-market-environment",
        "tab-hot-stock-discovery",
        "tab-screening",
        "tab-news-fetch"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: 터틀 지표 갱신. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"tab-turtle-refresh\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"tab-dualma-refresh\", \"tab-bollinger-refresh\", \"tab-strategy-comparison\", \"tab-market-breadth\", \"tab-llm-agents\", \"tab-technical-analysis\", \"tab-fundamental-sync\", \"tab-report-generate\", \"tab-market-environment\", \"tab-hot-stock-discovery\", \"tab-screening\", \"tab-news-fetch\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0278",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: identify assumptions. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "pm-product-discovery"
      ],
      "distractor_skills": [
        "pm-market-research",
        "pm-marketing-growth",
        "pm-product-strategy",
        "pm-execution",
        "pm-go-to-market",
        "pm-data-analytics",
        "pm-toolkit",
        "qa-test-expert",
        "kwp-product-management-user-research-synthesis",
        "nlm-discovery-lab",
        "e2e-testing",
        "agency-sprint-prioritizer",
        "agency-ux-researcher",
        "role-pm",
        "role-cso",
        "kwp-marketing-campaign-planning",
        "pm-sentinel",
        "kwp-product-management-competitive-analysis",
        "agency-executive-summary-generator",
        "agency-experiment-tracker",
        "evals-skills",
        "kwp-marketing-competitive-analysis",
        "kwp-product-management-metrics-tracking",
        "role-ceo",
        "frontend-expert",
        "kwp-human-resources-interview-prep",
        "kwp-product-management-roadmap-management",
        "kwp-design-user-research",
        "agency-trend-researcher",
        "kwp-engineering-testing-strategy",
        "kwp-engineering-tech-debt",
        "role-ux-designer",
        "ci-quality-gate",
        "kwp-customer-support-response-drafting",
        "hf-hub",
        "x-to-slack",
        "skill-optimizer",
        "kwp-human-resources-org-planning"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: identify assumptions. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"pm-product-discovery\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"pm-market-research\", \"pm-marketing-growth\", \"pm-product-strategy\", \"pm-execution\", \"pm-go-to-market\", \"pm-data-analytics\", \"pm-toolkit\", \"qa-test-expert\", \"kwp-product-management-user-research-synthesis\", \"nlm-discovery-lab\", \"e2e-testing\", \"agency-sprint-prioritizer\", \"agency-ux-researcher\", \"role-pm\", \"role-cso\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0279",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: People Analytics. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-human-resources-people-analytics"
      ],
      "distractor_skills": [
        "kwp-human-resources-interview-prep",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-org-planning",
        "kwp-operations-vendor-management",
        "kwp-data-data-context-extractor",
        "kwp-operations-process-optimization",
        "kwp-product-management-competitive-analysis",
        "kwp-data-data-exploration",
        "kwp-marketing-performance-analytics",
        "kwp-data-data-validation",
        "kwp-design-design-handoff",
        "kwp-product-management-feature-spec",
        "kwp-operations-change-management",
        "kwp-operations-resource-planning",
        "kwp-data-statistical-analysis",
        "kwp-finance-variance-analysis",
        "kwp-finance-financial-statements",
        "kwp-slack-slack-search",
        "kwp-sales-call-prep",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-data-interactive-dashboard-builder",
        "kwp-finance-close-management",
        "kwp-design-system",
        "kwp-finance-reconciliation",
        "kwp-common-room-account-research",
        "kwp-sales-daily-briefing",
        "kwp-sales-account-research",
        "kwp-product-management-user-research-synthesis",
        "kwp-marketing-campaign-planning",
        "kwp-legal-legal-risk-assessment",
        "kwp-brand-voice-guideline-generation",
        "kwp-data-data-visualization",
        "kwp-sales-draft-outreach",
        "kwp-sales-competitive-intelligence",
        "kwp-legal-nda-triage",
        "kwp-operations-compliance-tracking",
        "kwp-legal-canned-responses",
        "kwp-finance-audit-support",
        "kwp-enterprise-search-knowledge-synthesis"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: People Analytics. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-human-resources-people-analytics\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-human-resources-interview-prep\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-product-management-metrics-tracking\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-human-resources-org-planning\", \"kwp-operations-vendor-management\", \"kwp-data-data-context-extractor\", \"kwp-operations-process-optimization\", \"kwp-product-management-competitive-analysis\", \"kwp-data-data-exploration\", \"kwp-marketing-performance-analytics\", \"kwp-data-data-validation\", \"kwp-design-design-handoff\", \"kwp-product-management-feature-spec\", \"kwp-operations-change-management\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0280",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Change Management. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "kwp-operations-change-management"
      ],
      "distractor_skills": [
        "kwp-operations-process-optimization",
        "kwp-operations-vendor-management",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-interview-prep",
        "kwp-operations-resource-planning",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-people-analytics",
        "kwp-slack-slack-search",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-finance-close-management",
        "kwp-design-system",
        "kwp-marketing-campaign-planning",
        "kwp-engineering-system-design",
        "kwp-sales-draft-outreach",
        "kwp-sales-competitive-intelligence",
        "kwp-engineering-testing-strategy",
        "kwp-finance-variance-analysis",
        "kwp-sales-call-prep",
        "kwp-product-management-feature-spec",
        "kwp-data-interactive-dashboard-builder",
        "kwp-design-design-handoff"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Change Management. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"kwp-operations-change-management\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"kwp-operations-process-optimization\", \"kwp-operations-vendor-management\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-interview-prep\", \"kwp-operations-resource-planning\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-human-resources-people-analytics\", \"kwp-slack-slack-search\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-finance-close-management\", \"kwp-design-system\", \"kwp-marketing-campaign-planning\", \"kwp-engineering-system-design\", \"kwp-sales-draft-outreach\", \"kwp-sales-competitive-intelligence\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0281",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: AI Platform FE - FSD 개발 및 마이그레이션. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "fsd-development"
      ],
      "distractor_skills": [
        "design-review",
        "implement-screen",
        "figma-to-tds",
        "screen-description",
        "overlay-layout-patterns",
        "swagger-api-doc-generator",
        "github-workflow-automation",
        "fe-pipeline",
        "notion-docs-sync",
        "local-dev-setup",
        "sp-request-review",
        "tab-llm-agents",
        "kwp-design-system",
        "google-daily",
        "tab-screening",
        "sp-writing-plans",
        "tab-genai-features",
        "anthropic-web-artifacts-builder",
        "sp-subagent-dev",
        "sp-parallel-agents",
        "kwp-slack-slack-messaging",
        "anthropic-frontend-design",
        "anthropic-webapp-testing",
        "sp-writing-skills",
        "sp-tdd",
        "sp-executing-plans",
        "kwp-slack-slack-search",
        "sp-brainstorming",
        "sp-debugging",
        "anthropic-theme-factory",
        "e2e-testing",
        "agency-frontend-developer",
        "tab-turtle-refresh",
        "tab-fundamental-sync",
        "kwp-data-data-context-extractor",
        "tab-event-detect",
        "tab-dualma-refresh",
        "tab-bollinger-refresh",
        "tab-analysis-run"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: AI Platform FE - FSD 개발 및 마이그레이션. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"fsd-development\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"design-review\", \"implement-screen\", \"figma-to-tds\", \"screen-description\", \"overlay-layout-patterns\", \"swagger-api-doc-generator\", \"github-workflow-automation\", \"fe-pipeline\", \"notion-docs-sync\", \"local-dev-setup\", \"sp-request-review\", \"tab-llm-agents\", \"kwp-design-system\", \"google-daily\", \"tab-screening\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0282",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: RSS event scan. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "tab-event-detect"
      ],
      "distractor_skills": [
        "tab-analysis-run",
        "tab-patterns",
        "tab-fundamental-sync",
        "tab-genai-features",
        "tab-screening",
        "tab-turtle-refresh",
        "tab-strategy-comparison",
        "tab-dualma-refresh",
        "tab-technical-analysis",
        "tab-llm-agents",
        "tab-hot-stock-discovery",
        "tab-market-breadth",
        "tab-dashboard-summary",
        "tab-news-fetch",
        "tab-report-generate",
        "tab-stock-sync",
        "tab-bollinger-refresh",
        "tab-glossary",
        "tab-market-environment",
        "tab-kiwoom",
        "gws-recipe-events-from-sheet",
        "kwp-human-resources-people-analytics",
        "kwp-data-statistical-analysis",
        "skill-seekers",
        "hf-dataset-viewer",
        "autoskill-pipeline",
        "daily-stock-check",
        "anthropic-skill-creator",
        "agency-rapid-prototyper",
        "email-research-dispatcher",
        "kwp-data-data-context-extractor",
        "fe-pipeline",
        "kwp-design-design-handoff",
        "morning-ship",
        "swagger-api-doc-generator",
        "kwp-bio-research-nextflow-development",
        "sefo-orchestrator",
        "x-to-slack",
        "kwp-data-data-validation",
        "gws-recipe-doc-from-template",
        "e2e-testing",
        "anthropic-claude-api",
        "trading-market-news-analyst",
        "cursor-automations",
        "hf-cli",
        "sp-writing-plans",
        "sp-subagent-dev",
        "sp-parallel-agents"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: RSS event scan. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"tab-event-detect\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"tab-analysis-run\", \"tab-patterns\", \"tab-fundamental-sync\", \"tab-genai-features\", \"tab-screening\", \"tab-turtle-refresh\", \"tab-strategy-comparison\", \"tab-dualma-refresh\", \"tab-technical-analysis\", \"tab-llm-agents\", \"tab-hot-stock-discovery\", \"tab-market-breadth\", \"tab-dashboard-summary\", \"tab-news-fetch\", \"tab-report-generate\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0283",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: ad-hoc historical price queries. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "alphaear-stock"
      ],
      "distractor_skills": [
        "stock-csv-downloader",
        "alphaear-search",
        "alphaear-deepear-lite",
        "weekly-stock-update",
        "daily-stock-check",
        "alphaear-news",
        "tab-stock-sync",
        "alphaear-predictor",
        "alphaear-sentiment",
        "trading-sector-analyst",
        "alphaear-reporter",
        "alphaear-signal-tracker",
        "trading-technical-analyst",
        "trading-us-stock-analysis"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: ad-hoc historical price queries. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-stock\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"stock-csv-downloader\", \"alphaear-search\", \"alphaear-deepear-lite\", \"weekly-stock-update\", \"daily-stock-check\", \"alphaear-news\", \"tab-stock-sync\", \"alphaear-predictor\", \"alphaear-sentiment\", \"trading-sector-analyst\", \"alphaear-reporter\", \"alphaear-signal-tracker\", \"trading-technical-analyst\", \"trading-us-stock-analysis\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0284",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: multi-engine aggregation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "alphaear-search"
      ],
      "distractor_skills": [
        "alphaear-news",
        "alphaear-stock",
        "alphaear-deepear-lite",
        "alphaear-sentiment",
        "alphaear-predictor",
        "stock-csv-downloader",
        "alphaear-signal-tracker",
        "alphaear-logic-visualizer",
        "alphaear-reporter",
        "daily-stock-check",
        "trading-market-news-analyst",
        "hf-papers",
        "trading-market-environment-analysis",
        "cognee",
        "tab-stock-sync",
        "weekly-stock-update",
        "trading-technical-analyst",
        "trading-breadth-chart-analyst",
        "today",
        "tab-news-fetch",
        "trading-sector-analyst",
        "paper-auto-classifier",
        "email-research-dispatcher",
        "deep-research-pipeline",
        "recall",
        "daiso-mcp",
        "agency-ux-researcher",
        "kwp-human-resources-people-analytics",
        "mirofish-financial-sim",
        "kwp-engineering-documentation",
        "gws-recipe-events-from-sheet",
        "gws-recipe-doc-from-template",
        "gws-drive",
        "agency-frontend-developer",
        "local-dev-runner",
        "deep-review",
        "daily-strategy-post",
        "gws-docs"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: multi-engine aggregation. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-search\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"alphaear-news\", \"alphaear-stock\", \"alphaear-deepear-lite\", \"alphaear-sentiment\", \"alphaear-predictor\", \"stock-csv-downloader\", \"alphaear-signal-tracker\", \"alphaear-logic-visualizer\", \"alphaear-reporter\", \"daily-stock-check\", \"trading-market-news-analyst\", \"hf-papers\", \"trading-market-environment-analysis\", \"cognee\", \"tab-stock-sync\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0285",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: start of day sync. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "sod-ship"
      ],
      "distractor_skills": [
        "eod-ship",
        "morning-ship",
        "morning",
        "domain-commit",
        "release-ship",
        "daily-am-orchestrator",
        "ship",
        "commit-to-issue",
        "cursor-sync",
        "jira-linear-adapter",
        "portfolio-report-generator",
        "sp-git-worktrees",
        "release-commander",
        "daily-strategy-post",
        "calendar-daily-briefing",
        "gws-workspace",
        "kwp-sync",
        "hf-model-arena",
        "kwp-sales-daily-briefing",
        "trading-data-quality-checker",
        "knowledge-daily-aggregator",
        "kwp-operations-resource-planning",
        "pr-to-issue-linker",
        "plans",
        "hf-buckets",
        "deep-research-pipeline"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: start of day sync. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"sod-ship\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"eod-ship\", \"morning-ship\", \"morning\", \"domain-commit\", \"release-ship\", \"daily-am-orchestrator\", \"ship\", \"commit-to-issue\", \"cursor-sync\", \"jira-linear-adapter\", \"portfolio-report-generator\", \"sp-git-worktrees\", \"release-commander\", \"daily-strategy-post\", \"calendar-daily-briefing\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0286",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: create a bucket. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "hf-buckets"
      ],
      "distractor_skills": [
        "hf-repos",
        "hf-discussions",
        "hf-models",
        "hf-collections",
        "hf-endpoints",
        "hf-cli",
        "hf-hub",
        "hf-spaces",
        "hf-papers",
        "hf-dataset-viewer",
        "gws-drive",
        "hf-datasets",
        "gws-docs",
        "gws-sheets",
        "hf-auto-finetune-planner",
        "gws-chat",
        "hf-model-trainer",
        "hf-evaluation",
        "hf-model-arena",
        "hf-trackio",
        "hf-jobs",
        "gws-gmail",
        "hf-trending-intelligence",
        "domain-commit",
        "gws-recipe-doc-from-template",
        "gws-calendar",
        "skill-optimizer",
        "gws-workspace",
        "commit-to-issue",
        "md-to-slack-canvas",
        "gws-recipe-audit-sharing",
        "docs-tutor-setup",
        "kwp-design-design-system-management",
        "paperclip-tasks",
        "gws-recipe-events-from-sheet",
        "ship",
        "kwp-engineering-code-review",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: create a bucket. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"hf-buckets\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"hf-repos\", \"hf-discussions\", \"hf-models\", \"hf-collections\", \"hf-endpoints\", \"hf-cli\", \"hf-hub\", \"hf-spaces\", \"hf-papers\", \"hf-dataset-viewer\", \"gws-drive\", \"hf-datasets\", \"gws-docs\", \"gws-sheets\", \"hf-auto-finetune-planner\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0287",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: wants to view/manage RTK analytics and configuration. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "rtk"
      ],
      "distractor_skills": [
        "plans",
        "skill-optimizer",
        "workflow-miner",
        "gws-gmail",
        "paperclip-control",
        "gws-chat",
        "context-engineer",
        "intent-alignment-tracker",
        "gws-calendar",
        "hf-models",
        "gws-drive",
        "gws-docs",
        "hf-discussions",
        "hf-papers",
        "trading-stanley-druckenmiller-investment",
        "gws-sheets",
        "standup-digest",
        "paper-auto-classifier",
        "kwp-design-design-system-management",
        "hf-endpoints",
        "daiso-mcp",
        "service-health-doctor",
        "paperclip-tasks",
        "pr-to-issue-linker",
        "notebooklm-research",
        "trading-market-top-detector",
        "trading-edge-candidate-agent",
        "semantic-guard",
        "pm-toolkit",
        "local-dev-runner",
        "agency-roster",
        "sefo-governance",
        "hf-hub",
        "air-gap-orchestrator",
        "trading-breadth-chart-analyst",
        "kwp-human-resources-recruiting-pipeline",
        "agency-frontend-developer",
        "kwp-design-system",
        "paperclip-setup",
        "agency-terminal-integration-specialist",
        "kwp-engineering-code-review",
        "gws-workspace",
        "agency-performance-benchmarker",
        "agency-app-store-optimizer",
        "sefo-federation",
        "weekly-status-report",
        "i18n-sync",
        "hf-collections",
        "hf-cli",
        "feedback-meeting-scheduler"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: wants to view/manage RTK analytics and configuration. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"rtk\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"plans\", \"skill-optimizer\", \"workflow-miner\", \"gws-gmail\", \"paperclip-control\", \"gws-chat\", \"context-engineer\", \"intent-alignment-tracker\", \"gws-calendar\", \"hf-models\", \"gws-drive\", \"gws-docs\", \"hf-discussions\", \"hf-papers\", \"trading-stanley-druckenmiller-investment\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0288",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: evaluating fine-tuned financial sentiment models. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "hf-evaluation"
      ],
      "distractor_skills": [
        "hf-trackio",
        "hf-jobs",
        "hf-cli",
        "hf-model-trainer",
        "hf-datasets",
        "hf-dataset-viewer",
        "hf-models",
        "hf-endpoints",
        "hf-model-arena",
        "hf-hub",
        "hf-auto-finetune-planner",
        "ecc-eval-harness",
        "hf-buckets",
        "hf-repos",
        "hf-discussions",
        "hf-papers",
        "hf-collections",
        "hf-spaces",
        "hf-trending-intelligence",
        "alphaear-sentiment",
        "tab-report-generate"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: evaluating fine-tuned financial sentiment models. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"hf-evaluation\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"hf-trackio\", \"hf-jobs\", \"hf-cli\", \"hf-model-trainer\", \"hf-datasets\", \"hf-dataset-viewer\", \"hf-models\", \"hf-endpoints\", \"hf-model-arena\", \"hf-hub\", \"hf-auto-finetune-planner\", \"ecc-eval-harness\", \"hf-buckets\", \"hf-repos\", \"hf-discussions\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0289",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: wants to turn local markdown into structured Notion pages. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "md-to-notion"
      ],
      "distractor_skills": [
        "md-to-slack-canvas",
        "md-enhance-publish",
        "meeting-digest",
        "gws-docs",
        "docs-tutor-setup",
        "hf-buckets",
        "deep-research-pipeline",
        "notion-meeting-sync",
        "paper-review",
        "release-ship",
        "nlm-video",
        "github-sprint-digest",
        "gws-drive",
        "issue-to-code",
        "commit-to-issue",
        "cursor-sync",
        "ai-chief-of-staff",
        "transcribee",
        "weekly-status-report",
        "gws-sheets",
        "defuddle",
        "kwp-brand-voice-guideline-generation",
        "visual-explainer",
        "knowledge-daily-aggregator",
        "agency-data-consolidation-agent",
        "anthropic-pdf",
        "onboarding-accelerator",
        "nlm-slides",
        "nlm-arxiv-slides",
        "demo-forge",
        "anthropic-claude-api",
        "agency-xr-immersive-developer",
        "ralph-loop"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: wants to turn local markdown into structured Notion pages. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"md-to-notion\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"md-to-slack-canvas\", \"md-enhance-publish\", \"meeting-digest\", \"gws-docs\", \"docs-tutor-setup\", \"hf-buckets\", \"deep-research-pipeline\", \"notion-meeting-sync\", \"paper-review\", \"release-ship\", \"nlm-video\", \"github-sprint-digest\", \"gws-drive\", \"issue-to-code\", \"commit-to-issue\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0290",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: push cursor config to other repos. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "cursor-sync"
      ],
      "distractor_skills": [
        "eod-ship",
        "domain-commit",
        "simplify",
        "jira-linear-adapter",
        "daily-am-orchestrator",
        "plans",
        "commit-to-issue",
        "cursor-automations",
        "sod-ship",
        "full-stack-planner",
        "k8s-manifest-validator",
        "daily-stock-check",
        "refactor-simulator",
        "daily-strategy-post",
        "auto-research",
        "auto-research-distribute",
        "helm-validator",
        "setup-doctor",
        "ci-quality-gate",
        "code-review-all",
        "related-papers-scout",
        "feedback-meeting-scheduler",
        "ppt-template-engine",
        "codebase-archaeologist",
        "performance-profiler",
        "alphaxiv-paper-lookup",
        "pipeline-builder",
        "problem-definition",
        "scrapling",
        "agent-browser"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: push cursor config to other repos. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"cursor-sync\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"eod-ship\", \"domain-commit\", \"simplify\", \"jira-linear-adapter\", \"daily-am-orchestrator\", \"plans\", \"commit-to-issue\", \"cursor-automations\", \"sod-ship\", \"full-stack-planner\", \"k8s-manifest-validator\", \"daily-stock-check\", \"refactor-simulator\", \"daily-strategy-post\", \"auto-research\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0291",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: references agency-mobile-app-builder. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-mobile-app-builder"
      ],
      "distractor_skills": [
        "agency-xr-immersive-developer",
        "agency-rapid-prototyper",
        "agency-xr-interface-architect",
        "agency-brand-guardian",
        "agency-visionos-spatial-engineer",
        "agency-app-store-optimizer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-terminal-integration-specialist",
        "agency-ux-architect",
        "agency-data-consolidation-agent",
        "agency-senior-developer",
        "agency-macos-spatial-metal-engineer",
        "agency-report-distribution-agent",
        "agency-sales-data-extraction-agent",
        "agency-behavioral-nudge-engine",
        "agency-agents-orchestrator",
        "agency-developer-advocate",
        "agency-social-media-strategist",
        "agency-tool-evaluator",
        "agency-senior-project-manager",
        "agency-lsp-index-engineer",
        "agency-evidence-collector",
        "agency-reddit-community-builder",
        "agency-instagram-curator",
        "agency-test-results-analyzer",
        "agency-inclusive-visuals-specialist",
        "agency-api-tester",
        "agency-infrastructure-maintainer",
        "agency-workflow-optimizer",
        "agency-cultural-intelligence-strategist",
        "agency-ai-engineer",
        "agency-reality-checker",
        "agency-legal-compliance-checker",
        "agency-content-creator",
        "agency-tiktok-strategist",
        "agency-feedback-synthesizer",
        "agency-security-engineer",
        "agency-zhihu-strategist",
        "agency-frontend-developer",
        "agency-finance-tracker",
        "agency-autonomous-optimization-architect",
        "agency-studio-operations",
        "agency-xiaohongshu-specialist",
        "agency-growth-hacker"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: references agency-mobile-app-builder. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-mobile-app-builder\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-xr-immersive-developer\", \"agency-rapid-prototyper\", \"agency-xr-interface-architect\", \"agency-brand-guardian\", \"agency-visionos-spatial-engineer\", \"agency-app-store-optimizer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-terminal-integration-specialist\", \"agency-ux-architect\", \"agency-data-consolidation-agent\", \"agency-senior-developer\", \"agency-macos-spatial-metal-engineer\", \"agency-report-distribution-agent\", \"agency-sales-data-extraction-agent\", \"agency-behavioral-nudge-engine\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0292",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: Theme Factory Skill. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "anthropic-theme-factory"
      ],
      "distractor_skills": [
        "anthropic-brand-guidelines",
        "anthropic-web-artifacts-builder",
        "anthropic-claude-api",
        "anthropic-frontend-design",
        "anthropic-internal-comms",
        "anthropic-slack-gif-creator",
        "anthropic-webapp-testing",
        "anthropic-canvas-design",
        "anthropic-xlsx",
        "anthropic-docx",
        "anthropic-algorithmic-art",
        "anthropic-doc-coauthoring",
        "anthropic-pdf",
        "anthropic-skill-creator",
        "kwp-design-system",
        "anthropic-pptx",
        "kwp-sales-competitive-intelligence",
        "gws-recipe-doc-from-template",
        "kwp-sales-create-an-asset",
        "sp-subagent-dev",
        "anthropic-mcp-builder",
        "kwp-common-room-account-research",
        "kwp-finance-financial-statements",
        "kwp-data-interactive-dashboard-builder",
        "gws-workspace",
        "kwp-common-room-compose-outreach",
        "kwp-sync",
        "gws-sheets",
        "kwp-data-data-context-extractor",
        "kwp-engineering-system-design",
        "kwp-design-design-handoff",
        "kwp-design-user-research",
        "kwp-design-design-system-management",
        "kwp-common-room-weekly-prep-brief",
        "ux-expert",
        "kwp-engineering-documentation",
        "kwp-cowork-create-plugin",
        "deep-research-pipeline",
        "kwp-marketing-brand-voice",
        "kwp-customer-support-response-drafting",
        "notebooklm-studio",
        "ship",
        "kwp-product-management-metrics-tracking"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: Theme Factory Skill. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-theme-factory\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"anthropic-brand-guidelines\", \"anthropic-web-artifacts-builder\", \"anthropic-claude-api\", \"anthropic-frontend-design\", \"anthropic-internal-comms\", \"anthropic-slack-gif-creator\", \"anthropic-webapp-testing\", \"anthropic-canvas-design\", \"anthropic-xlsx\", \"anthropic-docx\", \"anthropic-algorithmic-art\", \"anthropic-doc-coauthoring\", \"anthropic-pdf\", \"anthropic-skill-creator\", \"kwp-design-system\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0293",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: PM 관점. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "role-pm"
      ],
      "distractor_skills": [
        "role-ceo",
        "role-cso",
        "role-finance",
        "role-ux-designer",
        "role-hr",
        "role-developer",
        "role-data-scientist",
        "role-sales",
        "role-security-engineer",
        "role-cto",
        "role-trading-expert",
        "agency-sprint-prioritizer",
        "pm-execution",
        "deep-research-pipeline",
        "role-dispatcher",
        "pm-data-analytics",
        "agency-experiment-tracker",
        "executive-briefing",
        "agency-project-shepherd",
        "qa-test-expert",
        "pm-product-strategy",
        "agency-test-results-analyzer",
        "problem-definition",
        "kwp-product-management-feature-spec",
        "agency-agents-orchestrator",
        "agency-finance-tracker",
        "kwp-product-management-competitive-analysis",
        "agency-data-consolidation-agent",
        "agency-legal-compliance-checker",
        "kwp-human-resources-people-analytics",
        "agency-senior-project-manager",
        "prd-research-factory",
        "e2e-testing",
        "agency-xr-immersive-developer",
        "kwp-operations-change-management",
        "pm-market-research",
        "paper-review",
        "pm-product-discovery",
        "agency-feedback-synthesizer",
        "agency-data-analytics-reporter",
        "agency-analytics-reporter",
        "feature-intake-router",
        "agency-infrastructure-maintainer",
        "critical-review",
        "agency-ux-researcher"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: PM 관점. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"role-pm\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"role-ceo\", \"role-cso\", \"role-finance\", \"role-ux-designer\", \"role-hr\", \"role-developer\", \"role-data-scientist\", \"role-sales\", \"role-security-engineer\", \"role-cto\", \"role-trading-expert\", \"agency-sprint-prioritizer\", \"pm-execution\", \"deep-research-pipeline\", \"role-dispatcher\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0294",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: new lead intake. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "lead-lifecycle-manager"
      ],
      "distractor_skills": [
        "docs-freshness-guardian",
        "feature-intake-router",
        "sales-weekly-dashboard",
        "planning-meeting-sentinel",
        "incident-lifecycle-manager",
        "sales-meeting-pipeline",
        "dependency-radar",
        "prd-research-factory",
        "planning-weekly-pulse",
        "tech-support-brain",
        "proposal-factory",
        "deep-research-pipeline",
        "kwp-sales-call-prep",
        "kwp-sales-draft-outreach",
        "ship",
        "kwp-sales-account-research",
        "auto-research-distribute",
        "paper-review",
        "role-sales",
        "kwp-legal-nda-triage",
        "ui-suite",
        "ecc-search-first",
        "nlm-arxiv-slides",
        "x-to-slack",
        "twitter-timeline-to-slack",
        "decision-router",
        "ecc-strategic-compact",
        "md-enhance-publish"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: new lead intake. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"lead-lifecycle-manager\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"docs-freshness-guardian\", \"feature-intake-router\", \"sales-weekly-dashboard\", \"planning-meeting-sentinel\", \"incident-lifecycle-manager\", \"sales-meeting-pipeline\", \"dependency-radar\", \"prd-research-factory\", \"planning-weekly-pulse\", \"tech-support-brain\", \"proposal-factory\", \"deep-research-pipeline\", \"kwp-sales-call-prep\", \"kwp-sales-draft-outreach\", \"ship\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0295",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: post-mortem. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "incident-to-improvement"
      ],
      "distractor_skills": [
        "incident-lifecycle-manager",
        "kwp-engineering-incident-response",
        "role-cto",
        "test-suite",
        "deep-research-pipeline",
        "infra-runbook-generator",
        "ui-suite",
        "diagnose",
        "release-commander",
        "pr-to-issue-linker",
        "deep-review",
        "e2e-overhaul",
        "refactor-simulator",
        "onboarding-accelerator",
        "autoskill-pipeline",
        "problem-definition",
        "infra-drift-detector",
        "iac-review-agent",
        "release-ship",
        "skill-optimizer",
        "hf-trending-intelligence",
        "helm-validator",
        "bespin-news-digest",
        "issue-to-code",
        "role-security-engineer",
        "qa-test-expert",
        "code-review-all",
        "role-developer",
        "intent-alignment-tracker",
        "backend-expert"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: post-mortem. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"incident-to-improvement\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"incident-lifecycle-manager\", \"kwp-engineering-incident-response\", \"role-cto\", \"test-suite\", \"deep-research-pipeline\", \"infra-runbook-generator\", \"ui-suite\", \"diagnose\", \"release-commander\", \"pr-to-issue-linker\", \"deep-review\", \"e2e-overhaul\", \"refactor-simulator\", \"onboarding-accelerator\", \"autoskill-pipeline\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0296",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Technical Writer agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-technical-writer"
      ],
      "distractor_skills": [
        "agency-xr-immersive-developer",
        "agency-frontend-developer",
        "agency-developer-advocate",
        "agency-brand-guardian",
        "agency-ux-architect",
        "agency-devops-automator",
        "agency-ux-researcher",
        "agency-data-consolidation-agent",
        "agency-feedback-synthesizer",
        "agency-app-store-optimizer",
        "agency-xr-interface-architect",
        "agency-visual-storyteller",
        "agency-inclusive-visuals-specialist",
        "agency-api-tester",
        "agency-security-engineer",
        "agency-image-prompt-engineer",
        "agency-ui-designer",
        "agency-backend-architect",
        "agency-behavioral-nudge-engine",
        "agency-xr-cockpit-interaction-specialist",
        "agency-report-distribution-agent",
        "agency-trend-researcher",
        "agency-rapid-prototyper",
        "agency-mobile-app-builder",
        "agency-terminal-integration-specialist",
        "agency-tool-evaluator",
        "agency-visionos-spatial-engineer",
        "agency-senior-developer",
        "agency-performance-benchmarker",
        "agency-ai-engineer",
        "agency-infrastructure-maintainer",
        "agency-finance-tracker",
        "agency-test-results-analyzer",
        "agency-sales-data-extraction-agent",
        "agency-growth-hacker",
        "agency-workflow-optimizer",
        "agency-cultural-intelligence-strategist",
        "agency-agents-orchestrator",
        "agency-whimsy-injector",
        "agency-wechat-official-account-manager",
        "agency-support-responder",
        "agency-experiment-tracker",
        "agency-tiktok-strategist"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Technical Writer agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-technical-writer\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-xr-immersive-developer\", \"agency-frontend-developer\", \"agency-developer-advocate\", \"agency-brand-guardian\", \"agency-ux-architect\", \"agency-devops-automator\", \"agency-ux-researcher\", \"agency-data-consolidation-agent\", \"agency-feedback-synthesizer\", \"agency-app-store-optimizer\", \"agency-xr-interface-architect\", \"agency-visual-storyteller\", \"agency-inclusive-visuals-specialist\", \"agency-api-tester\", \"agency-security-engineer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0297",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Finance Tracker agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-finance-tracker"
      ],
      "distractor_skills": [
        "agency-brand-guardian",
        "agency-infrastructure-maintainer",
        "agency-growth-hacker",
        "agency-xr-immersive-developer",
        "agency-xr-interface-architect",
        "agency-app-store-optimizer",
        "agency-tool-evaluator",
        "agency-trend-researcher",
        "agency-feedback-synthesizer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-api-tester",
        "agency-terminal-integration-specialist",
        "agency-workflow-optimizer",
        "agency-ux-architect",
        "agency-visionos-spatial-engineer",
        "agency-tiktok-strategist",
        "agency-macos-spatial-metal-engineer",
        "agency-experiment-tracker"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Finance Tracker agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-finance-tracker\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-brand-guardian\", \"agency-infrastructure-maintainer\", \"agency-growth-hacker\", \"agency-xr-immersive-developer\", \"agency-xr-interface-architect\", \"agency-app-store-optimizer\", \"agency-tool-evaluator\", \"agency-trend-researcher\", \"agency-feedback-synthesizer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-api-tester\", \"agency-terminal-integration-specialist\", \"agency-workflow-optimizer\", \"agency-ux-architect\", \"agency-visionos-spatial-engineer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0298",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: managing glossary. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "tab-glossary"
      ],
      "distractor_skills": [
        "tab-report-generate",
        "tab-news-fetch",
        "tab-patterns",
        "tab-market-breadth",
        "tab-market-environment",
        "tab-fundamental-sync",
        "tab-technical-analysis",
        "tab-llm-agents",
        "tab-dashboard-summary",
        "tab-strategy-comparison",
        "tab-kiwoom",
        "tab-screening"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: managing glossary. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"tab-glossary\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"tab-report-generate\", \"tab-news-fetch\", \"tab-patterns\", \"tab-market-breadth\", \"tab-market-environment\", \"tab-fundamental-sync\", \"tab-technical-analysis\", \"tab-llm-agents\", \"tab-dashboard-summary\", \"tab-strategy-comparison\", \"tab-kiwoom\", \"tab-screening\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0299",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: activate the Trend Researcher agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-trend-researcher"
      ],
      "distractor_skills": [
        "agency-brand-guardian",
        "agency-app-store-optimizer",
        "agency-tool-evaluator",
        "agency-xr-immersive-developer",
        "agency-feedback-synthesizer",
        "agency-test-results-analyzer",
        "agency-finance-tracker",
        "agency-xr-interface-architect",
        "agency-ux-researcher",
        "agency-xr-cockpit-interaction-specialist",
        "agency-report-distribution-agent",
        "agency-api-tester",
        "agency-behavioral-nudge-engine",
        "agency-terminal-integration-specialist",
        "agency-workflow-optimizer",
        "agency-cultural-intelligence-strategist",
        "agency-xiaohongshu-specialist",
        "agency-data-consolidation-agent",
        "agency-tiktok-strategist",
        "agency-visionos-spatial-engineer",
        "agency-whimsy-injector",
        "agency-experiment-tracker",
        "agency-senior-project-manager",
        "agency-lsp-index-engineer",
        "agency-reddit-community-builder",
        "agency-instagram-curator",
        "agency-zhihu-strategist",
        "agency-inclusive-visuals-specialist",
        "agency-infrastructure-maintainer",
        "agency-agents-orchestrator",
        "agency-studio-operations",
        "agency-developer-advocate",
        "agency-ai-engineer",
        "agency-visual-storyteller",
        "agency-growth-hacker",
        "agency-ux-architect",
        "agency-rapid-prototyper",
        "agency-mobile-app-builder",
        "agency-project-shepherd",
        "agency-wechat-official-account-manager",
        "agency-senior-developer"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: activate the Trend Researcher agent persona. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-trend-researcher\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-brand-guardian\", \"agency-app-store-optimizer\", \"agency-tool-evaluator\", \"agency-xr-immersive-developer\", \"agency-feedback-synthesizer\", \"agency-test-results-analyzer\", \"agency-finance-tracker\", \"agency-xr-interface-architect\", \"agency-ux-researcher\", \"agency-xr-cockpit-interaction-specialist\", \"agency-report-distribution-agent\", \"agency-api-tester\", \"agency-behavioral-nudge-engine\", \"agency-terminal-integration-specialist\", \"agency-workflow-optimizer\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "DS-0300",
      "setting": "distractor",
      "query": "Several skills look interchangeable for: references agency-terminal-integration-specialist. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).",
      "gold_skills": [
        "agency-terminal-integration-specialist"
      ],
      "distractor_skills": [
        "agency-app-store-optimizer",
        "agency-xr-immersive-developer",
        "agency-xr-cockpit-interaction-specialist",
        "agency-macos-spatial-metal-engineer",
        "agency-xr-interface-architect",
        "agency-senior-developer",
        "agency-tool-evaluator",
        "agency-brand-guardian",
        "agency-behavioral-nudge-engine",
        "agency-visionos-spatial-engineer",
        "agency-ux-architect",
        "agency-tiktok-strategist"
      ],
      "budget_tokens": null,
      "difficulty": "hard",
      "source": "synthetic",
      "judge_verdict": "uncertain",
      "judge_confidence": 0.68,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nSeveral skills look interchangeable for: references agency-terminal-integration-specialist. Select ONLY the canonical skill id for this repo’s Cursor skill library (not a generic agency persona if a project-specific skill exists).\n\nProposed gold_skills (ordered if composition):\n[\"agency-terminal-integration-specialist\"]\n\nSetting: distractor\nDistractors (if any, first 15 shown):\n[\"agency-app-store-optimizer\", \"agency-xr-immersive-developer\", \"agency-xr-cockpit-interaction-specialist\", \"agency-macos-spatial-metal-engineer\", \"agency-xr-interface-architect\", \"agency-senior-developer\", \"agency-tool-evaluator\", \"agency-brand-guardian\", \"agency-behavioral-nudge-engine\", \"agency-visionos-spatial-engineer\", \"agency-ux-architect\", \"agency-tiktok-strategist\"]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0001",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Design Critique. Ignore any legacy alias like `kwp-design-design-critique__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-design-design-critique"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-design-design-critique__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Design Critique. Ignore any legacy alias like `kwp-design-design-critique__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-design-design-critique\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0002",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: validate K8s manifests. Ignore any legacy alias like `k8s-manifest-validator__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "k8s-manifest-validator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "k8s-manifest-validator__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: validate K8s manifests. Ignore any legacy alias like `k8s-manifest-validator__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"k8s-manifest-validator\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0003",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: UI 변경 요청 시. Ignore any legacy alias like `fe-pipeline__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "fe-pipeline"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "fe-pipeline__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: UI 변경 요청 시. Ignore any legacy alias like `fe-pipeline__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"fe-pipeline\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0004",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: users request QC analysis. Ignore any legacy alias like `kwp-bio-research-single-cell-rna-qc__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-bio-research-single-cell-rna-qc"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-bio-research-single-cell-rna-qc__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: users request QC analysis. Ignore any legacy alias like `kwp-bio-research-single-cell-rna-qc__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-bio-research-single-cell-rna-qc\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0005",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Behavioral Nudge Engine agent persona. Ignore any legacy alias like `agency-behavioral-nudge-engine__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-behavioral-nudge-engine"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-behavioral-nudge-engine__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Behavioral Nudge Engine agent persona. Ignore any legacy alias like `agency-behavioral-nudge-engine__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-behavioral-nudge-engine\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0006",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: run autoskill pipeline. Ignore any legacy alias like `autoskill-pipeline__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "autoskill-pipeline"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "autoskill-pipeline__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: run autoskill pipeline. Ignore any legacy alias like `autoskill-pipeline__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"autoskill-pipeline\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0007",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: writing weekly status updates. Ignore any legacy alias like `kwp-product-management-stakeholder-comms__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-product-management-stakeholder-comms"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-product-management-stakeholder-comms__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: writing weekly status updates. Ignore any legacy alias like `kwp-product-management-stakeholder-comms__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-product-management-stakeholder-comms\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0008",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: create an automation. Ignore any legacy alias like `cursor-automations__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "cursor-automations"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "cursor-automations__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: create an automation. Ignore any legacy alias like `cursor-automations__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"cursor-automations\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0009",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: transcribe a video URL. Ignore any legacy alias like `transcribee__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "transcribee"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "transcribee__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: transcribe a video URL. Ignore any legacy alias like `transcribee__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"transcribee\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0010",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Mobile App Builder agent persona. Ignore any legacy alias like `agency-mobile-app-builder__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-mobile-app-builder"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-mobile-app-builder__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Mobile App Builder agent persona. Ignore any legacy alias like `agency-mobile-app-builder__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-mobile-app-builder\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0011",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: latency problems. Ignore any legacy alias like `performance-profiler__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "performance-profiler"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "performance-profiler__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: latency problems. Ignore any legacy alias like `performance-profiler__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"performance-profiler\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0012",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: incident lifecycle. Ignore any legacy alias like `incident-lifecycle-manager__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "incident-lifecycle-manager"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "incident-lifecycle-manager__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: incident lifecycle. Ignore any legacy alias like `incident-lifecycle-manager__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"incident-lifecycle-manager\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0013",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: refreshing Turtle indicators. Ignore any legacy alias like `tab-turtle-refresh__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "tab-turtle-refresh"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "tab-turtle-refresh__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: refreshing Turtle indicators. Ignore any legacy alias like `tab-turtle-refresh__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"tab-turtle-refresh\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0014",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: run sector rotation analysis. Ignore any legacy alias like `trading-sector-analyst__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "trading-sector-analyst"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "trading-sector-analyst__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: run sector rotation analysis. Ignore any legacy alias like `trading-sector-analyst__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"trading-sector-analyst\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0015",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: triage issues. Ignore any legacy alias like `sprint-orchestrator__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "sprint-orchestrator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "sprint-orchestrator__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: triage issues. Ignore any legacy alias like `sprint-orchestrator__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"sprint-orchestrator\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0016",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Social Media Strategist agent persona. Ignore any legacy alias like `agency-social-media-strategist__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-social-media-strategist"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-social-media-strategist__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Social Media Strategist agent persona. Ignore any legacy alias like `agency-social-media-strategist__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-social-media-strategist\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0017",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: who owns this code. Ignore any legacy alias like `codebase-archaeologist__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "codebase-archaeologist"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "codebase-archaeologist__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: who owns this code. Ignore any legacy alias like `codebase-archaeologist__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"codebase-archaeologist\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0018",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: responding to customer tickets. Ignore any legacy alias like `kwp-customer-support-response-drafting__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-customer-support-response-drafting"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-customer-support-response-drafting__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: responding to customer tickets. Ignore any legacy alias like `kwp-customer-support-response-drafting__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-customer-support-response-drafting\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0019",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Agentic Identity Trust Architect agent persona. Ignore any legacy alias like `agency-agentic-identity-trust-architect__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-agentic-identity-trust-architect"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-agentic-identity-trust-architect__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Agentic Identity Trust Architect agent persona. Ignore any legacy alias like `agency-agentic-identity-trust-architect__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-agentic-identity-trust-architect\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0020",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: generating testing workpapers. Ignore any legacy alias like `kwp-finance-audit-support__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-finance-audit-support"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-finance-audit-support__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: generating testing workpapers. Ignore any legacy alias like `kwp-finance-audit-support__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-finance-audit-support\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0021",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: analyzing interview notes. Ignore any legacy alias like `kwp-product-management-user-research-synthesis__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-product-management-user-research-synthesis"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-product-management-user-research-synthesis__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: analyzing interview notes. Ignore any legacy alias like `kwp-product-management-user-research-synthesis__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-product-management-user-research-synthesis\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0022",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: starting feature work that needs isolation from current workspace. Ignore any legacy alias like `sp-git-worktrees__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "sp-git-worktrees"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "sp-git-worktrees__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: starting feature work that needs isolation from current workspace. Ignore any legacy alias like `sp-git-worktrees__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"sp-git-worktrees\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0023",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: requests professional documents with tables of contents. Ignore any legacy alias like `anthropic-docx__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "anthropic-docx"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "anthropic-docx__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: requests professional documents with tables of contents. Ignore any legacy alias like `anthropic-docx__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-docx\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0024",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: plan automation. Ignore any legacy alias like `automation-strategist__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "automation-strategist"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "automation-strategist__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: plan automation. Ignore any legacy alias like `automation-strategist__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"automation-strategist\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0025",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Performance Benchmarker agent persona. Ignore any legacy alias like `agency-performance-benchmarker__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-performance-benchmarker"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-performance-benchmarker__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Performance Benchmarker agent persona. Ignore any legacy alias like `agency-performance-benchmarker__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-performance-benchmarker\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0026",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: creating a roadmap. Ignore any legacy alias like `kwp-product-management-roadmap-management__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-product-management-roadmap-management"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-product-management-roadmap-management__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: creating a roadmap. Ignore any legacy alias like `kwp-product-management-roadmap-management__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-product-management-roadmap-management\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0027",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: recent price changes. Ignore any legacy alias like `alphaear-stock__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "alphaear-stock"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "alphaear-stock__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: recent price changes. Ignore any legacy alias like `alphaear-stock__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-stock\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0028",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: generating GenAI features. Ignore any legacy alias like `tab-genai-features__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "tab-genai-features"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "tab-genai-features__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: generating GenAI features. Ignore any legacy alias like `tab-genai-features__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"tab-genai-features\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0029",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: : customize plugin. Ignore any legacy alias like `kwp-cowork-plugin-customizer__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-cowork-plugin-customizer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-cowork-plugin-customizer__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: : customize plugin. Ignore any legacy alias like `kwp-cowork-plugin-customizer__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-cowork-plugin-customizer\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0030",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: sync with Jira. Ignore any legacy alias like `jira-linear-adapter__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "jira-linear-adapter"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "jira-linear-adapter__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: sync with Jira. Ignore any legacy alias like `jira-linear-adapter__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"jira-linear-adapter\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0031",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Google Daily Automation. Ignore any legacy alias like `google-daily__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "google-daily"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "google-daily__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Google Daily Automation. Ignore any legacy alias like `google-daily__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"google-daily\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0032",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Tech Debt Management. Ignore any legacy alias like `kwp-engineering-tech-debt__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-engineering-tech-debt"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-engineering-tech-debt__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Tech Debt Management. Ignore any legacy alias like `kwp-engineering-tech-debt__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-engineering-tech-debt\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0033",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: monitor papers. Ignore any legacy alias like `paper-auto-classifier__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "paper-auto-classifier"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "paper-auto-classifier__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: monitor papers. Ignore any legacy alias like `paper-auto-classifier__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"paper-auto-classifier\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0034",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: process this link. Ignore any legacy alias like `unified-intel-intake__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "unified-intel-intake"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "unified-intel-intake__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: process this link. Ignore any legacy alias like `unified-intel-intake__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"unified-intel-intake\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0035",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: produce one. Ignore any legacy alias like `anthropic-pdf__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "anthropic-pdf"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "anthropic-pdf__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: produce one. Ignore any legacy alias like `anthropic-pdf__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-pdf\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0036",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: create E2E tests. Ignore any legacy alias like `e2e-testing__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "e2e-testing"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "e2e-testing__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: create E2E tests. Ignore any legacy alias like `e2e-testing__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"e2e-testing\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0037",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Design System Management. Ignore any legacy alias like `kwp-design-design-system-management__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-design-design-system-management"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-design-design-system-management__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Design System Management. Ignore any legacy alias like `kwp-design-design-system-management__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-design-design-system-management\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0038",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: wrap up for the day. Ignore any legacy alias like `eod-ship__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "eod-ship"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "eod-ship__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: wrap up for the day. Ignore any legacy alias like `eod-ship__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"eod-ship\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0039",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Brand Discovery. Ignore any legacy alias like `kwp-brand-voice-discover-brand__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-brand-voice-discover-brand"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-brand-voice-discover-brand__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Brand Discovery. Ignore any legacy alias like `kwp-brand-voice-discover-brand__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-brand-voice-discover-brand\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0040",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: upload files. Ignore any legacy alias like `gws-drive__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "gws-drive"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "gws-drive__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: upload files. Ignore any legacy alias like `gws-drive__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"gws-drive\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0041",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: subagents lack sufficient context. Ignore any legacy alias like `ecc-iterative-retrieval__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "ecc-iterative-retrieval"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "ecc-iterative-retrieval__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: subagents lack sufficient context. Ignore any legacy alias like `ecc-iterative-retrieval__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"ecc-iterative-retrieval\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0042",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Technical Documentation. Ignore any legacy alias like `kwp-engineering-documentation__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-engineering-documentation"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-engineering-documentation__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Technical Documentation. Ignore any legacy alias like `kwp-engineering-documentation__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-engineering-documentation\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0043",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: screen stocks. Ignore any legacy alias like `trading-finviz-screener__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "trading-finviz-screener"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "trading-finviz-screener__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: screen stocks. Ignore any legacy alias like `trading-finviz-screener__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"trading-finviz-screener\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0044",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: sync skills. Ignore any legacy alias like `cursor-sync__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "cursor-sync"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "cursor-sync__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: sync skills. Ignore any legacy alias like `cursor-sync__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"cursor-sync\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0045",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: evaluating fine-tuned financial sentiment models. Ignore any legacy alias like `hf-evaluation__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "hf-evaluation"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "hf-evaluation__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: evaluating fine-tuned financial sentiment models. Ignore any legacy alias like `hf-evaluation__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"hf-evaluation\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0046",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: refreshing breadth. Ignore any legacy alias like `tab-market-breadth__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "tab-market-breadth"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "tab-market-breadth__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: refreshing breadth. Ignore any legacy alias like `tab-market-breadth__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"tab-market-breadth\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0047",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: commit my changes. Ignore any legacy alias like `domain-commit__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "domain-commit"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "domain-commit__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: commit my changes. Ignore any legacy alias like `domain-commit__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"domain-commit\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0048",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: CEO perspective. Ignore any legacy alias like `role-ceo__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "role-ceo"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "role-ceo__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: CEO perspective. Ignore any legacy alias like `role-ceo__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"role-ceo\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0049",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: create a podcast. Ignore any legacy alias like `notebooklm-studio__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "notebooklm-studio"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "notebooklm-studio__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: create a podcast. Ignore any legacy alias like `notebooklm-studio__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"notebooklm-studio\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0050",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: discovering hot stocks. Ignore any legacy alias like `tab-hot-stock-discovery__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "tab-hot-stock-discovery"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "tab-hot-stock-discovery__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: discovering hot stocks. Ignore any legacy alias like `tab-hot-stock-discovery__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"tab-hot-stock-discovery\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0051",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: run ralph loop. Ignore any legacy alias like `ralph-loop__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "ralph-loop"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "ralph-loop__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: run ralph loop. Ignore any legacy alias like `ralph-loop__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"ralph-loop\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0052",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Data Consolidation Agent agent persona. Ignore any legacy alias like `agency-data-consolidation-agent__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-data-consolidation-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-data-consolidation-agent__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Data Consolidation Agent agent persona. Ignore any legacy alias like `agency-data-consolidation-agent__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-data-consolidation-agent\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0053",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Incident Response. Ignore any legacy alias like `kwp-engineering-incident-response__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-engineering-incident-response"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-engineering-incident-response__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Incident Response. Ignore any legacy alias like `kwp-engineering-incident-response__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-engineering-incident-response\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0054",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: extract skills from sessions. Ignore any legacy alias like `autoskill-extractor__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "autoskill-extractor"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "autoskill-extractor__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: extract skills from sessions. Ignore any legacy alias like `autoskill-extractor__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"autoskill-extractor\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0055",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: CSO perspective. Ignore any legacy alias like `role-cso__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "role-cso"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "role-cso__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: CSO perspective. Ignore any legacy alias like `role-cso__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"role-cso\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0056",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Lsp Index Engineer agent persona. Ignore any legacy alias like `agency-lsp-index-engineer__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-lsp-index-engineer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-lsp-index-engineer__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Lsp Index Engineer agent persona. Ignore any legacy alias like `agency-lsp-index-engineer__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-lsp-index-engineer\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0057",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: is the market topping. Ignore any legacy alias like `trading-market-top-detector__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "trading-market-top-detector"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "trading-market-top-detector__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: is the market topping. Ignore any legacy alias like `trading-market-top-detector__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"trading-market-top-detector\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0058",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: write PRD. Ignore any legacy alias like `pm-execution__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "pm-execution"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "pm-execution__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: write PRD. Ignore any legacy alias like `pm-execution__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"pm-execution\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0059",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: check paperclip status. Ignore any legacy alias like `paperclip-control__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "paperclip-control"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "paperclip-control__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: check paperclip status. Ignore any legacy alias like `paperclip-control__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"paperclip-control\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0060",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: find models. Ignore any legacy alias like `hf-models__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "hf-models"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "hf-models__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: find models. Ignore any legacy alias like `hf-models__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"hf-models\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0061",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: update KWP rules. Ignore any legacy alias like `kwp-sync__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-sync"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-sync__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: update KWP rules. Ignore any legacy alias like `kwp-sync__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-sync\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0062",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Security Engineer agent persona. Ignore any legacy alias like `agency-security-engineer__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-security-engineer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-security-engineer__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Security Engineer agent persona. Ignore any legacy alias like `agency-security-engineer__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-security-engineer\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0063",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: deck. Ignore any legacy alias like `anthropic-pptx__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "anthropic-pptx"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "anthropic-pptx__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: deck. Ignore any legacy alias like `anthropic-pptx__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-pptx\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0064",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: designing long-running AFK agent workflows. Ignore any legacy alias like `ecc-autonomous-loops__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "ecc-autonomous-loops"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "ecc-autonomous-loops__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: designing long-running AFK agent workflows. Ignore any legacy alias like `ecc-autonomous-loops__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"ecc-autonomous-loops\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0065",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Xiaohongshu Specialist agent persona. Ignore any legacy alias like `agency-xiaohongshu-specialist__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-xiaohongshu-specialist"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-xiaohongshu-specialist__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Xiaohongshu Specialist agent persona. Ignore any legacy alias like `agency-xiaohongshu-specialist__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-xiaohongshu-specialist\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0066",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: evolve skills. Ignore any legacy alias like `autoskill-evolve__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "autoskill-evolve"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "autoskill-evolve__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: evolve skills. Ignore any legacy alias like `autoskill-evolve__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"autoskill-evolve\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0067",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: deep research pipeline. Ignore any legacy alias like `deep-research-pipeline__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "deep-research-pipeline"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "deep-research-pipeline__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: deep research pipeline. Ignore any legacy alias like `deep-research-pipeline__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"deep-research-pipeline\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0068",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: check for drift. Ignore any legacy alias like `infra-drift-detector__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "infra-drift-detector"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "infra-drift-detector__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: check for drift. Ignore any legacy alias like `infra-drift-detector__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"infra-drift-detector\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0069",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: refreshing DualMA. Ignore any legacy alias like `tab-dualma-refresh__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "tab-dualma-refresh"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "tab-dualma-refresh__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: refreshing DualMA. Ignore any legacy alias like `tab-dualma-refresh__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"tab-dualma-refresh\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0070",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: researching a competitor. Ignore any legacy alias like `kwp-product-management-competitive-analysis__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-product-management-competitive-analysis"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-product-management-competitive-analysis__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: researching a competitor. Ignore any legacy alias like `kwp-product-management-competitive-analysis__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-product-management-competitive-analysis\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0071",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: onboarding. Ignore any legacy alias like `onboarding-accelerator__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "onboarding-accelerator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "onboarding-accelerator__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: onboarding. Ignore any legacy alias like `onboarding-accelerator__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"onboarding-accelerator\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0072",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Contact Research. Ignore any legacy alias like `kwp-common-room-contact-research__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-common-room-contact-research"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-common-room-contact-research__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Contact Research. Ignore any legacy alias like `kwp-common-room-contact-research__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-common-room-contact-research\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0073",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: test review. Ignore any legacy alias like `test-suite__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "test-suite"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "test-suite__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: test review. Ignore any legacy alias like `test-suite__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"test-suite\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0074",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: link PRs to issues. Ignore any legacy alias like `pr-to-issue-linker__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "pr-to-issue-linker"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "pr-to-issue-linker__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: link PRs to issues. Ignore any legacy alias like `pr-to-issue-linker__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"pr-to-issue-linker\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0075",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Project Shepherd agent persona. Ignore any legacy alias like `agency-project-shepherd__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-project-shepherd"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-project-shepherd__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Project Shepherd agent persona. Ignore any legacy alias like `agency-project-shepherd__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-project-shepherd\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0076",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: merge skill candidate. Ignore any legacy alias like `autoskill-merger__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "autoskill-merger"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "autoskill-merger__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: merge skill candidate. Ignore any legacy alias like `autoskill-merger__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"autoskill-merger\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0077",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Swagger API Documentation Generator. Ignore any legacy alias like `swagger-api-doc-generator__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "swagger-api-doc-generator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "swagger-api-doc-generator__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Swagger API Documentation Generator. Ignore any legacy alias like `swagger-api-doc-generator__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"swagger-api-doc-generator\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0078",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: performing bank reconciliations. Ignore any legacy alias like `kwp-finance-reconciliation__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-finance-reconciliation"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-finance-reconciliation__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: performing bank reconciliations. Ignore any legacy alias like `kwp-finance-reconciliation__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-finance-reconciliation\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0079",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: check data quality. Ignore any legacy alias like `trading-data-quality-checker__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "trading-data-quality-checker"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "trading-data-quality-checker__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: check data quality. Ignore any legacy alias like `trading-data-quality-checker__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"trading-data-quality-checker\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0080",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Growth Hacker agent persona. Ignore any legacy alias like `agency-growth-hacker__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-growth-hacker"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-growth-hacker__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Growth Hacker agent persona. Ignore any legacy alias like `agency-growth-hacker__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-growth-hacker\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0081",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: create strategy hub. Ignore any legacy alias like `nlm-strategy-hub__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "nlm-strategy-hub"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "nlm-strategy-hub__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: create strategy hub. Ignore any legacy alias like `nlm-strategy-hub__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"nlm-strategy-hub\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0082",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: list today's papers. Ignore any legacy alias like `hf-papers__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "hf-papers"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "hf-papers__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: list today's papers. Ignore any legacy alias like `hf-papers__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"hf-papers\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0083",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: monitoring finance signals. Ignore any legacy alias like `alphaear-signal-tracker__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "alphaear-signal-tracker"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "alphaear-signal-tracker__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: monitoring finance signals. Ignore any legacy alias like `alphaear-signal-tracker__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-signal-tracker\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0084",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Ui Designer agent persona. Ignore any legacy alias like `agency-ui-designer__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-ui-designer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-ui-designer__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Ui Designer agent persona. Ignore any legacy alias like `agency-ui-designer__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-ui-designer\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0085",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Prospect. Ignore any legacy alias like `kwp-apollo-prospect__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-apollo-prospect"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-apollo-prospect__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Prospect. Ignore any legacy alias like `kwp-apollo-prospect__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-apollo-prospect\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0086",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: update recent stock data. Ignore any legacy alias like `weekly-stock-update__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "weekly-stock-update"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "weekly-stock-update__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: update recent stock data. Ignore any legacy alias like `weekly-stock-update__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"weekly-stock-update\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0087",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: convert knowledge to paper. Ignore any legacy alias like `knowledge-to-paper__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "knowledge-to-paper"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "knowledge-to-paper__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: convert knowledge to paper. Ignore any legacy alias like `knowledge-to-paper__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"knowledge-to-paper\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0088",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Task Management. Ignore any legacy alias like `kwp-productivity-task-management__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-productivity-task-management"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-productivity-task-management__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Task Management. Ignore any legacy alias like `kwp-productivity-task-management__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-productivity-task-management\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0089",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: data scientist perspective. Ignore any legacy alias like `role-data-scientist__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "role-data-scientist"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "role-data-scientist__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: data scientist perspective. Ignore any legacy alias like `role-data-scientist__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"role-data-scientist\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0090",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Change Management. Ignore any legacy alias like `kwp-operations-change-management__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-operations-change-management"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-operations-change-management__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Change Management. Ignore any legacy alias like `kwp-operations-change-management__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-operations-change-management\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0091",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: create demo. Ignore any legacy alias like `demo-forge__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "demo-forge"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "demo-forge__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: create demo. Ignore any legacy alias like `demo-forge__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"demo-forge\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0092",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Frontend Developer agent persona. Ignore any legacy alias like `agency-frontend-developer__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-frontend-developer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-frontend-developer__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Frontend Developer agent persona. Ignore any legacy alias like `agency-frontend-developer__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-frontend-developer\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0093",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: generate standup. Ignore any legacy alias like `standup-digest__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "standup-digest"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "standup-digest__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: generate standup. Ignore any legacy alias like `standup-digest__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"standup-digest\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0094",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: check setup. Ignore any legacy alias like `setup-doctor__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "setup-doctor"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "setup-doctor__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: check setup. Ignore any legacy alias like `setup-doctor__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"setup-doctor\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0095",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Zhihu Strategist agent persona. Ignore any legacy alias like `agency-zhihu-strategist__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-zhihu-strategist"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-zhihu-strategist__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Zhihu Strategist agent persona. Ignore any legacy alias like `agency-zhihu-strategist__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-zhihu-strategist\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0096",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Source Management. Ignore any legacy alias like `kwp-enterprise-search-source-management__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-enterprise-search-source-management"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-enterprise-search-source-management__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Source Management. Ignore any legacy alias like `kwp-enterprise-search-source-management__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-enterprise-search-source-management\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0097",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: recall. Ignore any legacy alias like `recall__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "recall"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "recall__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: recall. Ignore any legacy alias like `recall__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"recall\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0098",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: archive an email to Docs. Ignore any legacy alias like `gws-recipe-save-email-to-doc__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "gws-recipe-save-email-to-doc"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "gws-recipe-save-email-to-doc__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: archive an email to Docs. Ignore any legacy alias like `gws-recipe-save-email-to-doc__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"gws-recipe-save-email-to-doc\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0099",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: computing TA. Ignore any legacy alias like `tab-technical-analysis__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "tab-technical-analysis"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "tab-technical-analysis__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: computing TA. Ignore any legacy alias like `tab-technical-analysis__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"tab-technical-analysis\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0100",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: design a system. Ignore any legacy alias like `system-thinker__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "system-thinker"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "system-thinker__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: design a system. Ignore any legacy alias like `system-thinker__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"system-thinker\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0101",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: schedule feedback meetings. Ignore any legacy alias like `feedback-meeting-scheduler__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "feedback-meeting-scheduler"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "feedback-meeting-scheduler__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: schedule feedback meetings. Ignore any legacy alias like `feedback-meeting-scheduler__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"feedback-meeting-scheduler\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0102",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Evidence Collector agent persona. Ignore any legacy alias like `agency-evidence-collector__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-evidence-collector"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-evidence-collector__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Evidence Collector agent persona. Ignore any legacy alias like `agency-evidence-collector__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-evidence-collector\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0103",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: creating dashboards. Ignore any legacy alias like `kwp-data-interactive-dashboard-builder__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-data-interactive-dashboard-builder"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-data-interactive-dashboard-builder__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: creating dashboards. Ignore any legacy alias like `kwp-data-interactive-dashboard-builder__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-data-interactive-dashboard-builder\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0104",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Twitter Engager agent persona. Ignore any legacy alias like `agency-twitter-engager__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-twitter-engager"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-twitter-engager__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Twitter Engager agent persona. Ignore any legacy alias like `agency-twitter-engager__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-twitter-engager\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0105",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: meeting sentinel. Ignore any legacy alias like `planning-meeting-sentinel__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "planning-meeting-sentinel"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "planning-meeting-sentinel__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: meeting sentinel. Ignore any legacy alias like `planning-meeting-sentinel__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"planning-meeting-sentinel\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0106",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: create NLM slides. Ignore any legacy alias like `nlm-slides__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "nlm-slides"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "nlm-slides__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: create NLM slides. Ignore any legacy alias like `nlm-slides__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"nlm-slides\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0107",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: PRD from meeting. Ignore any legacy alias like `prd-research-factory__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "prd-research-factory"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "prd-research-factory__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: PRD from meeting. Ignore any legacy alias like `prd-research-factory__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"prd-research-factory\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0108",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: evaluating agent output quality. Ignore any legacy alias like `ecc-eval-harness__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "ecc-eval-harness"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "ecc-eval-harness__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: evaluating agent output quality. Ignore any legacy alias like `ecc-eval-harness__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"ecc-eval-harness\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0109",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: find an API. Ignore any legacy alias like `public-apis__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "public-apis"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "public-apis__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: find an API. Ignore any legacy alias like `public-apis__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"public-apis\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0110",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Interview Prep. Ignore any legacy alias like `kwp-human-resources-interview-prep__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-human-resources-interview-prep"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-human-resources-interview-prep__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Interview Prep. Ignore any legacy alias like `kwp-human-resources-interview-prep__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-human-resources-interview-prep\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0111",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: fetching news. Ignore any legacy alias like `tab-news-fetch__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "tab-news-fetch"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "tab-news-fetch__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: fetching news. Ignore any legacy alias like `tab-news-fetch__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"tab-news-fetch\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0112",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: US Market Bubble Detection Skill (Revised v2.1). Ignore any legacy alias like `trading-us-market-bubble-detector__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "trading-us-market-bubble-detector"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "trading-us-market-bubble-detector__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: US Market Bubble Detection Skill (Revised v2.1). Ignore any legacy alias like `trading-us-market-bubble-detector__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"trading-us-market-bubble-detector\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0113",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: distribute research results. Ignore any legacy alias like `auto-research-distribute__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "auto-research-distribute"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "auto-research-distribute__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: distribute research results. Ignore any legacy alias like `auto-research-distribute__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"auto-research-distribute\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0114",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: market breadth assessment. Ignore any legacy alias like `trading-breadth-chart-analyst__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "trading-breadth-chart-analyst"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "trading-breadth-chart-analyst__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: market breadth assessment. Ignore any legacy alias like `trading-breadth-chart-analyst__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"trading-breadth-chart-analyst\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0115",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: scvi-tools Deep Learning Skill. Ignore any legacy alias like `kwp-bio-research-scvi-tools__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-bio-research-scvi-tools"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-bio-research-scvi-tools__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: scvi-tools Deep Learning Skill. Ignore any legacy alias like `kwp-bio-research-scvi-tools__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-bio-research-scvi-tools\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0116",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Screen Implementation Orchestrator. Ignore any legacy alias like `implement-screen__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "implement-screen"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "implement-screen__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Screen Implementation Orchestrator. Ignore any legacy alias like `implement-screen__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"implement-screen\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0117",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Show companies hiring AI engineers this month. Ignore any legacy alias like `kwp-common-room-prospect__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-common-room-prospect"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-common-room-prospect__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Show companies hiring AI engineers this month. Ignore any legacy alias like `kwp-common-room-prospect__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-common-room-prospect\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0118",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: explain this paper. Ignore any legacy alias like `alphaxiv-paper-lookup__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "alphaxiv-paper-lookup"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "alphaxiv-paper-lookup__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: explain this paper. Ignore any legacy alias like `alphaxiv-paper-lookup__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"alphaxiv-paper-lookup\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0119",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Inclusive Visuals Specialist agent persona. Ignore any legacy alias like `agency-inclusive-visuals-specialist__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-inclusive-visuals-specialist"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-inclusive-visuals-specialist__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Inclusive Visuals Specialist agent persona. Ignore any legacy alias like `agency-inclusive-visuals-specialist__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-inclusive-visuals-specialist\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0120",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Developer Advocate agent persona. Ignore any legacy alias like `agency-developer-advocate__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-developer-advocate"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-developer-advocate__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Developer Advocate agent persona. Ignore any legacy alias like `agency-developer-advocate__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-developer-advocate\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0121",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: completing tasks. Ignore any legacy alias like `sp-request-review__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "sp-request-review"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "sp-request-review__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: completing tasks. Ignore any legacy alias like `sp-request-review__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"sp-request-review\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0122",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: docs freshness. Ignore any legacy alias like `docs-freshness-guardian__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "docs-freshness-guardian"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "docs-freshness-guardian__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: docs freshness. Ignore any legacy alias like `docs-freshness-guardian__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"docs-freshness-guardian\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0123",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Resource Planning. Ignore any legacy alias like `kwp-operations-resource-planning__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-operations-resource-planning"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-operations-resource-planning__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Resource Planning. Ignore any legacy alias like `kwp-operations-resource-planning__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-operations-resource-planning\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0124",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: enforcing consistent style. Ignore any legacy alias like `ecc-coding-standards__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "ecc-coding-standards"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "ecc-coding-standards__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: enforcing consistent style. Ignore any legacy alias like `ecc-coding-standards__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"ecc-coding-standards\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0125",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: simulate public reaction. Ignore any legacy alias like `mirofish-opinion-sim__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "mirofish-opinion-sim"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "mirofish-opinion-sim__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: simulate public reaction. Ignore any legacy alias like `mirofish-opinion-sim__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"mirofish-opinion-sim\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0126",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: finance perspective. Ignore any legacy alias like `role-finance__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "role-finance"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "role-finance__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: finance perspective. Ignore any legacy alias like `role-finance__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"role-finance\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0127",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: GTM strategy. Ignore any legacy alias like `pm-go-to-market__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "pm-go-to-market"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "pm-go-to-market__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: GTM strategy. Ignore any legacy alias like `pm-go-to-market__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"pm-go-to-market\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0128",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: define the problem. Ignore any legacy alias like `problem-definition__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "problem-definition"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "problem-definition__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: define the problem. Ignore any legacy alias like `problem-definition__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"problem-definition\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0129",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: sync skills across instances. Ignore any legacy alias like `sefo-federation__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "sefo-federation"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "sefo-federation__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: sync skills across instances. Ignore any legacy alias like `sefo-federation__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"sefo-federation\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0130",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: a security review. Ignore any legacy alias like `security-expert__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "security-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "security-expert__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: a security review. Ignore any legacy alias like `security-expert__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"security-expert\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0131",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: market breadth healthy. Ignore any legacy alias like `trading-uptrend-analyzer__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "trading-uptrend-analyzer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "trading-uptrend-analyzer__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: market breadth healthy. Ignore any legacy alias like `trading-uptrend-analyzer__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"trading-uptrend-analyzer\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0132",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Ux Researcher agent persona. Ignore any legacy alias like `agency-ux-researcher__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-ux-researcher"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-ux-researcher__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Ux Researcher agent persona. Ignore any legacy alias like `agency-ux-researcher__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-ux-researcher\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0133",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: create a Google Doc. Ignore any legacy alias like `gws-docs__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "gws-docs"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "gws-docs__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: create a Google Doc. Ignore any legacy alias like `gws-docs__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"gws-docs\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0134",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: fill a PPT template. Ignore any legacy alias like `ppt-template-engine__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "ppt-template-engine"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "ppt-template-engine__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: fill a PPT template. Ignore any legacy alias like `ppt-template-engine__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"ppt-template-engine\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0135",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Rapid Prototyper agent persona. Ignore any legacy alias like `agency-rapid-prototyper__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-rapid-prototyper"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-rapid-prototyper__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Rapid Prototyper agent persona. Ignore any legacy alias like `agency-rapid-prototyper__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-rapid-prototyper\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0136",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: create NLM video. Ignore any legacy alias like `nlm-video__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "nlm-video"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "nlm-video__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: create NLM video. Ignore any legacy alias like `nlm-video__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"nlm-video\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0137",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Call Prep. Ignore any legacy alias like `kwp-common-room-call-prep__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-common-room-call-prep"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-common-room-call-prep__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Call Prep. Ignore any legacy alias like `kwp-common-room-call-prep__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-common-room-call-prep\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0138",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Report Distribution Agent agent persona. Ignore any legacy alias like `agency-report-distribution-agent__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-report-distribution-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-report-distribution-agent__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Report Distribution Agent agent persona. Ignore any legacy alias like `agency-report-distribution-agent__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-report-distribution-agent\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0139",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: run the project locally. Ignore any legacy alias like `local-dev-runner__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "local-dev-runner"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "local-dev-runner__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: run the project locally. Ignore any legacy alias like `local-dev-runner__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"local-dev-runner\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0140",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: check report quality. Ignore any legacy alias like `ai-quality-evaluator__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "ai-quality-evaluator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "ai-quality-evaluator__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: check report quality. Ignore any legacy alias like `ai-quality-evaluator__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"ai-quality-evaluator\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0141",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Autonomous Optimization Architect agent persona. Ignore any legacy alias like `agency-autonomous-optimization-architect__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-autonomous-optimization-architect"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-autonomous-optimization-architect__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Autonomous Optimization Architect agent persona. Ignore any legacy alias like `agency-autonomous-optimization-architect__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-autonomous-optimization-architect\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0142",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: validate Helm chart. Ignore any legacy alias like `helm-validator__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "helm-validator"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "helm-validator__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: validate Helm chart. Ignore any legacy alias like `helm-validator__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"helm-validator\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0143",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: fine-tuning financial sentiment models. Ignore any legacy alias like `hf-model-trainer__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "hf-model-trainer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "hf-model-trainer__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: fine-tuning financial sentiment models. Ignore any legacy alias like `hf-model-trainer__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"hf-model-trainer\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0144",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: running GPU workloads for financial model inference. Ignore any legacy alias like `hf-jobs__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "hf-jobs"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "hf-jobs__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: running GPU workloads for financial model inference. Ignore any legacy alias like `hf-jobs__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"hf-jobs\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0145",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: SQL query. Ignore any legacy alias like `pm-data-analytics__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "pm-data-analytics"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "pm-data-analytics__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: SQL query. Ignore any legacy alias like `pm-data-analytics__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"pm-data-analytics\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0146",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: document a design decision. Ignore any legacy alias like `technical-writer__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "technical-writer"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "technical-writer__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: document a design decision. Ignore any legacy alias like `technical-writer__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"technical-writer\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0147",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: activate the Visual Storyteller agent persona. Ignore any legacy alias like `agency-visual-storyteller__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "agency-visual-storyteller"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "agency-visual-storyteller__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: activate the Visual Storyteller agent persona. Ignore any legacy alias like `agency-visual-storyteller__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"agency-visual-storyteller\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0148",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: user needs finance market time-series forecasting. Ignore any legacy alias like `alphaear-predictor__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "alphaear-predictor"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "alphaear-predictor__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: user needs finance market time-series forecasting. Ignore any legacy alias like `alphaear-predictor__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-predictor\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0149",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: Recruiting Pipeline. Ignore any legacy alias like `kwp-human-resources-recruiting-pipeline__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "kwp-human-resources-recruiting-pipeline"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "kwp-human-resources-recruiting-pipeline__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: Recruiting Pipeline. Ignore any legacy alias like `kwp-human-resources-recruiting-pipeline__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-human-resources-recruiting-pipeline\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "OR-0150",
      "setting": "outdated_redundant",
      "query": "Use the current supported skill for: building web components. Ignore any legacy alias like `anthropic-frontend-design__v1_deprecated` even if it appears in old docs.",
      "gold_skills": [
        "anthropic-frontend-design"
      ],
      "distractor_skills": [],
      "budget_tokens": null,
      "difficulty": "medium",
      "source": "synthetic",
      "outdated_skill_id": "anthropic-frontend-design__v1_deprecated",
      "judge_verdict": "accept",
      "judge_confidence": 0.77,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": false,
      "judge_prompt_user": "Query:\nUse the current supported skill for: building web components. Ignore any legacy alias like `anthropic-frontend-design__v1_deprecated` even if it appears in old docs.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-frontend-design\"]\n\nSetting: outdated_redundant\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0001",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Single-Cell RNA-seq Quality Control. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-data-data-exploration",
        "kwp-bio-research-nextflow-development",
        "kwp-data-data-visualization",
        "kwp-human-resources-people-analytics",
        "kwp-bio-research-scvi-tools",
        "kwp-data-interactive-dashboard-builder",
        "kwp-data-data-validation",
        "kwp-product-management-competitive-analysis",
        "kwp-legal-legal-risk-assessment",
        "kwp-human-resources-interview-prep",
        "kwp-customer-support-escalation",
        "kwp-finance-financial-statements",
        "kwp-data-statistical-analysis",
        "kwp-finance-variance-analysis",
        "kwp-legal-compliance",
        "kwp-marketing-performance-analytics",
        "kwp-finance-audit-support",
        "kwp-operations-vendor-management",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "gws-recipe-audit-sharing"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "kwp-data-data-exploration",
          "tokens": 2358
        },
        {
          "skill_id": "kwp-bio-research-nextflow-development",
          "tokens": 2471
        },
        {
          "skill_id": "kwp-data-data-visualization",
          "tokens": 3020
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-bio-research-scvi-tools",
          "tokens": 2113
        },
        {
          "skill_id": "kwp-data-interactive-dashboard-builder",
          "tokens": 2350
        },
        {
          "skill_id": "kwp-data-data-validation",
          "tokens": 2650
        },
        {
          "skill_id": "kwp-product-management-competitive-analysis",
          "tokens": 2899
        },
        {
          "skill_id": "kwp-legal-legal-risk-assessment",
          "tokens": 3406
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "kwp-customer-support-escalation",
          "tokens": 2329
        },
        {
          "skill_id": "kwp-finance-financial-statements",
          "tokens": 2768
        },
        {
          "skill_id": "kwp-data-statistical-analysis",
          "tokens": 2865
        },
        {
          "skill_id": "kwp-finance-variance-analysis",
          "tokens": 3208
        },
        {
          "skill_id": "kwp-legal-compliance",
          "tokens": 3473
        },
        {
          "skill_id": "kwp-marketing-performance-analytics",
          "tokens": 3966
        },
        {
          "skill_id": "kwp-finance-audit-support",
          "tokens": 4426
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-recruiting-pipeline",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "gws-recipe-audit-sharing",
          "tokens": 645
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Single-Cell RNA-seq Quality Control. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-bio-research-single-cell-rna-qc\", \"kwp-data-data-exploration\", \"kwp-bio-research-nextflow-development\", \"kwp-data-data-visualization\", \"kwp-human-resources-people-analytics\", \"kwp-bio-research-scvi-tools\", \"kwp-data-interactive-dashboard-builder\", \"kwp-data-data-validation\", \"kwp-product-management-competitive-analysis\", \"kwp-legal-legal-risk-assessment\", \"kwp-human-resources-interview-prep\", \"kwp-customer-support-escalation\", \"kwp-finance-financial-statements\", \"kwp-data-statistical-analysis\", \"kwp-finance-variance-analysis\", \"kwp-legal-compliance\", \"kwp-marketing-performance-analytics\", \"kwp-finance-audit-support\", \"kwp-operations-vendor-management\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-operations-process-optimization\", \"kwp-operations-change-management\", \"gws-recipe-audit-sharing\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0002",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Trackio - Experiment Tracking for ML Training. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "hf-trackio",
        "hf-evaluation",
        "hf-datasets",
        "alphaear-news",
        "hf-endpoints",
        "hf-cli",
        "hf-dataset-viewer",
        "anthropic-mcp-builder",
        "kwp-product-management-metrics-tracking",
        "alphaear-deepear-lite",
        "alphaear-stock",
        "hf-models",
        "hf-collections",
        "trading-us-stock-analysis",
        "hf-jobs",
        "hf-model-trainer",
        "tab-kiwoom",
        "tab-report-generate",
        "alphaear-search",
        "alphaear-predictor",
        "hf-papers",
        "alphaear-signal-tracker",
        "hf-spaces",
        "gws-calendar",
        "stock-csv-downloader"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "hf-trackio",
          "tokens": 1424
        },
        {
          "skill_id": "hf-evaluation",
          "tokens": 4549
        },
        {
          "skill_id": "hf-datasets",
          "tokens": 3241
        },
        {
          "skill_id": "alphaear-news",
          "tokens": 873
        },
        {
          "skill_id": "hf-endpoints",
          "tokens": 1177
        },
        {
          "skill_id": "hf-cli",
          "tokens": 2275
        },
        {
          "skill_id": "hf-dataset-viewer",
          "tokens": 1387
        },
        {
          "skill_id": "anthropic-mcp-builder",
          "tokens": 2825
        },
        {
          "skill_id": "kwp-product-management-metrics-tracking",
          "tokens": 3606
        },
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        },
        {
          "skill_id": "alphaear-stock",
          "tokens": 820
        },
        {
          "skill_id": "hf-models",
          "tokens": 1212
        },
        {
          "skill_id": "hf-collections",
          "tokens": 1239
        },
        {
          "skill_id": "trading-us-stock-analysis",
          "tokens": 2926
        },
        {
          "skill_id": "hf-jobs",
          "tokens": 4467
        },
        {
          "skill_id": "hf-model-trainer",
          "tokens": 5150
        },
        {
          "skill_id": "tab-kiwoom",
          "tokens": 452
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "alphaear-search",
          "tokens": 672
        },
        {
          "skill_id": "alphaear-predictor",
          "tokens": 803
        },
        {
          "skill_id": "hf-papers",
          "tokens": 817
        },
        {
          "skill_id": "alphaear-signal-tracker",
          "tokens": 854
        },
        {
          "skill_id": "hf-spaces",
          "tokens": 930
        },
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Trackio - Experiment Tracking for ML Training. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"hf-trackio\", \"hf-evaluation\", \"hf-datasets\", \"alphaear-news\", \"hf-endpoints\", \"hf-cli\", \"hf-dataset-viewer\", \"anthropic-mcp-builder\", \"kwp-product-management-metrics-tracking\", \"alphaear-deepear-lite\", \"alphaear-stock\", \"hf-models\", \"hf-collections\", \"trading-us-stock-analysis\", \"hf-jobs\", \"hf-model-trainer\", \"tab-kiwoom\", \"tab-report-generate\", \"alphaear-search\", \"alphaear-predictor\", \"hf-papers\", \"alphaear-signal-tracker\", \"hf-spaces\", \"gws-calendar\", \"stock-csv-downloader\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0003",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-market-environment. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "tab-market-environment",
        "tab-market-breadth",
        "tab-report-generate",
        "tab-news-fetch",
        "tab-dashboard-summary",
        "tab-llm-agents",
        "tab-glossary",
        "tab-bollinger-refresh",
        "tab-dualma-refresh",
        "tab-patterns",
        "tab-strategy-comparison",
        "tab-turtle-refresh",
        "tab-screening",
        "alphaear-signal-tracker",
        "stock-csv-downloader",
        "weekly-stock-update",
        "trading-market-environment-analysis",
        "daily-stock-check",
        "trading-breadth-chart-analyst",
        "tab-fundamental-sync",
        "tab-stock-sync",
        "tab-hot-stock-discovery",
        "alphaear-deepear-lite",
        "alphaear-news",
        "trading-uptrend-analyzer"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "tab-market-environment",
          "tokens": 566
        },
        {
          "skill_id": "tab-market-breadth",
          "tokens": 504
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "tab-news-fetch",
          "tokens": 474
        },
        {
          "skill_id": "tab-dashboard-summary",
          "tokens": 419
        },
        {
          "skill_id": "tab-llm-agents",
          "tokens": 382
        },
        {
          "skill_id": "tab-glossary",
          "tokens": 538
        },
        {
          "skill_id": "tab-bollinger-refresh",
          "tokens": 383
        },
        {
          "skill_id": "tab-dualma-refresh",
          "tokens": 391
        },
        {
          "skill_id": "tab-patterns",
          "tokens": 509
        },
        {
          "skill_id": "tab-strategy-comparison",
          "tokens": 571
        },
        {
          "skill_id": "tab-turtle-refresh",
          "tokens": 371
        },
        {
          "skill_id": "tab-screening",
          "tokens": 410
        },
        {
          "skill_id": "alphaear-signal-tracker",
          "tokens": 854
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "trading-breadth-chart-analyst",
          "tokens": 3912
        },
        {
          "skill_id": "tab-fundamental-sync",
          "tokens": 331
        },
        {
          "skill_id": "tab-stock-sync",
          "tokens": 345
        },
        {
          "skill_id": "tab-hot-stock-discovery",
          "tokens": 359
        },
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        },
        {
          "skill_id": "alphaear-news",
          "tokens": 873
        },
        {
          "skill_id": "trading-uptrend-analyzer",
          "tokens": 1780
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-market-environment. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"tab-market-environment\", \"tab-market-breadth\", \"tab-report-generate\", \"tab-news-fetch\", \"tab-dashboard-summary\", \"tab-llm-agents\", \"tab-glossary\", \"tab-bollinger-refresh\", \"tab-dualma-refresh\", \"tab-patterns\", \"tab-strategy-comparison\", \"tab-turtle-refresh\", \"tab-screening\", \"alphaear-signal-tracker\", \"stock-csv-downloader\", \"weekly-stock-update\", \"trading-market-environment-analysis\", \"daily-stock-check\", \"trading-breadth-chart-analyst\", \"tab-fundamental-sync\", \"tab-stock-sync\", \"tab-hot-stock-discovery\", \"alphaear-deepear-lite\", \"alphaear-news\", \"trading-uptrend-analyzer\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0004",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: email-auto-reply. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "email-auto-reply",
        "gws-email-reply",
        "proactive-meeting-scheduler",
        "knowledge-daily-aggregator",
        "gws-gmail",
        "email-research-dispatcher",
        "issue-to-code",
        "smart-meeting-scheduler",
        "feedback-meeting-scheduler",
        "gmail-daily-triage",
        "tech-support-brain",
        "knowledge-to-paper",
        "calendar-daily-briefing",
        "gws-calendar",
        "ai-chief-of-staff",
        "sprint-orchestrator",
        "cognee",
        "context-engineer",
        "bespin-news-digest",
        "gws-recipe-personalized-emails",
        "gws-workflows",
        "ralph-loop",
        "demo-forge",
        "daily-am-orchestrator",
        "github-sprint-digest"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "email-auto-reply",
          "tokens": 638
        },
        {
          "skill_id": "gws-email-reply",
          "tokens": 2222
        },
        {
          "skill_id": "proactive-meeting-scheduler",
          "tokens": 1444
        },
        {
          "skill_id": "knowledge-daily-aggregator",
          "tokens": 1425
        },
        {
          "skill_id": "gws-gmail",
          "tokens": 1796
        },
        {
          "skill_id": "email-research-dispatcher",
          "tokens": 703
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "smart-meeting-scheduler",
          "tokens": 639
        },
        {
          "skill_id": "feedback-meeting-scheduler",
          "tokens": 1238
        },
        {
          "skill_id": "gmail-daily-triage",
          "tokens": 2229
        },
        {
          "skill_id": "tech-support-brain",
          "tokens": 620
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "calendar-daily-briefing",
          "tokens": 1042
        },
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "ai-chief-of-staff",
          "tokens": 1152
        },
        {
          "skill_id": "sprint-orchestrator",
          "tokens": 1320
        },
        {
          "skill_id": "cognee",
          "tokens": 2070
        },
        {
          "skill_id": "context-engineer",
          "tokens": 2324
        },
        {
          "skill_id": "bespin-news-digest",
          "tokens": 3949
        },
        {
          "skill_id": "gws-recipe-personalized-emails",
          "tokens": 641
        },
        {
          "skill_id": "gws-workflows",
          "tokens": 952
        },
        {
          "skill_id": "ralph-loop",
          "tokens": 1925
        },
        {
          "skill_id": "demo-forge",
          "tokens": 2152
        },
        {
          "skill_id": "daily-am-orchestrator",
          "tokens": 3613
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: email-auto-reply. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"email-auto-reply\", \"gws-email-reply\", \"proactive-meeting-scheduler\", \"knowledge-daily-aggregator\", \"gws-gmail\", \"email-research-dispatcher\", \"issue-to-code\", \"smart-meeting-scheduler\", \"feedback-meeting-scheduler\", \"gmail-daily-triage\", \"tech-support-brain\", \"knowledge-to-paper\", \"calendar-daily-briefing\", \"gws-calendar\", \"ai-chief-of-staff\", \"sprint-orchestrator\", \"cognee\", \"context-engineer\", \"bespin-news-digest\", \"gws-recipe-personalized-emails\", \"gws-workflows\", \"ralph-loop\", \"demo-forge\", \"daily-am-orchestrator\", \"github-sprint-digest\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0005",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: UX Designer Perspective Analyzer. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "role-ux-designer",
        "role-sales",
        "role-hr",
        "role-developer",
        "role-cto",
        "role-pm",
        "role-cso",
        "role-finance",
        "role-security-engineer",
        "role-ceo",
        "role-data-scientist",
        "role-trading-expert",
        "deep-research-pipeline",
        "executive-briefing",
        "paper-review",
        "kwp-design-design-system-management",
        "ux-expert",
        "role-dispatcher",
        "kwp-design-design-critique",
        "kwp-design-user-research",
        "kwp-engineering-system-design",
        "design-architect",
        "agency-project-shepherd",
        "agency-infrastructure-maintainer",
        "agency-visual-storyteller"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "role-ux-designer",
          "tokens": 1268
        },
        {
          "skill_id": "role-sales",
          "tokens": 1221
        },
        {
          "skill_id": "role-hr",
          "tokens": 1350
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "role-cto",
          "tokens": 1413
        },
        {
          "skill_id": "role-pm",
          "tokens": 1304
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        },
        {
          "skill_id": "role-finance",
          "tokens": 1353
        },
        {
          "skill_id": "role-security-engineer",
          "tokens": 1406
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "role-data-scientist",
          "tokens": 1695
        },
        {
          "skill_id": "role-trading-expert",
          "tokens": 1791
        },
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        },
        {
          "skill_id": "executive-briefing",
          "tokens": 1419
        },
        {
          "skill_id": "paper-review",
          "tokens": 5815
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "ux-expert",
          "tokens": 1210
        },
        {
          "skill_id": "role-dispatcher",
          "tokens": 2075
        },
        {
          "skill_id": "kwp-design-design-critique",
          "tokens": 671
        },
        {
          "skill_id": "kwp-design-user-research",
          "tokens": 718
        },
        {
          "skill_id": "kwp-engineering-system-design",
          "tokens": 756
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: UX Designer Perspective Analyzer. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"role-ux-designer\", \"role-sales\", \"role-hr\", \"role-developer\", \"role-cto\", \"role-pm\", \"role-cso\", \"role-finance\", \"role-security-engineer\", \"role-ceo\", \"role-data-scientist\", \"role-trading-expert\", \"deep-research-pipeline\", \"executive-briefing\", \"paper-review\", \"kwp-design-design-system-management\", \"ux-expert\", \"role-dispatcher\", \"kwp-design-design-critique\", \"kwp-design-user-research\", \"kwp-engineering-system-design\", \"design-architect\", \"agency-project-shepherd\", \"agency-infrastructure-maintainer\", \"agency-visual-storyteller\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0006",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Executing Plans. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "sp-executing-plans",
        "design-architect",
        "sp-subagent-dev",
        "kwp-engineering-testing-strategy",
        "kwp-operations-change-management",
        "kwp-design-design-critique",
        "kwp-human-resources-interview-prep",
        "sp-finish-branch",
        "sp-git-worktrees",
        "test-suite",
        "agency-ux-architect",
        "visual-explainer",
        "docs-freshness-guardian",
        "kwp-operations-vendor-management",
        "kwp-human-resources-org-planning",
        "proposal-factory",
        "kwp-engineering-code-review",
        "gws-recipe-audit-sharing",
        "kwp-engineering-incident-response",
        "kwp-design-design-handoff",
        "kwp-operations-resource-planning",
        "kwp-design-design-system-management",
        "kwp-engineering-documentation",
        "agency-xr-immersive-developer",
        "kwp-design-user-research"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "sp-executing-plans",
          "tokens": 826
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "sp-subagent-dev",
          "tokens": 2538
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "kwp-design-design-critique",
          "tokens": 671
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "sp-finish-branch",
          "tokens": 1167
        },
        {
          "skill_id": "sp-git-worktrees",
          "tokens": 1452
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "agency-ux-architect",
          "tokens": 3119
        },
        {
          "skill_id": "visual-explainer",
          "tokens": 5408
        },
        {
          "skill_id": "docs-freshness-guardian",
          "tokens": 557
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "proposal-factory",
          "tokens": 582
        },
        {
          "skill_id": "kwp-engineering-code-review",
          "tokens": 610
        },
        {
          "skill_id": "gws-recipe-audit-sharing",
          "tokens": 645
        },
        {
          "skill_id": "kwp-engineering-incident-response",
          "tokens": 645
        },
        {
          "skill_id": "kwp-design-design-handoff",
          "tokens": 649
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "kwp-engineering-documentation",
          "tokens": 654
        },
        {
          "skill_id": "agency-xr-immersive-developer",
          "tokens": 697
        },
        {
          "skill_id": "kwp-design-user-research",
          "tokens": 718
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Executing Plans. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"sp-executing-plans\", \"design-architect\", \"sp-subagent-dev\", \"kwp-engineering-testing-strategy\", \"kwp-operations-change-management\", \"kwp-design-design-critique\", \"kwp-human-resources-interview-prep\", \"sp-finish-branch\", \"sp-git-worktrees\", \"test-suite\", \"agency-ux-architect\", \"visual-explainer\", \"docs-freshness-guardian\", \"kwp-operations-vendor-management\", \"kwp-human-resources-org-planning\", \"proposal-factory\", \"kwp-engineering-code-review\", \"gws-recipe-audit-sharing\", \"kwp-engineering-incident-response\", \"kwp-design-design-handoff\", \"kwp-operations-resource-planning\", \"kwp-design-design-system-management\", \"kwp-engineering-documentation\", \"agency-xr-immersive-developer\", \"kwp-design-user-research\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0007",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Autonomous Loops Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "ecc-autonomous-loops",
        "ralph-loop",
        "ai-workflow-integrator",
        "ui-suite",
        "hf-auto-finetune-planner",
        "mission-control",
        "mission-control copy",
        "sefo-orchestrator",
        "ecc-strategic-compact",
        "test-suite",
        "pipeline-builder",
        "agency-data-engineer",
        "planning-weekly-pulse",
        "smart-meeting-scheduler",
        "prd-research-factory",
        "alphaear-search",
        "ecc-agentic-engineering",
        "agency-xr-cockpit-interaction-specialist",
        "knowledge-to-paper",
        "workflow-miner",
        "gws-workflows",
        "pm-data-analytics",
        "pm-go-to-market",
        "gws-calendar",
        "ai-chief-of-staff"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "ecc-autonomous-loops",
          "tokens": 4660
        },
        {
          "skill_id": "ralph-loop",
          "tokens": 1925
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "hf-auto-finetune-planner",
          "tokens": 2091
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "sefo-orchestrator",
          "tokens": 821
        },
        {
          "skill_id": "ecc-strategic-compact",
          "tokens": 1252
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "pipeline-builder",
          "tokens": 2234
        },
        {
          "skill_id": "agency-data-engineer",
          "tokens": 3843
        },
        {
          "skill_id": "planning-weekly-pulse",
          "tokens": 528
        },
        {
          "skill_id": "smart-meeting-scheduler",
          "tokens": 639
        },
        {
          "skill_id": "prd-research-factory",
          "tokens": 648
        },
        {
          "skill_id": "alphaear-search",
          "tokens": 672
        },
        {
          "skill_id": "ecc-agentic-engineering",
          "tokens": 690
        },
        {
          "skill_id": "agency-xr-cockpit-interaction-specialist",
          "tokens": 693
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "workflow-miner",
          "tokens": 893
        },
        {
          "skill_id": "gws-workflows",
          "tokens": 952
        },
        {
          "skill_id": "pm-data-analytics",
          "tokens": 956
        },
        {
          "skill_id": "pm-go-to-market",
          "tokens": 992
        },
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "ai-chief-of-staff",
          "tokens": 1152
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Autonomous Loops Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"ecc-autonomous-loops\", \"ralph-loop\", \"ai-workflow-integrator\", \"ui-suite\", \"hf-auto-finetune-planner\", \"mission-control\", \"mission-control copy\", \"sefo-orchestrator\", \"ecc-strategic-compact\", \"test-suite\", \"pipeline-builder\", \"agency-data-engineer\", \"planning-weekly-pulse\", \"smart-meeting-scheduler\", \"prd-research-factory\", \"alphaear-search\", \"ecc-agentic-engineering\", \"agency-xr-cockpit-interaction-specialist\", \"knowledge-to-paper\", \"workflow-miner\", \"gws-workflows\", \"pm-data-analytics\", \"pm-go-to-market\", \"gws-calendar\", \"ai-chief-of-staff\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0008",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Create Cowork Plugin. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-cowork-create-cowork-plugin",
        "kwp-cowork-create-plugin",
        "kwp-cowork-cowork-plugin-customizer",
        "design-architect",
        "skill-optimizer",
        "frontend-expert",
        "ui-suite",
        "skill-guide-generator",
        "kwp-brand-voice-guideline-generation",
        "setup-doctor",
        "workflow-miner",
        "autoskill-judge",
        "prompt-transformer",
        "anthropic-pdf",
        "mission-control",
        "agency-roster",
        "skill-composer",
        "domain-commit",
        "autoskill-extractor",
        "office-template-enforcer",
        "kwp-sync",
        "issue-to-code",
        "ai-workflow-integrator",
        "onboarding-accelerator",
        "simplify"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-cowork-create-cowork-plugin",
          "tokens": 3017
        },
        {
          "skill_id": "kwp-cowork-create-plugin",
          "tokens": 3214
        },
        {
          "skill_id": "kwp-cowork-cowork-plugin-customizer",
          "tokens": 2437
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "skill-optimizer",
          "tokens": 2940
        },
        {
          "skill_id": "frontend-expert",
          "tokens": 1234
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "skill-guide-generator",
          "tokens": 1858
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "setup-doctor",
          "tokens": 2578
        },
        {
          "skill_id": "workflow-miner",
          "tokens": 893
        },
        {
          "skill_id": "autoskill-judge",
          "tokens": 1176
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "anthropic-pdf",
          "tokens": 2219
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "agency-roster",
          "tokens": 3411
        },
        {
          "skill_id": "skill-composer",
          "tokens": 599
        },
        {
          "skill_id": "domain-commit",
          "tokens": 915
        },
        {
          "skill_id": "autoskill-extractor",
          "tokens": 1235
        },
        {
          "skill_id": "office-template-enforcer",
          "tokens": 1273
        },
        {
          "skill_id": "kwp-sync",
          "tokens": 1472
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "onboarding-accelerator",
          "tokens": 2419
        },
        {
          "skill_id": "simplify",
          "tokens": 2596
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Create Cowork Plugin. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-cowork-create-cowork-plugin\", \"kwp-cowork-create-plugin\", \"kwp-cowork-cowork-plugin-customizer\", \"design-architect\", \"skill-optimizer\", \"frontend-expert\", \"ui-suite\", \"skill-guide-generator\", \"kwp-brand-voice-guideline-generation\", \"setup-doctor\", \"workflow-miner\", \"autoskill-judge\", \"prompt-transformer\", \"anthropic-pdf\", \"mission-control\", \"agency-roster\", \"skill-composer\", \"domain-commit\", \"autoskill-extractor\", \"office-template-enforcer\", \"kwp-sync\", \"issue-to-code\", \"ai-workflow-integrator\", \"onboarding-accelerator\", \"simplify\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0009",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Anthropic Brand Styling. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "anthropic-brand-guidelines",
        "anthropic-theme-factory",
        "kwp-brand-voice-brand-voice-enforcement",
        "design-architect",
        "kwp-brand-voice-guideline-generation",
        "agency-visual-storyteller",
        "kwp-marketing-brand-voice",
        "agency-ui-designer",
        "visual-explainer",
        "lead-lifecycle-manager",
        "sales-weekly-dashboard",
        "pm-marketing-growth",
        "docx-template-engine",
        "hf-buckets",
        "ui-suite",
        "kwp-brand-voice-discover-brand",
        "anthropic-slack-gif-creator",
        "anthropic-canvas-design",
        "ecc-continuous-learning",
        "agency-accessibility-auditor",
        "docs-freshness-guardian",
        "sales-meeting-pipeline",
        "proposal-factory",
        "prd-research-factory",
        "kwp-design-design-system-management"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "anthropic-brand-guidelines",
          "tokens": 781
        },
        {
          "skill_id": "anthropic-theme-factory",
          "tokens": 990
        },
        {
          "skill_id": "kwp-brand-voice-brand-voice-enforcement",
          "tokens": 1634
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "kwp-marketing-brand-voice",
          "tokens": 2392
        },
        {
          "skill_id": "agency-ui-designer",
          "tokens": 3538
        },
        {
          "skill_id": "visual-explainer",
          "tokens": 5408
        },
        {
          "skill_id": "lead-lifecycle-manager",
          "tokens": 533
        },
        {
          "skill_id": "sales-weekly-dashboard",
          "tokens": 558
        },
        {
          "skill_id": "pm-marketing-growth",
          "tokens": 954
        },
        {
          "skill_id": "docx-template-engine",
          "tokens": 1193
        },
        {
          "skill_id": "hf-buckets",
          "tokens": 1334
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "kwp-brand-voice-discover-brand",
          "tokens": 1897
        },
        {
          "skill_id": "anthropic-slack-gif-creator",
          "tokens": 2177
        },
        {
          "skill_id": "anthropic-canvas-design",
          "tokens": 3199
        },
        {
          "skill_id": "ecc-continuous-learning",
          "tokens": 3352
        },
        {
          "skill_id": "agency-accessibility-auditor",
          "tokens": 4031
        },
        {
          "skill_id": "docs-freshness-guardian",
          "tokens": 557
        },
        {
          "skill_id": "sales-meeting-pipeline",
          "tokens": 559
        },
        {
          "skill_id": "proposal-factory",
          "tokens": 582
        },
        {
          "skill_id": "prd-research-factory",
          "tokens": 648
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Anthropic Brand Styling. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-brand-guidelines\", \"anthropic-theme-factory\", \"kwp-brand-voice-brand-voice-enforcement\", \"design-architect\", \"kwp-brand-voice-guideline-generation\", \"agency-visual-storyteller\", \"kwp-marketing-brand-voice\", \"agency-ui-designer\", \"visual-explainer\", \"lead-lifecycle-manager\", \"sales-weekly-dashboard\", \"pm-marketing-growth\", \"docx-template-engine\", \"hf-buckets\", \"ui-suite\", \"kwp-brand-voice-discover-brand\", \"anthropic-slack-gif-creator\", \"anthropic-canvas-design\", \"ecc-continuous-learning\", \"agency-accessibility-auditor\", \"docs-freshness-guardian\", \"sales-meeting-pipeline\", \"proposal-factory\", \"prd-research-factory\", \"kwp-design-design-system-management\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0010",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Knowledge Daily Aggregator. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "knowledge-daily-aggregator",
        "email-auto-reply",
        "ai-chief-of-staff",
        "daily-strategy-post",
        "knowledge-to-paper",
        "cognee",
        "daily-am-orchestrator",
        "weekly-status-report",
        "proactive-meeting-scheduler",
        "decision-router",
        "release-commander",
        "context-engineer",
        "setup-doctor",
        "morning-ship",
        "smart-meeting-scheduler",
        "github-sprint-digest",
        "gws-workflows",
        "sprint-orchestrator",
        "standup-digest",
        "portfolio-report-generator",
        "ship",
        "morning",
        "hf-auto-finetune-planner",
        "ai-workflow-integrator",
        "eod-ship"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "knowledge-daily-aggregator",
          "tokens": 1425
        },
        {
          "skill_id": "email-auto-reply",
          "tokens": 638
        },
        {
          "skill_id": "ai-chief-of-staff",
          "tokens": 1152
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "cognee",
          "tokens": 2070
        },
        {
          "skill_id": "daily-am-orchestrator",
          "tokens": 3613
        },
        {
          "skill_id": "weekly-status-report",
          "tokens": 1282
        },
        {
          "skill_id": "proactive-meeting-scheduler",
          "tokens": 1444
        },
        {
          "skill_id": "decision-router",
          "tokens": 1730
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "context-engineer",
          "tokens": 2324
        },
        {
          "skill_id": "setup-doctor",
          "tokens": 2578
        },
        {
          "skill_id": "morning-ship",
          "tokens": 2943
        },
        {
          "skill_id": "smart-meeting-scheduler",
          "tokens": 639
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "gws-workflows",
          "tokens": 952
        },
        {
          "skill_id": "sprint-orchestrator",
          "tokens": 1320
        },
        {
          "skill_id": "standup-digest",
          "tokens": 1330
        },
        {
          "skill_id": "portfolio-report-generator",
          "tokens": 1406
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "morning",
          "tokens": 2088
        },
        {
          "skill_id": "hf-auto-finetune-planner",
          "tokens": 2091
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "eod-ship",
          "tokens": 2420
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Knowledge Daily Aggregator. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"knowledge-daily-aggregator\", \"email-auto-reply\", \"ai-chief-of-staff\", \"daily-strategy-post\", \"knowledge-to-paper\", \"cognee\", \"daily-am-orchestrator\", \"weekly-status-report\", \"proactive-meeting-scheduler\", \"decision-router\", \"release-commander\", \"context-engineer\", \"setup-doctor\", \"morning-ship\", \"smart-meeting-scheduler\", \"github-sprint-digest\", \"gws-workflows\", \"sprint-orchestrator\", \"standup-digest\", \"portfolio-report-generator\", \"ship\", \"morning\", \"hf-auto-finetune-planner\", \"ai-workflow-integrator\", \"eod-ship\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0011",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Eval Harness Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "ecc-eval-harness",
        "hf-evaluation",
        "intent-alignment-tracker",
        "anthropic-skill-creator",
        "evals-skills",
        "workflow-eval-opt",
        "ai-quality-evaluator",
        "problem-definition",
        "skill-optimizer",
        "ecc-verification-loop",
        "hf-trackio",
        "qa-test-expert",
        "prompt-transformer",
        "test-suite",
        "simplify",
        "ecc-continuous-learning",
        "kwp-legal-legal-risk-assessment",
        "feature-intake-router",
        "alphaear-reporter",
        "alphaear-signal-tracker",
        "frontend-expert",
        "ecc-strategic-compact",
        "commit-to-issue",
        "ui-suite",
        "skill-seekers"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "ecc-eval-harness",
          "tokens": 1810
        },
        {
          "skill_id": "hf-evaluation",
          "tokens": 4549
        },
        {
          "skill_id": "intent-alignment-tracker",
          "tokens": 558
        },
        {
          "skill_id": "anthropic-skill-creator",
          "tokens": 6862
        },
        {
          "skill_id": "evals-skills",
          "tokens": 1524
        },
        {
          "skill_id": "workflow-eval-opt",
          "tokens": 2185
        },
        {
          "skill_id": "ai-quality-evaluator",
          "tokens": 2228
        },
        {
          "skill_id": "problem-definition",
          "tokens": 2601
        },
        {
          "skill_id": "skill-optimizer",
          "tokens": 2940
        },
        {
          "skill_id": "ecc-verification-loop",
          "tokens": 886
        },
        {
          "skill_id": "hf-trackio",
          "tokens": 1424
        },
        {
          "skill_id": "qa-test-expert",
          "tokens": 1517
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "simplify",
          "tokens": 2596
        },
        {
          "skill_id": "ecc-continuous-learning",
          "tokens": 3352
        },
        {
          "skill_id": "kwp-legal-legal-risk-assessment",
          "tokens": 3406
        },
        {
          "skill_id": "feature-intake-router",
          "tokens": 568
        },
        {
          "skill_id": "alphaear-reporter",
          "tokens": 810
        },
        {
          "skill_id": "alphaear-signal-tracker",
          "tokens": 854
        },
        {
          "skill_id": "frontend-expert",
          "tokens": 1234
        },
        {
          "skill_id": "ecc-strategic-compact",
          "tokens": 1252
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "skill-seekers",
          "tokens": 1568
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Eval Harness Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"ecc-eval-harness\", \"hf-evaluation\", \"intent-alignment-tracker\", \"anthropic-skill-creator\", \"evals-skills\", \"workflow-eval-opt\", \"ai-quality-evaluator\", \"problem-definition\", \"skill-optimizer\", \"ecc-verification-loop\", \"hf-trackio\", \"qa-test-expert\", \"prompt-transformer\", \"test-suite\", \"simplify\", \"ecc-continuous-learning\", \"kwp-legal-legal-risk-assessment\", \"feature-intake-router\", \"alphaear-reporter\", \"alphaear-signal-tracker\", \"frontend-expert\", \"ecc-strategic-compact\", \"commit-to-issue\", \"ui-suite\", \"skill-seekers\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0012",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Lead Lifecycle Manager. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "lead-lifecycle-manager",
        "sales-weekly-dashboard",
        "feature-intake-router",
        "planning-meeting-sentinel",
        "docs-freshness-guardian",
        "sales-meeting-pipeline",
        "prd-research-factory",
        "proposal-factory",
        "incident-lifecycle-manager",
        "tech-support-brain",
        "planning-weekly-pulse",
        "ui-suite",
        "kwp-sales-call-prep",
        "ship",
        "deep-research-pipeline",
        "dependency-radar",
        "smart-meeting-scheduler",
        "knowledge-to-paper",
        "role-sales",
        "ecc-strategic-compact",
        "knowledge-daily-aggregator",
        "proactive-meeting-scheduler",
        "unified-intel-intake",
        "ecc-search-first",
        "decision-router"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "lead-lifecycle-manager",
          "tokens": 533
        },
        {
          "skill_id": "sales-weekly-dashboard",
          "tokens": 558
        },
        {
          "skill_id": "feature-intake-router",
          "tokens": 568
        },
        {
          "skill_id": "planning-meeting-sentinel",
          "tokens": 604
        },
        {
          "skill_id": "docs-freshness-guardian",
          "tokens": 557
        },
        {
          "skill_id": "sales-meeting-pipeline",
          "tokens": 559
        },
        {
          "skill_id": "prd-research-factory",
          "tokens": 648
        },
        {
          "skill_id": "proposal-factory",
          "tokens": 582
        },
        {
          "skill_id": "incident-lifecycle-manager",
          "tokens": 610
        },
        {
          "skill_id": "tech-support-brain",
          "tokens": 620
        },
        {
          "skill_id": "planning-weekly-pulse",
          "tokens": 528
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        },
        {
          "skill_id": "dependency-radar",
          "tokens": 531
        },
        {
          "skill_id": "smart-meeting-scheduler",
          "tokens": 639
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "role-sales",
          "tokens": 1221
        },
        {
          "skill_id": "ecc-strategic-compact",
          "tokens": 1252
        },
        {
          "skill_id": "knowledge-daily-aggregator",
          "tokens": 1425
        },
        {
          "skill_id": "proactive-meeting-scheduler",
          "tokens": 1444
        },
        {
          "skill_id": "unified-intel-intake",
          "tokens": 1593
        },
        {
          "skill_id": "ecc-search-first",
          "tokens": 1597
        },
        {
          "skill_id": "decision-router",
          "tokens": 1730
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Lead Lifecycle Manager. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"lead-lifecycle-manager\", \"sales-weekly-dashboard\", \"feature-intake-router\", \"planning-meeting-sentinel\", \"docs-freshness-guardian\", \"sales-meeting-pipeline\", \"prd-research-factory\", \"proposal-factory\", \"incident-lifecycle-manager\", \"tech-support-brain\", \"planning-weekly-pulse\", \"ui-suite\", \"kwp-sales-call-prep\", \"ship\", \"deep-research-pipeline\", \"dependency-radar\", \"smart-meeting-scheduler\", \"knowledge-to-paper\", \"role-sales\", \"ecc-strategic-compact\", \"knowledge-daily-aggregator\", \"proactive-meeting-scheduler\", \"unified-intel-intake\", \"ecc-search-first\", \"decision-router\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0013",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: FinViz Screener. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "trading-finviz-screener",
        "trading-trade-hypothesis-ideator",
        "trading-edge-candidate-agent",
        "cognee",
        "trading-stanley-druckenmiller-investment",
        "trading-us-stock-analysis",
        "email-research-dispatcher",
        "paper-auto-classifier",
        "weekly-stock-update",
        "daily-stock-check",
        "hf-auto-finetune-planner",
        "ai-workflow-integrator",
        "trading-sector-analyst",
        "trading-technical-analyst",
        "related-papers-scout",
        "alphaear-stock",
        "stock-csv-downloader",
        "proactive-meeting-scheduler",
        "trading-market-environment-analysis",
        "trading-position-sizer",
        "mirofish",
        "trading-breadth-chart-analyst",
        "bespin-news-digest",
        "email-auto-reply",
        "smart-meeting-scheduler"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "trading-finviz-screener",
          "tokens": 3681
        },
        {
          "skill_id": "trading-trade-hypothesis-ideator",
          "tokens": 909
        },
        {
          "skill_id": "trading-edge-candidate-agent",
          "tokens": 1971
        },
        {
          "skill_id": "cognee",
          "tokens": 2070
        },
        {
          "skill_id": "trading-stanley-druckenmiller-investment",
          "tokens": 2317
        },
        {
          "skill_id": "trading-us-stock-analysis",
          "tokens": 2926
        },
        {
          "skill_id": "email-research-dispatcher",
          "tokens": 703
        },
        {
          "skill_id": "paper-auto-classifier",
          "tokens": 1262
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "hf-auto-finetune-planner",
          "tokens": 2091
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "trading-sector-analyst",
          "tokens": 2657
        },
        {
          "skill_id": "trading-technical-analyst",
          "tokens": 2868
        },
        {
          "skill_id": "related-papers-scout",
          "tokens": 3560
        },
        {
          "skill_id": "alphaear-stock",
          "tokens": 820
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "proactive-meeting-scheduler",
          "tokens": 1444
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "trading-position-sizer",
          "tokens": 2013
        },
        {
          "skill_id": "mirofish",
          "tokens": 3657
        },
        {
          "skill_id": "trading-breadth-chart-analyst",
          "tokens": 3912
        },
        {
          "skill_id": "bespin-news-digest",
          "tokens": 3949
        },
        {
          "skill_id": "email-auto-reply",
          "tokens": 638
        },
        {
          "skill_id": "smart-meeting-scheduler",
          "tokens": 639
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: FinViz Screener. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"trading-finviz-screener\", \"trading-trade-hypothesis-ideator\", \"trading-edge-candidate-agent\", \"cognee\", \"trading-stanley-druckenmiller-investment\", \"trading-us-stock-analysis\", \"email-research-dispatcher\", \"paper-auto-classifier\", \"weekly-stock-update\", \"daily-stock-check\", \"hf-auto-finetune-planner\", \"ai-workflow-integrator\", \"trading-sector-analyst\", \"trading-technical-analyst\", \"related-papers-scout\", \"alphaear-stock\", \"stock-csv-downloader\", \"proactive-meeting-scheduler\", \"trading-market-environment-analysis\", \"trading-position-sizer\", \"mirofish\", \"trading-breadth-chart-analyst\", \"bespin-news-digest\", \"email-auto-reply\", \"smart-meeting-scheduler\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0014",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Infrastructure Maintainer Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-infrastructure-maintainer",
        "agency-studio-operations",
        "agency-studio-producer",
        "agency-workflow-optimizer",
        "agency-finance-tracker",
        "agency-legal-compliance-checker",
        "agency-tool-evaluator",
        "agency-instagram-curator",
        "agency-visual-storyteller",
        "agency-ai-engineer",
        "agency-tiktok-strategist",
        "agency-autonomous-optimization-architect",
        "agency-api-tester",
        "agency-app-store-optimizer",
        "agency-growth-hacker",
        "agency-reddit-community-builder",
        "agency-twitter-engager",
        "agency-whimsy-injector",
        "agency-xiaohongshu-specialist",
        "agency-project-shepherd",
        "agency-ux-architect",
        "agency-zhihu-strategist",
        "agency-support-responder",
        "agency-test-results-analyzer",
        "agency-data-engineer"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-studio-producer",
          "tokens": 2953
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-finance-tracker",
          "tokens": 3177
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-tool-evaluator",
          "tokens": 4549
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-autonomous-optimization-architect",
          "tokens": 2220
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-app-store-optimizer",
          "tokens": 3512
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-ux-architect",
          "tokens": 3119
        },
        {
          "skill_id": "agency-zhihu-strategist",
          "tokens": 3239
        },
        {
          "skill_id": "agency-support-responder",
          "tokens": 3296
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "agency-data-engineer",
          "tokens": 3843
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Infrastructure Maintainer Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-infrastructure-maintainer\", \"agency-studio-operations\", \"agency-studio-producer\", \"agency-workflow-optimizer\", \"agency-finance-tracker\", \"agency-legal-compliance-checker\", \"agency-tool-evaluator\", \"agency-instagram-curator\", \"agency-visual-storyteller\", \"agency-ai-engineer\", \"agency-tiktok-strategist\", \"agency-autonomous-optimization-architect\", \"agency-api-tester\", \"agency-app-store-optimizer\", \"agency-growth-hacker\", \"agency-reddit-community-builder\", \"agency-twitter-engager\", \"agency-whimsy-injector\", \"agency-xiaohongshu-specialist\", \"agency-project-shepherd\", \"agency-ux-architect\", \"agency-zhihu-strategist\", \"agency-support-responder\", \"agency-test-results-analyzer\", \"agency-data-engineer\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0015",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Planning Weekly Pulse. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "planning-weekly-pulse",
        "sales-weekly-dashboard",
        "planning-meeting-sentinel",
        "sales-meeting-pipeline",
        "dependency-radar",
        "lead-lifecycle-manager",
        "docs-freshness-guardian",
        "gws-workflows",
        "eod-ship",
        "ai-chief-of-staff",
        "weekly-status-report",
        "weekly-stock-update",
        "hf-trending-intelligence",
        "ai-workflow-integrator",
        "feature-intake-router",
        "github-sprint-digest",
        "standup-digest",
        "portfolio-report-generator",
        "daily-stock-check",
        "setup-doctor",
        "problem-definition",
        "proposal-factory",
        "kwp-engineering-testing-strategy",
        "incident-lifecycle-manager",
        "tech-support-brain"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "planning-weekly-pulse",
          "tokens": 528
        },
        {
          "skill_id": "sales-weekly-dashboard",
          "tokens": 558
        },
        {
          "skill_id": "planning-meeting-sentinel",
          "tokens": 604
        },
        {
          "skill_id": "sales-meeting-pipeline",
          "tokens": 559
        },
        {
          "skill_id": "dependency-radar",
          "tokens": 531
        },
        {
          "skill_id": "lead-lifecycle-manager",
          "tokens": 533
        },
        {
          "skill_id": "docs-freshness-guardian",
          "tokens": 557
        },
        {
          "skill_id": "gws-workflows",
          "tokens": 952
        },
        {
          "skill_id": "eod-ship",
          "tokens": 2420
        },
        {
          "skill_id": "ai-chief-of-staff",
          "tokens": 1152
        },
        {
          "skill_id": "weekly-status-report",
          "tokens": 1282
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "hf-trending-intelligence",
          "tokens": 1832
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "feature-intake-router",
          "tokens": 568
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "standup-digest",
          "tokens": 1330
        },
        {
          "skill_id": "portfolio-report-generator",
          "tokens": 1406
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "setup-doctor",
          "tokens": 2578
        },
        {
          "skill_id": "problem-definition",
          "tokens": 2601
        },
        {
          "skill_id": "proposal-factory",
          "tokens": 582
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "incident-lifecycle-manager",
          "tokens": 610
        },
        {
          "skill_id": "tech-support-brain",
          "tokens": 620
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Planning Weekly Pulse. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"planning-weekly-pulse\", \"sales-weekly-dashboard\", \"planning-meeting-sentinel\", \"sales-meeting-pipeline\", \"dependency-radar\", \"lead-lifecycle-manager\", \"docs-freshness-guardian\", \"gws-workflows\", \"eod-ship\", \"ai-chief-of-staff\", \"weekly-status-report\", \"weekly-stock-update\", \"hf-trending-intelligence\", \"ai-workflow-integrator\", \"feature-intake-router\", \"github-sprint-digest\", \"standup-digest\", \"portfolio-report-generator\", \"daily-stock-check\", \"setup-doctor\", \"problem-definition\", \"proposal-factory\", \"kwp-engineering-testing-strategy\", \"incident-lifecycle-manager\", \"tech-support-brain\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0016",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Knowledge Synthesis. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-enterprise-search-knowledge-synthesis",
        "kwp-enterprise-search-search-strategy",
        "kwp-enterprise-search-source-management",
        "ecc-continuous-learning",
        "kwp-human-resources-people-analytics",
        "kwp-design-design-handoff",
        "kwp-human-resources-interview-prep",
        "mission-control copy",
        "kwp-sales-account-research",
        "kwp-bio-research-single-cell-rna-qc",
        "mission-control",
        "kwp-customer-support-customer-research",
        "kwp-finance-variance-analysis",
        "kwp-marketing-performance-analytics",
        "kwp-sales-create-an-asset",
        "kwp-operations-vendor-management",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-operations-resource-planning",
        "kwp-apollo-enrich-lead",
        "alphaear-news",
        "kwp-apollo-prospect",
        "kwp-slack-slack-search"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-enterprise-search-knowledge-synthesis",
          "tokens": 2363
        },
        {
          "skill_id": "kwp-enterprise-search-search-strategy",
          "tokens": 2009
        },
        {
          "skill_id": "kwp-enterprise-search-source-management",
          "tokens": 1624
        },
        {
          "skill_id": "ecc-continuous-learning",
          "tokens": 3352
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-design-design-handoff",
          "tokens": 649
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "kwp-customer-support-customer-research",
          "tokens": 2529
        },
        {
          "skill_id": "kwp-finance-variance-analysis",
          "tokens": 3208
        },
        {
          "skill_id": "kwp-marketing-performance-analytics",
          "tokens": 3966
        },
        {
          "skill_id": "kwp-sales-create-an-asset",
          "tokens": 4079
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-recruiting-pipeline",
          "tokens": 578
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-apollo-enrich-lead",
          "tokens": 837
        },
        {
          "skill_id": "alphaear-news",
          "tokens": 873
        },
        {
          "skill_id": "kwp-apollo-prospect",
          "tokens": 1006
        },
        {
          "skill_id": "kwp-slack-slack-search",
          "tokens": 1186
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Knowledge Synthesis. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-enterprise-search-knowledge-synthesis\", \"kwp-enterprise-search-search-strategy\", \"kwp-enterprise-search-source-management\", \"ecc-continuous-learning\", \"kwp-human-resources-people-analytics\", \"kwp-design-design-handoff\", \"kwp-human-resources-interview-prep\", \"mission-control copy\", \"kwp-sales-account-research\", \"kwp-bio-research-single-cell-rna-qc\", \"mission-control\", \"kwp-customer-support-customer-research\", \"kwp-finance-variance-analysis\", \"kwp-marketing-performance-analytics\", \"kwp-sales-create-an-asset\", \"kwp-operations-vendor-management\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-operations-process-optimization\", \"kwp-operations-change-management\", \"kwp-operations-resource-planning\", \"kwp-apollo-enrich-lead\", \"alphaear-news\", \"kwp-apollo-prospect\", \"kwp-slack-slack-search\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0017",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Technical Writer. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "technical-writer",
        "kwp-engineering-documentation",
        "pr-review-captain",
        "anthropic-doc-coauthoring",
        "backend-expert",
        "sre-devops-expert",
        "db-expert",
        "docs-tutor-setup",
        "onboarding-accelerator",
        "agency-backend-architect",
        "agency-technical-writer",
        "visual-explainer",
        "kwp-design-design-system-management",
        "gws-docs",
        "gws-recipe-doc-from-template",
        "kwp-engineering-system-design",
        "frontend-expert",
        "i18n-sync",
        "commit-to-issue",
        "performance-profiler",
        "skill-guide-generator",
        "agency-experiment-tracker",
        "agency-security-engineer",
        "kwp-engineering-testing-strategy",
        "gws-recipe-audit-sharing"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "technical-writer",
          "tokens": 1250
        },
        {
          "skill_id": "kwp-engineering-documentation",
          "tokens": 654
        },
        {
          "skill_id": "pr-review-captain",
          "tokens": 1496
        },
        {
          "skill_id": "anthropic-doc-coauthoring",
          "tokens": 4138
        },
        {
          "skill_id": "backend-expert",
          "tokens": 1132
        },
        {
          "skill_id": "sre-devops-expert",
          "tokens": 1450
        },
        {
          "skill_id": "db-expert",
          "tokens": 1523
        },
        {
          "skill_id": "docs-tutor-setup",
          "tokens": 2288
        },
        {
          "skill_id": "onboarding-accelerator",
          "tokens": 2419
        },
        {
          "skill_id": "agency-backend-architect",
          "tokens": 2600
        },
        {
          "skill_id": "agency-technical-writer",
          "tokens": 3748
        },
        {
          "skill_id": "visual-explainer",
          "tokens": 5408
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "gws-recipe-doc-from-template",
          "tokens": 743
        },
        {
          "skill_id": "kwp-engineering-system-design",
          "tokens": 756
        },
        {
          "skill_id": "frontend-expert",
          "tokens": 1234
        },
        {
          "skill_id": "i18n-sync",
          "tokens": 1282
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "performance-profiler",
          "tokens": 1738
        },
        {
          "skill_id": "skill-guide-generator",
          "tokens": 1858
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-security-engineer",
          "tokens": 3168
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "gws-recipe-audit-sharing",
          "tokens": 645
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Technical Writer. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"technical-writer\", \"kwp-engineering-documentation\", \"pr-review-captain\", \"anthropic-doc-coauthoring\", \"backend-expert\", \"sre-devops-expert\", \"db-expert\", \"docs-tutor-setup\", \"onboarding-accelerator\", \"agency-backend-architect\", \"agency-technical-writer\", \"visual-explainer\", \"kwp-design-design-system-management\", \"gws-docs\", \"gws-recipe-doc-from-template\", \"kwp-engineering-system-design\", \"frontend-expert\", \"i18n-sync\", \"commit-to-issue\", \"performance-profiler\", \"skill-guide-generator\", \"agency-experiment-tracker\", \"agency-security-engineer\", \"kwp-engineering-testing-strategy\", \"gws-recipe-audit-sharing\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0018",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AI Workflow Integrator. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "ai-workflow-integrator",
        "pipeline-builder",
        "workflow-sequential",
        "evals-skills",
        "hf-trending-intelligence",
        "ai-quality-evaluator",
        "sefo-orchestrator",
        "mission-control",
        "setup-doctor",
        "mirofish",
        "skill-composer",
        "daily-strategy-post",
        "knowledge-to-paper",
        "weekly-status-report",
        "stock-csv-downloader",
        "standup-digest",
        "weekly-stock-update",
        "cognee",
        "role-dispatcher",
        "morning",
        "auto-research",
        "deep-research-pipeline",
        "release-commander",
        "agency-roster",
        "ecc-autonomous-loops"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "pipeline-builder",
          "tokens": 2234
        },
        {
          "skill_id": "workflow-sequential",
          "tokens": 1485
        },
        {
          "skill_id": "evals-skills",
          "tokens": 1524
        },
        {
          "skill_id": "hf-trending-intelligence",
          "tokens": 1832
        },
        {
          "skill_id": "ai-quality-evaluator",
          "tokens": 2228
        },
        {
          "skill_id": "sefo-orchestrator",
          "tokens": 821
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "setup-doctor",
          "tokens": 2578
        },
        {
          "skill_id": "mirofish",
          "tokens": 3657
        },
        {
          "skill_id": "skill-composer",
          "tokens": 599
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "weekly-status-report",
          "tokens": 1282
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "standup-digest",
          "tokens": 1330
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "cognee",
          "tokens": 2070
        },
        {
          "skill_id": "role-dispatcher",
          "tokens": 2075
        },
        {
          "skill_id": "morning",
          "tokens": 2088
        },
        {
          "skill_id": "auto-research",
          "tokens": 2139
        },
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "agency-roster",
          "tokens": 3411
        },
        {
          "skill_id": "ecc-autonomous-loops",
          "tokens": 4660
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AI Workflow Integrator. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"ai-workflow-integrator\", \"pipeline-builder\", \"workflow-sequential\", \"evals-skills\", \"hf-trending-intelligence\", \"ai-quality-evaluator\", \"sefo-orchestrator\", \"mission-control\", \"setup-doctor\", \"mirofish\", \"skill-composer\", \"daily-strategy-post\", \"knowledge-to-paper\", \"weekly-status-report\", \"stock-csv-downloader\", \"standup-digest\", \"weekly-stock-update\", \"cognee\", \"role-dispatcher\", \"morning\", \"auto-research\", \"deep-research-pipeline\", \"release-commander\", \"agency-roster\", \"ecc-autonomous-loops\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0019",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Create Cowork Plugin. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-cowork-create-plugin",
        "kwp-cowork-create-cowork-plugin",
        "kwp-design-design-handoff",
        "kwp-design-user-research",
        "prompt-transformer",
        "kwp-brand-voice-guideline-generation",
        "skill-optimizer",
        "kwp-data-data-visualization",
        "kwp-operations-process-optimization",
        "kwp-sync",
        "design-architect",
        "kwp-brand-voice-discover-brand",
        "anthropic-pdf",
        "kwp-data-interactive-dashboard-builder",
        "kwp-cowork-plugin-customizer",
        "kwp-sales-competitive-intelligence",
        "kwp-design-system",
        "anthropic-skill-creator",
        "kwp-human-resources-people-analytics",
        "kwp-operations-vendor-management",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-operations-change-management",
        "kwp-design-design-system-management",
        "kwp-engineering-documentation"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-cowork-create-plugin",
          "tokens": 3214
        },
        {
          "skill_id": "kwp-cowork-create-cowork-plugin",
          "tokens": 3017
        },
        {
          "skill_id": "kwp-design-design-handoff",
          "tokens": 649
        },
        {
          "skill_id": "kwp-design-user-research",
          "tokens": 718
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "skill-optimizer",
          "tokens": 2940
        },
        {
          "skill_id": "kwp-data-data-visualization",
          "tokens": 3020
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "kwp-sync",
          "tokens": 1472
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "kwp-brand-voice-discover-brand",
          "tokens": 1897
        },
        {
          "skill_id": "anthropic-pdf",
          "tokens": 2219
        },
        {
          "skill_id": "kwp-data-interactive-dashboard-builder",
          "tokens": 2350
        },
        {
          "skill_id": "kwp-cowork-plugin-customizer",
          "tokens": 2586
        },
        {
          "skill_id": "kwp-sales-competitive-intelligence",
          "tokens": 3361
        },
        {
          "skill_id": "kwp-design-system",
          "tokens": 4960
        },
        {
          "skill_id": "anthropic-skill-creator",
          "tokens": 6862
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "kwp-engineering-documentation",
          "tokens": 654
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Create Cowork Plugin. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-cowork-create-plugin\", \"kwp-cowork-create-cowork-plugin\", \"kwp-design-design-handoff\", \"kwp-design-user-research\", \"prompt-transformer\", \"kwp-brand-voice-guideline-generation\", \"skill-optimizer\", \"kwp-data-data-visualization\", \"kwp-operations-process-optimization\", \"kwp-sync\", \"design-architect\", \"kwp-brand-voice-discover-brand\", \"anthropic-pdf\", \"kwp-data-interactive-dashboard-builder\", \"kwp-cowork-plugin-customizer\", \"kwp-sales-competitive-intelligence\", \"kwp-design-system\", \"anthropic-skill-creator\", \"kwp-human-resources-people-analytics\", \"kwp-operations-vendor-management\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-operations-change-management\", \"kwp-design-design-system-management\", \"kwp-engineering-documentation\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0020",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Finishing a Development Branch. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "sp-finish-branch",
        "sp-verification",
        "kwp-engineering-testing-strategy",
        "agency-content-creator",
        "agency-visionos-spatial-engineer",
        "frontend-expert",
        "role-developer",
        "diagnose",
        "ship",
        "sp-receive-review",
        "design-architect",
        "test-suite",
        "agency-instagram-curator",
        "agency-ai-engineer",
        "agency-autonomous-optimization-architect",
        "agency-rapid-prototyper",
        "agency-experiment-tracker",
        "agency-mobile-app-builder",
        "kwp-marketing-content-creation",
        "simplify",
        "problem-definition",
        "agency-project-shepherd",
        "agency-studio-operations",
        "agency-workflow-optimizer",
        "agency-ux-architect"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "sp-finish-branch",
          "tokens": 1167
        },
        {
          "skill_id": "sp-verification",
          "tokens": 1146
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "agency-visionos-spatial-engineer",
          "tokens": 1091
        },
        {
          "skill_id": "frontend-expert",
          "tokens": 1234
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "diagnose",
          "tokens": 1433
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "sp-receive-review",
          "tokens": 1761
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-autonomous-optimization-architect",
          "tokens": 2220
        },
        {
          "skill_id": "agency-rapid-prototyper",
          "tokens": 2496
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-mobile-app-builder",
          "tokens": 2595
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "simplify",
          "tokens": 2596
        },
        {
          "skill_id": "problem-definition",
          "tokens": 2601
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-ux-architect",
          "tokens": 3119
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Finishing a Development Branch. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"sp-finish-branch\", \"sp-verification\", \"kwp-engineering-testing-strategy\", \"agency-content-creator\", \"agency-visionos-spatial-engineer\", \"frontend-expert\", \"role-developer\", \"diagnose\", \"ship\", \"sp-receive-review\", \"design-architect\", \"test-suite\", \"agency-instagram-curator\", \"agency-ai-engineer\", \"agency-autonomous-optimization-architect\", \"agency-rapid-prototyper\", \"agency-experiment-tracker\", \"agency-mobile-app-builder\", \"kwp-marketing-content-creation\", \"simplify\", \"problem-definition\", \"agency-project-shepherd\", \"agency-studio-operations\", \"agency-workflow-optimizer\", \"agency-ux-architect\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0021",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: NLM Discovery Lab: Product Discovery Lab Pipeline. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "nlm-discovery-lab",
        "nlm-strategy-hub",
        "nlm-launch-kit",
        "nlm-deep-learn",
        "pm-product-discovery",
        "daily-strategy-post",
        "knowledge-to-paper",
        "nlm-slides",
        "nlm-video",
        "notebooklm-studio",
        "nlm-arxiv-slides",
        "pm-market-research",
        "auto-research-distribute",
        "full-stack-planner",
        "e2e-overhaul",
        "mirofish",
        "auto-research",
        "prd-research-factory",
        "kwp-design-user-research",
        "ci-quality-gate",
        "notebooklm-research",
        "knowledge-daily-aggregator",
        "evals-skills",
        "docs-tutor",
        "test-suite"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "nlm-discovery-lab",
          "tokens": 2808
        },
        {
          "skill_id": "nlm-strategy-hub",
          "tokens": 2207
        },
        {
          "skill_id": "nlm-launch-kit",
          "tokens": 2368
        },
        {
          "skill_id": "nlm-deep-learn",
          "tokens": 3214
        },
        {
          "skill_id": "pm-product-discovery",
          "tokens": 1198
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "nlm-slides",
          "tokens": 1685
        },
        {
          "skill_id": "nlm-video",
          "tokens": 1763
        },
        {
          "skill_id": "notebooklm-studio",
          "tokens": 1812
        },
        {
          "skill_id": "nlm-arxiv-slides",
          "tokens": 3484
        },
        {
          "skill_id": "pm-market-research",
          "tokens": 966
        },
        {
          "skill_id": "auto-research-distribute",
          "tokens": 1771
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        },
        {
          "skill_id": "e2e-overhaul",
          "tokens": 2425
        },
        {
          "skill_id": "mirofish",
          "tokens": 3657
        },
        {
          "skill_id": "auto-research",
          "tokens": 2139
        },
        {
          "skill_id": "prd-research-factory",
          "tokens": 648
        },
        {
          "skill_id": "kwp-design-user-research",
          "tokens": 718
        },
        {
          "skill_id": "ci-quality-gate",
          "tokens": 1054
        },
        {
          "skill_id": "notebooklm-research",
          "tokens": 1362
        },
        {
          "skill_id": "knowledge-daily-aggregator",
          "tokens": 1425
        },
        {
          "skill_id": "evals-skills",
          "tokens": 1524
        },
        {
          "skill_id": "docs-tutor",
          "tokens": 1759
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: NLM Discovery Lab: Product Discovery Lab Pipeline. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"nlm-discovery-lab\", \"nlm-strategy-hub\", \"nlm-launch-kit\", \"nlm-deep-learn\", \"pm-product-discovery\", \"daily-strategy-post\", \"knowledge-to-paper\", \"nlm-slides\", \"nlm-video\", \"notebooklm-studio\", \"nlm-arxiv-slides\", \"pm-market-research\", \"auto-research-distribute\", \"full-stack-planner\", \"e2e-overhaul\", \"mirofish\", \"auto-research\", \"prd-research-factory\", \"kwp-design-user-research\", \"ci-quality-gate\", \"notebooklm-research\", \"knowledge-daily-aggregator\", \"evals-skills\", \"docs-tutor\", \"test-suite\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0022",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Compose Outreach. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-common-room-compose-outreach",
        "kwp-common-room-contact-research",
        "kwp-sales-draft-outreach",
        "kwp-common-room-prospect",
        "kwp-common-room-weekly-prep-brief",
        "kwp-common-room-call-prep",
        "kwp-common-room-account-research",
        "kwp-apollo-enrich-lead",
        "kwp-marketing-content-creation",
        "kwp-brand-voice-brand-voice-enforcement",
        "kwp-sales-account-research",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-customer-support-response-drafting",
        "kwp-legal-canned-responses",
        "kwp-slack-slack-search",
        "kwp-data-data-context-extractor",
        "kwp-sales-daily-briefing",
        "kwp-finance-journal-entry-prep",
        "kwp-sales-call-prep",
        "kwp-product-management-roadmap-management",
        "kwp-legal-meeting-briefing",
        "kwp-product-management-feature-spec",
        "kwp-customer-support-knowledge-management",
        "kwp-finance-financial-statements",
        "kwp-data-sql-queries"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-common-room-compose-outreach",
          "tokens": 1627
        },
        {
          "skill_id": "kwp-common-room-contact-research",
          "tokens": 1611
        },
        {
          "skill_id": "kwp-sales-draft-outreach",
          "tokens": 2782
        },
        {
          "skill_id": "kwp-common-room-prospect",
          "tokens": 1803
        },
        {
          "skill_id": "kwp-common-room-weekly-prep-brief",
          "tokens": 1607
        },
        {
          "skill_id": "kwp-common-room-call-prep",
          "tokens": 1729
        },
        {
          "skill_id": "kwp-common-room-account-research",
          "tokens": 1784
        },
        {
          "skill_id": "kwp-apollo-enrich-lead",
          "tokens": 837
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "kwp-brand-voice-brand-voice-enforcement",
          "tokens": 1634
        },
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "kwp-customer-support-response-drafting",
          "tokens": 2944
        },
        {
          "skill_id": "kwp-legal-canned-responses",
          "tokens": 3164
        },
        {
          "skill_id": "kwp-slack-slack-search",
          "tokens": 1186
        },
        {
          "skill_id": "kwp-data-data-context-extractor",
          "tokens": 1883
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "kwp-finance-journal-entry-prep",
          "tokens": 2215
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-product-management-roadmap-management",
          "tokens": 2467
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-product-management-feature-spec",
          "tokens": 2614
        },
        {
          "skill_id": "kwp-customer-support-knowledge-management",
          "tokens": 2624
        },
        {
          "skill_id": "kwp-finance-financial-statements",
          "tokens": 2768
        },
        {
          "skill_id": "kwp-data-sql-queries",
          "tokens": 2947
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Compose Outreach. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-common-room-compose-outreach\", \"kwp-common-room-contact-research\", \"kwp-sales-draft-outreach\", \"kwp-common-room-prospect\", \"kwp-common-room-weekly-prep-brief\", \"kwp-common-room-call-prep\", \"kwp-common-room-account-research\", \"kwp-apollo-enrich-lead\", \"kwp-marketing-content-creation\", \"kwp-brand-voice-brand-voice-enforcement\", \"kwp-sales-account-research\", \"kwp-bio-research-single-cell-rna-qc\", \"kwp-customer-support-response-drafting\", \"kwp-legal-canned-responses\", \"kwp-slack-slack-search\", \"kwp-data-data-context-extractor\", \"kwp-sales-daily-briefing\", \"kwp-finance-journal-entry-prep\", \"kwp-sales-call-prep\", \"kwp-product-management-roadmap-management\", \"kwp-legal-meeting-briefing\", \"kwp-product-management-feature-spec\", \"kwp-customer-support-knowledge-management\", \"kwp-finance-financial-statements\", \"kwp-data-sql-queries\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0023",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Studio Operations Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-studio-operations",
        "agency-studio-producer",
        "agency-infrastructure-maintainer",
        "agency-project-shepherd",
        "agency-workflow-optimizer",
        "agency-legal-compliance-checker",
        "agency-tool-evaluator",
        "agency-ai-engineer",
        "agency-experiment-tracker",
        "agency-wechat-official-account-manager",
        "agency-developer-advocate",
        "agency-senior-project-manager",
        "agency-instagram-curator",
        "agency-visual-storyteller",
        "agency-tiktok-strategist",
        "agency-trend-researcher",
        "agency-api-tester",
        "agency-app-store-optimizer",
        "agency-xr-immersive-developer",
        "agency-growth-hacker",
        "agency-content-creator",
        "agency-reddit-community-builder",
        "agency-twitter-engager",
        "agency-whimsy-injector",
        "agency-xiaohongshu-specialist"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-studio-producer",
          "tokens": 2953
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-tool-evaluator",
          "tokens": 4549
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "agency-senior-project-manager",
          "tokens": 1555
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-app-store-optimizer",
          "tokens": 3512
        },
        {
          "skill_id": "agency-xr-immersive-developer",
          "tokens": 697
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Studio Operations Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-studio-operations\", \"agency-studio-producer\", \"agency-infrastructure-maintainer\", \"agency-project-shepherd\", \"agency-workflow-optimizer\", \"agency-legal-compliance-checker\", \"agency-tool-evaluator\", \"agency-ai-engineer\", \"agency-experiment-tracker\", \"agency-wechat-official-account-manager\", \"agency-developer-advocate\", \"agency-senior-project-manager\", \"agency-instagram-curator\", \"agency-visual-storyteller\", \"agency-tiktok-strategist\", \"agency-trend-researcher\", \"agency-api-tester\", \"agency-app-store-optimizer\", \"agency-xr-immersive-developer\", \"agency-growth-hacker\", \"agency-content-creator\", \"agency-reddit-community-builder\", \"agency-twitter-engager\", \"agency-whimsy-injector\", \"agency-xiaohongshu-specialist\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0024",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PRD Research Factory. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "prd-research-factory",
        "proposal-factory",
        "lead-lifecycle-manager",
        "feature-intake-router",
        "role-cso",
        "nlm-arxiv-slides",
        "sales-meeting-pipeline",
        "planning-meeting-sentinel",
        "role-sales",
        "ui-suite",
        "test-suite",
        "auto-research",
        "nlm-launch-kit",
        "docs-freshness-guardian",
        "sales-weekly-dashboard",
        "incident-lifecycle-manager",
        "tech-support-brain",
        "role-ceo",
        "ship",
        "nlm-slides",
        "nlm-strategy-hub",
        "plans",
        "nlm-discovery-lab",
        "bespin-news-digest",
        "knowledge-to-paper"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "prd-research-factory",
          "tokens": 648
        },
        {
          "skill_id": "proposal-factory",
          "tokens": 582
        },
        {
          "skill_id": "lead-lifecycle-manager",
          "tokens": 533
        },
        {
          "skill_id": "feature-intake-router",
          "tokens": 568
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        },
        {
          "skill_id": "nlm-arxiv-slides",
          "tokens": 3484
        },
        {
          "skill_id": "sales-meeting-pipeline",
          "tokens": 559
        },
        {
          "skill_id": "planning-meeting-sentinel",
          "tokens": 604
        },
        {
          "skill_id": "role-sales",
          "tokens": 1221
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "auto-research",
          "tokens": 2139
        },
        {
          "skill_id": "nlm-launch-kit",
          "tokens": 2368
        },
        {
          "skill_id": "docs-freshness-guardian",
          "tokens": 557
        },
        {
          "skill_id": "sales-weekly-dashboard",
          "tokens": 558
        },
        {
          "skill_id": "incident-lifecycle-manager",
          "tokens": 610
        },
        {
          "skill_id": "tech-support-brain",
          "tokens": 620
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "nlm-slides",
          "tokens": 1685
        },
        {
          "skill_id": "nlm-strategy-hub",
          "tokens": 2207
        },
        {
          "skill_id": "plans",
          "tokens": 2297
        },
        {
          "skill_id": "nlm-discovery-lab",
          "tokens": 2808
        },
        {
          "skill_id": "bespin-news-digest",
          "tokens": 3949
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PRD Research Factory. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"prd-research-factory\", \"proposal-factory\", \"lead-lifecycle-manager\", \"feature-intake-router\", \"role-cso\", \"nlm-arxiv-slides\", \"sales-meeting-pipeline\", \"planning-meeting-sentinel\", \"role-sales\", \"ui-suite\", \"test-suite\", \"auto-research\", \"nlm-launch-kit\", \"docs-freshness-guardian\", \"sales-weekly-dashboard\", \"incident-lifecycle-manager\", \"tech-support-brain\", \"role-ceo\", \"ship\", \"nlm-slides\", \"nlm-strategy-hub\", \"plans\", \"nlm-discovery-lab\", \"bespin-news-digest\", \"knowledge-to-paper\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0025",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: User Research. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-design-user-research",
        "kwp-engineering-testing-strategy",
        "kwp-design-accessibility-review",
        "kwp-design-design-system-management",
        "kwp-engineering-documentation",
        "kwp-design-ux-writing",
        "kwp-engineering-system-design",
        "kwp-design-design-critique",
        "kwp-operations-compliance-tracking",
        "kwp-engineering-code-review",
        "kwp-engineering-tech-debt",
        "kwp-engineering-incident-response",
        "kwp-operations-risk-assessment",
        "ux-expert",
        "kwp-productivity-task-management",
        "role-ux-designer",
        "design-architect",
        "agency-visual-storyteller",
        "problem-definition",
        "agency-ux-researcher",
        "kwp-operations-resource-planning",
        "kwp-human-resources-interview-prep",
        "full-stack-planner",
        "agency-experiment-tracker",
        "kwp-cowork-create-plugin"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-design-user-research",
          "tokens": 718
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "kwp-design-accessibility-review",
          "tokens": 679
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "kwp-engineering-documentation",
          "tokens": 654
        },
        {
          "skill_id": "kwp-design-ux-writing",
          "tokens": 675
        },
        {
          "skill_id": "kwp-engineering-system-design",
          "tokens": 756
        },
        {
          "skill_id": "kwp-design-design-critique",
          "tokens": 671
        },
        {
          "skill_id": "kwp-operations-compliance-tracking",
          "tokens": 686
        },
        {
          "skill_id": "kwp-engineering-code-review",
          "tokens": 610
        },
        {
          "skill_id": "kwp-engineering-tech-debt",
          "tokens": 648
        },
        {
          "skill_id": "kwp-engineering-incident-response",
          "tokens": 645
        },
        {
          "skill_id": "kwp-operations-risk-assessment",
          "tokens": 808
        },
        {
          "skill_id": "ux-expert",
          "tokens": 1210
        },
        {
          "skill_id": "kwp-productivity-task-management",
          "tokens": 776
        },
        {
          "skill_id": "role-ux-designer",
          "tokens": 1268
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "problem-definition",
          "tokens": 2601
        },
        {
          "skill_id": "agency-ux-researcher",
          "tokens": 3473
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "kwp-cowork-create-plugin",
          "tokens": 3214
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: User Research. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-design-user-research\", \"kwp-engineering-testing-strategy\", \"kwp-design-accessibility-review\", \"kwp-design-design-system-management\", \"kwp-engineering-documentation\", \"kwp-design-ux-writing\", \"kwp-engineering-system-design\", \"kwp-design-design-critique\", \"kwp-operations-compliance-tracking\", \"kwp-engineering-code-review\", \"kwp-engineering-tech-debt\", \"kwp-engineering-incident-response\", \"kwp-operations-risk-assessment\", \"ux-expert\", \"kwp-productivity-task-management\", \"role-ux-designer\", \"design-architect\", \"agency-visual-storyteller\", \"problem-definition\", \"agency-ux-researcher\", \"kwp-operations-resource-planning\", \"kwp-human-resources-interview-prep\", \"full-stack-planner\", \"agency-experiment-tracker\", \"kwp-cowork-create-plugin\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0026",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: ArchitectUX Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-ux-architect",
        "agency-infrastructure-maintainer",
        "agency-developer-advocate",
        "agency-agentic-identity-trust-architect",
        "agency-xr-cockpit-interaction-specialist",
        "agency-senior-developer",
        "agency-inclusive-visuals-specialist",
        "agency-visual-storyteller",
        "agency-ai-engineer",
        "agency-studio-producer",
        "agency-lsp-index-engineer",
        "agency-macos-spatial-metal-engineer",
        "agency-api-tester",
        "agency-legal-compliance-checker",
        "agency-xr-interface-architect",
        "agency-xr-immersive-developer",
        "agency-data-consolidation-agent",
        "agency-visionos-spatial-engineer",
        "agency-terminal-integration-specialist",
        "agency-behavioral-nudge-engine",
        "agency-instagram-curator",
        "agency-cultural-intelligence-strategist",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "agency-twitter-engager"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-ux-architect",
          "tokens": 3119
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "agency-agentic-identity-trust-architect",
          "tokens": 4533
        },
        {
          "skill_id": "agency-xr-cockpit-interaction-specialist",
          "tokens": 693
        },
        {
          "skill_id": "agency-senior-developer",
          "tokens": 1781
        },
        {
          "skill_id": "agency-inclusive-visuals-specialist",
          "tokens": 1931
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-studio-producer",
          "tokens": 2953
        },
        {
          "skill_id": "agency-lsp-index-engineer",
          "tokens": 3006
        },
        {
          "skill_id": "agency-macos-spatial-metal-engineer",
          "tokens": 3229
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-xr-interface-architect",
          "tokens": 657
        },
        {
          "skill_id": "agency-xr-immersive-developer",
          "tokens": 697
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        },
        {
          "skill_id": "agency-visionos-spatial-engineer",
          "tokens": 1091
        },
        {
          "skill_id": "agency-terminal-integration-specialist",
          "tokens": 1273
        },
        {
          "skill_id": "agency-behavioral-nudge-engine",
          "tokens": 1553
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-cultural-intelligence-strategist",
          "tokens": 1892
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: ArchitectUX Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-ux-architect\", \"agency-infrastructure-maintainer\", \"agency-developer-advocate\", \"agency-agentic-identity-trust-architect\", \"agency-xr-cockpit-interaction-specialist\", \"agency-senior-developer\", \"agency-inclusive-visuals-specialist\", \"agency-visual-storyteller\", \"agency-ai-engineer\", \"agency-studio-producer\", \"agency-lsp-index-engineer\", \"agency-macos-spatial-metal-engineer\", \"agency-api-tester\", \"agency-legal-compliance-checker\", \"agency-xr-interface-architect\", \"agency-xr-immersive-developer\", \"agency-data-consolidation-agent\", \"agency-visionos-spatial-engineer\", \"agency-terminal-integration-specialist\", \"agency-behavioral-nudge-engine\", \"agency-instagram-curator\", \"agency-cultural-intelligence-strategist\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"agency-twitter-engager\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0027",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: MCP Server Development Guide. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "anthropic-mcp-builder",
        "hf-trackio",
        "anthropic-claude-api",
        "skill-seekers",
        "public-apis",
        "kwp-data-interactive-dashboard-builder",
        "setup-doctor",
        "ecc-coding-standards",
        "prompt-transformer",
        "mission-control copy",
        "mission-control",
        "kwp-data-data-visualization",
        "ecc-continuous-learning",
        "agency-api-tester",
        "ecc-configure",
        "hf-evaluation",
        "planning-meeting-sentinel",
        "gws-chat",
        "kwp-engineering-system-design",
        "ci-quality-gate",
        "ppt-template-engine",
        "hf-endpoints",
        "gws-workspace",
        "docx-template-engine",
        "anthropic-frontend-design"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "anthropic-mcp-builder",
          "tokens": 2825
        },
        {
          "skill_id": "hf-trackio",
          "tokens": 1424
        },
        {
          "skill_id": "anthropic-claude-api",
          "tokens": 4709
        },
        {
          "skill_id": "skill-seekers",
          "tokens": 1568
        },
        {
          "skill_id": "public-apis",
          "tokens": 2246
        },
        {
          "skill_id": "kwp-data-interactive-dashboard-builder",
          "tokens": 2350
        },
        {
          "skill_id": "setup-doctor",
          "tokens": 2578
        },
        {
          "skill_id": "ecc-coding-standards",
          "tokens": 1050
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "kwp-data-data-visualization",
          "tokens": 3020
        },
        {
          "skill_id": "ecc-continuous-learning",
          "tokens": 3352
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "ecc-configure",
          "tokens": 3534
        },
        {
          "skill_id": "hf-evaluation",
          "tokens": 4549
        },
        {
          "skill_id": "planning-meeting-sentinel",
          "tokens": 604
        },
        {
          "skill_id": "gws-chat",
          "tokens": 751
        },
        {
          "skill_id": "kwp-engineering-system-design",
          "tokens": 756
        },
        {
          "skill_id": "ci-quality-gate",
          "tokens": 1054
        },
        {
          "skill_id": "ppt-template-engine",
          "tokens": 1101
        },
        {
          "skill_id": "hf-endpoints",
          "tokens": 1177
        },
        {
          "skill_id": "gws-workspace",
          "tokens": 1186
        },
        {
          "skill_id": "docx-template-engine",
          "tokens": 1193
        },
        {
          "skill_id": "anthropic-frontend-design",
          "tokens": 1312
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: MCP Server Development Guide. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-mcp-builder\", \"hf-trackio\", \"anthropic-claude-api\", \"skill-seekers\", \"public-apis\", \"kwp-data-interactive-dashboard-builder\", \"setup-doctor\", \"ecc-coding-standards\", \"prompt-transformer\", \"mission-control copy\", \"mission-control\", \"kwp-data-data-visualization\", \"ecc-continuous-learning\", \"agency-api-tester\", \"ecc-configure\", \"hf-evaluation\", \"planning-meeting-sentinel\", \"gws-chat\", \"kwp-engineering-system-design\", \"ci-quality-gate\", \"ppt-template-engine\", \"hf-endpoints\", \"gws-workspace\", \"docx-template-engine\", \"anthropic-frontend-design\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0028",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Sales Meeting Pipeline. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "sales-meeting-pipeline",
        "planning-meeting-sentinel",
        "lead-lifecycle-manager",
        "planning-weekly-pulse",
        "sales-weekly-dashboard",
        "prd-research-factory",
        "proposal-factory",
        "gws-workflows",
        "weekly-status-report",
        "kwp-sales-call-prep",
        "docs-freshness-guardian",
        "feature-intake-router",
        "knowledge-daily-aggregator",
        "kwp-sales-daily-briefing",
        "dependency-radar",
        "incident-lifecycle-manager",
        "tech-support-brain",
        "gws-drive",
        "gws-calendar",
        "ai-chief-of-staff",
        "ship",
        "decision-router",
        "kwp-marketing-content-creation",
        "anthropic-internal-comms",
        "gws-recipe-personalized-emails"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "sales-meeting-pipeline",
          "tokens": 559
        },
        {
          "skill_id": "planning-meeting-sentinel",
          "tokens": 604
        },
        {
          "skill_id": "lead-lifecycle-manager",
          "tokens": 533
        },
        {
          "skill_id": "planning-weekly-pulse",
          "tokens": 528
        },
        {
          "skill_id": "sales-weekly-dashboard",
          "tokens": 558
        },
        {
          "skill_id": "prd-research-factory",
          "tokens": 648
        },
        {
          "skill_id": "proposal-factory",
          "tokens": 582
        },
        {
          "skill_id": "gws-workflows",
          "tokens": 952
        },
        {
          "skill_id": "weekly-status-report",
          "tokens": 1282
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "docs-freshness-guardian",
          "tokens": 557
        },
        {
          "skill_id": "feature-intake-router",
          "tokens": 568
        },
        {
          "skill_id": "knowledge-daily-aggregator",
          "tokens": 1425
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "dependency-radar",
          "tokens": 531
        },
        {
          "skill_id": "incident-lifecycle-manager",
          "tokens": 610
        },
        {
          "skill_id": "tech-support-brain",
          "tokens": 620
        },
        {
          "skill_id": "gws-drive",
          "tokens": 997
        },
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "ai-chief-of-staff",
          "tokens": 1152
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "decision-router",
          "tokens": 1730
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "anthropic-internal-comms",
          "tokens": 522
        },
        {
          "skill_id": "gws-recipe-personalized-emails",
          "tokens": 641
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Sales Meeting Pipeline. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"sales-meeting-pipeline\", \"planning-meeting-sentinel\", \"lead-lifecycle-manager\", \"planning-weekly-pulse\", \"sales-weekly-dashboard\", \"prd-research-factory\", \"proposal-factory\", \"gws-workflows\", \"weekly-status-report\", \"kwp-sales-call-prep\", \"docs-freshness-guardian\", \"feature-intake-router\", \"knowledge-daily-aggregator\", \"kwp-sales-daily-briefing\", \"dependency-radar\", \"incident-lifecycle-manager\", \"tech-support-brain\", \"gws-drive\", \"gws-calendar\", \"ai-chief-of-staff\", \"ship\", \"decision-router\", \"kwp-marketing-content-creation\", \"anthropic-internal-comms\", \"gws-recipe-personalized-emails\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0029",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Roadmap Management Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-product-management-roadmap-management",
        "kwp-data-data-visualization",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-org-planning",
        "kwp-marketing-campaign-planning",
        "kwp-product-management-user-research-synthesis",
        "kwp-operations-change-management",
        "kwp-operations-resource-planning",
        "kwp-human-resources-interview-prep",
        "pm-execution",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-customer-support-escalation",
        "kwp-data-interactive-dashboard-builder",
        "agency-executive-summary-generator",
        "kwp-product-management-feature-spec",
        "kwp-customer-support-knowledge-management",
        "kwp-data-data-validation",
        "kwp-finance-financial-statements",
        "kwp-finance-close-management",
        "kwp-product-management-competitive-analysis",
        "kwp-legal-legal-risk-assessment",
        "kwp-design-system",
        "kwp-human-resources-people-analytics",
        "kwp-operations-vendor-management",
        "kwp-human-resources-recruiting-pipeline"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-product-management-roadmap-management",
          "tokens": 2467
        },
        {
          "skill_id": "kwp-data-data-visualization",
          "tokens": 3020
        },
        {
          "skill_id": "kwp-product-management-metrics-tracking",
          "tokens": 3606
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-marketing-campaign-planning",
          "tokens": 2803
        },
        {
          "skill_id": "kwp-product-management-user-research-synthesis",
          "tokens": 3071
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "pm-execution",
          "tokens": 1178
        },
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "kwp-customer-support-escalation",
          "tokens": 2329
        },
        {
          "skill_id": "kwp-data-interactive-dashboard-builder",
          "tokens": 2350
        },
        {
          "skill_id": "agency-executive-summary-generator",
          "tokens": 2465
        },
        {
          "skill_id": "kwp-product-management-feature-spec",
          "tokens": 2614
        },
        {
          "skill_id": "kwp-customer-support-knowledge-management",
          "tokens": 2624
        },
        {
          "skill_id": "kwp-data-data-validation",
          "tokens": 2650
        },
        {
          "skill_id": "kwp-finance-financial-statements",
          "tokens": 2768
        },
        {
          "skill_id": "kwp-finance-close-management",
          "tokens": 2782
        },
        {
          "skill_id": "kwp-product-management-competitive-analysis",
          "tokens": 2899
        },
        {
          "skill_id": "kwp-legal-legal-risk-assessment",
          "tokens": 3406
        },
        {
          "skill_id": "kwp-design-system",
          "tokens": 4960
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-recruiting-pipeline",
          "tokens": 578
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Roadmap Management Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-product-management-roadmap-management\", \"kwp-data-data-visualization\", \"kwp-product-management-metrics-tracking\", \"kwp-human-resources-org-planning\", \"kwp-marketing-campaign-planning\", \"kwp-product-management-user-research-synthesis\", \"kwp-operations-change-management\", \"kwp-operations-resource-planning\", \"kwp-human-resources-interview-prep\", \"pm-execution\", \"kwp-bio-research-single-cell-rna-qc\", \"kwp-customer-support-escalation\", \"kwp-data-interactive-dashboard-builder\", \"agency-executive-summary-generator\", \"kwp-product-management-feature-spec\", \"kwp-customer-support-knowledge-management\", \"kwp-data-data-validation\", \"kwp-finance-financial-statements\", \"kwp-finance-close-management\", \"kwp-product-management-competitive-analysis\", \"kwp-legal-legal-risk-assessment\", \"kwp-design-system\", \"kwp-human-resources-people-analytics\", \"kwp-operations-vendor-management\", \"kwp-human-resources-recruiting-pipeline\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0030",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Druckenmiller Strategy Synthesizer. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "trading-stanley-druckenmiller-investment",
        "trading-position-sizer",
        "trading-us-stock-analysis",
        "trading-market-environment-analysis",
        "trading-breadth-chart-analyst",
        "trading-trade-hypothesis-ideator",
        "trading-edge-candidate-agent",
        "trading-technical-analyst",
        "trading-finviz-screener",
        "trading-sector-analyst",
        "alphaear-deepear-lite",
        "alphaear-stock",
        "stock-csv-downloader",
        "weekly-stock-update",
        "daily-stock-check",
        "trading-uptrend-analyzer",
        "role-trading-expert",
        "trading-scenario-analyzer",
        "trading-market-top-detector",
        "trading-backtest-expert",
        "trading-us-market-bubble-detector",
        "trading-theme-detector",
        "alphaear-reporter",
        "role-ceo",
        "weekly-status-report"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "trading-stanley-druckenmiller-investment",
          "tokens": 2317
        },
        {
          "skill_id": "trading-position-sizer",
          "tokens": 2013
        },
        {
          "skill_id": "trading-us-stock-analysis",
          "tokens": 2926
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "trading-breadth-chart-analyst",
          "tokens": 3912
        },
        {
          "skill_id": "trading-trade-hypothesis-ideator",
          "tokens": 909
        },
        {
          "skill_id": "trading-edge-candidate-agent",
          "tokens": 1971
        },
        {
          "skill_id": "trading-technical-analyst",
          "tokens": 2868
        },
        {
          "skill_id": "trading-finviz-screener",
          "tokens": 3681
        },
        {
          "skill_id": "trading-sector-analyst",
          "tokens": 2657
        },
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        },
        {
          "skill_id": "alphaear-stock",
          "tokens": 820
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "trading-uptrend-analyzer",
          "tokens": 1780
        },
        {
          "skill_id": "role-trading-expert",
          "tokens": 1791
        },
        {
          "skill_id": "trading-scenario-analyzer",
          "tokens": 2068
        },
        {
          "skill_id": "trading-market-top-detector",
          "tokens": 2295
        },
        {
          "skill_id": "trading-backtest-expert",
          "tokens": 2754
        },
        {
          "skill_id": "trading-us-market-bubble-detector",
          "tokens": 3062
        },
        {
          "skill_id": "trading-theme-detector",
          "tokens": 3487
        },
        {
          "skill_id": "alphaear-reporter",
          "tokens": 810
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "weekly-status-report",
          "tokens": 1282
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Druckenmiller Strategy Synthesizer. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"trading-stanley-druckenmiller-investment\", \"trading-position-sizer\", \"trading-us-stock-analysis\", \"trading-market-environment-analysis\", \"trading-breadth-chart-analyst\", \"trading-trade-hypothesis-ideator\", \"trading-edge-candidate-agent\", \"trading-technical-analyst\", \"trading-finviz-screener\", \"trading-sector-analyst\", \"alphaear-deepear-lite\", \"alphaear-stock\", \"stock-csv-downloader\", \"weekly-stock-update\", \"daily-stock-check\", \"trading-uptrend-analyzer\", \"role-trading-expert\", \"trading-scenario-analyzer\", \"trading-market-top-detector\", \"trading-backtest-expert\", \"trading-us-market-bubble-detector\", \"trading-theme-detector\", \"alphaear-reporter\", \"role-ceo\", \"weekly-status-report\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0031",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-stock-sync. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "tab-stock-sync",
        "weekly-stock-update",
        "alphaear-stock",
        "stock-csv-downloader",
        "daily-stock-check",
        "tab-fundamental-sync",
        "tab-kiwoom",
        "trading-sector-analyst",
        "tab-genai-features",
        "tab-llm-agents",
        "tab-dualma-refresh",
        "tab-strategy-comparison",
        "tab-report-generate",
        "alphaear-search",
        "trading-technical-analyst",
        "today",
        "tab-dashboard-summary",
        "tab-news-fetch",
        "tab-market-breadth",
        "alphaear-deepear-lite",
        "alphaear-news",
        "trading-backtest-expert",
        "trading-us-stock-analysis",
        "trading-us-market-bubble-detector",
        "tab-hot-stock-discovery"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "tab-stock-sync",
          "tokens": 345
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "alphaear-stock",
          "tokens": 820
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "tab-fundamental-sync",
          "tokens": 331
        },
        {
          "skill_id": "tab-kiwoom",
          "tokens": 452
        },
        {
          "skill_id": "trading-sector-analyst",
          "tokens": 2657
        },
        {
          "skill_id": "tab-genai-features",
          "tokens": 354
        },
        {
          "skill_id": "tab-llm-agents",
          "tokens": 382
        },
        {
          "skill_id": "tab-dualma-refresh",
          "tokens": 391
        },
        {
          "skill_id": "tab-strategy-comparison",
          "tokens": 571
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "alphaear-search",
          "tokens": 672
        },
        {
          "skill_id": "trading-technical-analyst",
          "tokens": 2868
        },
        {
          "skill_id": "today",
          "tokens": 6812
        },
        {
          "skill_id": "tab-dashboard-summary",
          "tokens": 419
        },
        {
          "skill_id": "tab-news-fetch",
          "tokens": 474
        },
        {
          "skill_id": "tab-market-breadth",
          "tokens": 504
        },
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        },
        {
          "skill_id": "alphaear-news",
          "tokens": 873
        },
        {
          "skill_id": "trading-backtest-expert",
          "tokens": 2754
        },
        {
          "skill_id": "trading-us-stock-analysis",
          "tokens": 2926
        },
        {
          "skill_id": "trading-us-market-bubble-detector",
          "tokens": 3062
        },
        {
          "skill_id": "tab-hot-stock-discovery",
          "tokens": 359
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-stock-sync. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"tab-stock-sync\", \"weekly-stock-update\", \"alphaear-stock\", \"stock-csv-downloader\", \"daily-stock-check\", \"tab-fundamental-sync\", \"tab-kiwoom\", \"trading-sector-analyst\", \"tab-genai-features\", \"tab-llm-agents\", \"tab-dualma-refresh\", \"tab-strategy-comparison\", \"tab-report-generate\", \"alphaear-search\", \"trading-technical-analyst\", \"today\", \"tab-dashboard-summary\", \"tab-news-fetch\", \"tab-market-breadth\", \"alphaear-deepear-lite\", \"alphaear-news\", \"trading-backtest-expert\", \"trading-us-stock-analysis\", \"trading-us-market-bubble-detector\", \"tab-hot-stock-discovery\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0032",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PM Data Analytics. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "pm-data-analytics",
        "pm-product-discovery",
        "evals-skills",
        "pm-toolkit",
        "pm-marketing-growth",
        "pm-market-research",
        "pm-go-to-market",
        "pm-execution",
        "pm-product-strategy",
        "kwp-human-resources-people-analytics",
        "agency-data-analytics-reporter",
        "role-pm",
        "e2e-testing",
        "ai-workflow-integrator",
        "mission-control",
        "agency-test-results-analyzer",
        "agency-analytics-reporter",
        "mirofish-graph-explorer",
        "role-cso",
        "role-hr",
        "role-developer",
        "ui-suite",
        "scrapling",
        "role-data-scientist",
        "daily-stock-check"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "pm-data-analytics",
          "tokens": 956
        },
        {
          "skill_id": "pm-product-discovery",
          "tokens": 1198
        },
        {
          "skill_id": "evals-skills",
          "tokens": 1524
        },
        {
          "skill_id": "pm-toolkit",
          "tokens": 935
        },
        {
          "skill_id": "pm-marketing-growth",
          "tokens": 954
        },
        {
          "skill_id": "pm-market-research",
          "tokens": 966
        },
        {
          "skill_id": "pm-go-to-market",
          "tokens": 992
        },
        {
          "skill_id": "pm-execution",
          "tokens": 1178
        },
        {
          "skill_id": "pm-product-strategy",
          "tokens": 1271
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "agency-data-analytics-reporter",
          "tokens": 1003
        },
        {
          "skill_id": "role-pm",
          "tokens": 1304
        },
        {
          "skill_id": "e2e-testing",
          "tokens": 1484
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "agency-analytics-reporter",
          "tokens": 4157
        },
        {
          "skill_id": "mirofish-graph-explorer",
          "tokens": 1175
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        },
        {
          "skill_id": "role-hr",
          "tokens": 1350
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "scrapling",
          "tokens": 1675
        },
        {
          "skill_id": "role-data-scientist",
          "tokens": 1695
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PM Data Analytics. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"pm-data-analytics\", \"pm-product-discovery\", \"evals-skills\", \"pm-toolkit\", \"pm-marketing-growth\", \"pm-market-research\", \"pm-go-to-market\", \"pm-execution\", \"pm-product-strategy\", \"kwp-human-resources-people-analytics\", \"agency-data-analytics-reporter\", \"role-pm\", \"e2e-testing\", \"ai-workflow-integrator\", \"mission-control\", \"agency-test-results-analyzer\", \"agency-analytics-reporter\", \"mirofish-graph-explorer\", \"role-cso\", \"role-hr\", \"role-developer\", \"ui-suite\", \"scrapling\", \"role-data-scientist\", \"daily-stock-check\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0033",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Canned Responses Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-legal-canned-responses",
        "kwp-legal-compliance",
        "kwp-customer-support-response-drafting",
        "kwp-product-management-user-research-synthesis",
        "kwp-legal-contract-review",
        "kwp-legal-nda-triage",
        "kwp-human-resources-people-analytics",
        "kwp-operations-vendor-management",
        "kwp-human-resources-interview-prep",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-finance-reconciliation",
        "kwp-data-interactive-dashboard-builder",
        "kwp-data-data-exploration",
        "kwp-legal-meeting-briefing",
        "kwp-customer-support-customer-research",
        "kwp-data-data-validation",
        "kwp-finance-financial-statements",
        "kwp-data-statistical-analysis",
        "kwp-data-data-visualization",
        "kwp-legal-legal-risk-assessment",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-legal-canned-responses",
          "tokens": 3164
        },
        {
          "skill_id": "kwp-legal-compliance",
          "tokens": 3473
        },
        {
          "skill_id": "kwp-customer-support-response-drafting",
          "tokens": 2944
        },
        {
          "skill_id": "kwp-product-management-user-research-synthesis",
          "tokens": 3071
        },
        {
          "skill_id": "kwp-legal-contract-review",
          "tokens": 3137
        },
        {
          "skill_id": "kwp-legal-nda-triage",
          "tokens": 3139
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "kwp-finance-reconciliation",
          "tokens": 2287
        },
        {
          "skill_id": "kwp-data-interactive-dashboard-builder",
          "tokens": 2350
        },
        {
          "skill_id": "kwp-data-data-exploration",
          "tokens": 2358
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-customer-support-customer-research",
          "tokens": 2529
        },
        {
          "skill_id": "kwp-data-data-validation",
          "tokens": 2650
        },
        {
          "skill_id": "kwp-finance-financial-statements",
          "tokens": 2768
        },
        {
          "skill_id": "kwp-data-statistical-analysis",
          "tokens": 2865
        },
        {
          "skill_id": "kwp-data-data-visualization",
          "tokens": 3020
        },
        {
          "skill_id": "kwp-legal-legal-risk-assessment",
          "tokens": 3406
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-recruiting-pipeline",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Canned Responses Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-legal-canned-responses\", \"kwp-legal-compliance\", \"kwp-customer-support-response-drafting\", \"kwp-product-management-user-research-synthesis\", \"kwp-legal-contract-review\", \"kwp-legal-nda-triage\", \"kwp-human-resources-people-analytics\", \"kwp-operations-vendor-management\", \"kwp-human-resources-interview-prep\", \"kwp-bio-research-single-cell-rna-qc\", \"kwp-finance-reconciliation\", \"kwp-data-interactive-dashboard-builder\", \"kwp-data-data-exploration\", \"kwp-legal-meeting-briefing\", \"kwp-customer-support-customer-research\", \"kwp-data-data-validation\", \"kwp-finance-financial-statements\", \"kwp-data-statistical-analysis\", \"kwp-data-data-visualization\", \"kwp-legal-legal-risk-assessment\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-operations-process-optimization\", \"kwp-operations-change-management\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0034",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Agent Browser. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agent-browser",
        "scrapling",
        "demo-forge",
        "e2e-overhaul",
        "defuddle",
        "cursor-automations",
        "playwright-runner",
        "ui-suite",
        "notebooklm-studio",
        "ralph-loop",
        "meeting-digest",
        "visual-explainer",
        "sefo-orchestrator",
        "hf-hub",
        "frontend-expert",
        "e2e-testing",
        "weekly-stock-update",
        "presentation-strategist",
        "cognee",
        "hf-auto-finetune-planner",
        "md-to-notion",
        "deep-research-pipeline",
        "public-apis",
        "slack-agent",
        "agency-mobile-app-builder"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agent-browser",
          "tokens": 3431
        },
        {
          "skill_id": "scrapling",
          "tokens": 1675
        },
        {
          "skill_id": "demo-forge",
          "tokens": 2152
        },
        {
          "skill_id": "e2e-overhaul",
          "tokens": 2425
        },
        {
          "skill_id": "defuddle",
          "tokens": 2409
        },
        {
          "skill_id": "cursor-automations",
          "tokens": 2911
        },
        {
          "skill_id": "playwright-runner",
          "tokens": 2641
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "notebooklm-studio",
          "tokens": 1812
        },
        {
          "skill_id": "ralph-loop",
          "tokens": 1925
        },
        {
          "skill_id": "meeting-digest",
          "tokens": 4491
        },
        {
          "skill_id": "visual-explainer",
          "tokens": 5408
        },
        {
          "skill_id": "sefo-orchestrator",
          "tokens": 821
        },
        {
          "skill_id": "hf-hub",
          "tokens": 1151
        },
        {
          "skill_id": "frontend-expert",
          "tokens": 1234
        },
        {
          "skill_id": "e2e-testing",
          "tokens": 1484
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "presentation-strategist",
          "tokens": 1990
        },
        {
          "skill_id": "cognee",
          "tokens": 2070
        },
        {
          "skill_id": "hf-auto-finetune-planner",
          "tokens": 2091
        },
        {
          "skill_id": "md-to-notion",
          "tokens": 2096
        },
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        },
        {
          "skill_id": "public-apis",
          "tokens": 2246
        },
        {
          "skill_id": "slack-agent",
          "tokens": 2564
        },
        {
          "skill_id": "agency-mobile-app-builder",
          "tokens": 2595
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Agent Browser. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agent-browser\", \"scrapling\", \"demo-forge\", \"e2e-overhaul\", \"defuddle\", \"cursor-automations\", \"playwright-runner\", \"ui-suite\", \"notebooklm-studio\", \"ralph-loop\", \"meeting-digest\", \"visual-explainer\", \"sefo-orchestrator\", \"hf-hub\", \"frontend-expert\", \"e2e-testing\", \"weekly-stock-update\", \"presentation-strategist\", \"cognee\", \"hf-auto-finetune-planner\", \"md-to-notion\", \"deep-research-pipeline\", \"public-apis\", \"slack-agent\", \"agency-mobile-app-builder\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0035",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Hugging Face Repos. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "hf-repos",
        "hf-buckets",
        "hf-discussions",
        "hf-models",
        "hf-collections",
        "hf-hub",
        "hf-cli",
        "hf-endpoints",
        "hf-papers",
        "hf-spaces",
        "gws-sheets",
        "gws-drive",
        "hf-datasets",
        "gws-docs",
        "hf-dataset-viewer",
        "gws-gmail",
        "anthropic-pdf",
        "skill-optimizer",
        "gws-recipe-doc-from-template",
        "domain-commit",
        "gws-calendar",
        "autoskill-merger",
        "transcribee",
        "prompt-transformer",
        "test-suite"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "hf-repos",
          "tokens": 1106
        },
        {
          "skill_id": "hf-buckets",
          "tokens": 1334
        },
        {
          "skill_id": "hf-discussions",
          "tokens": 1118
        },
        {
          "skill_id": "hf-models",
          "tokens": 1212
        },
        {
          "skill_id": "hf-collections",
          "tokens": 1239
        },
        {
          "skill_id": "hf-hub",
          "tokens": 1151
        },
        {
          "skill_id": "hf-cli",
          "tokens": 2275
        },
        {
          "skill_id": "hf-endpoints",
          "tokens": 1177
        },
        {
          "skill_id": "hf-papers",
          "tokens": 817
        },
        {
          "skill_id": "hf-spaces",
          "tokens": 930
        },
        {
          "skill_id": "gws-sheets",
          "tokens": 918
        },
        {
          "skill_id": "gws-drive",
          "tokens": 997
        },
        {
          "skill_id": "hf-datasets",
          "tokens": 3241
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "hf-dataset-viewer",
          "tokens": 1387
        },
        {
          "skill_id": "gws-gmail",
          "tokens": 1796
        },
        {
          "skill_id": "anthropic-pdf",
          "tokens": 2219
        },
        {
          "skill_id": "skill-optimizer",
          "tokens": 2940
        },
        {
          "skill_id": "gws-recipe-doc-from-template",
          "tokens": 743
        },
        {
          "skill_id": "domain-commit",
          "tokens": 915
        },
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "autoskill-merger",
          "tokens": 1142
        },
        {
          "skill_id": "transcribee",
          "tokens": 1255
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Hugging Face Repos. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"hf-repos\", \"hf-buckets\", \"hf-discussions\", \"hf-models\", \"hf-collections\", \"hf-hub\", \"hf-cli\", \"hf-endpoints\", \"hf-papers\", \"hf-spaces\", \"gws-sheets\", \"gws-drive\", \"hf-datasets\", \"gws-docs\", \"hf-dataset-viewer\", \"gws-gmail\", \"anthropic-pdf\", \"skill-optimizer\", \"gws-recipe-doc-from-template\", \"domain-commit\", \"gws-calendar\", \"autoskill-merger\", \"transcribee\", \"prompt-transformer\", \"test-suite\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0036",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Infrastructure Runbook Generator. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "infra-runbook-generator",
        "helm-validator",
        "sre-devops-expert",
        "docs-tutor-setup",
        "kwp-engineering-documentation",
        "infra-drift-detector",
        "role-developer",
        "k8s-manifest-validator",
        "terraform-reviewer",
        "ppt-template-engine",
        "docx-template-engine",
        "office-template-enforcer",
        "iac-review-agent",
        "skill-guide-generator",
        "full-stack-planner",
        "onboarding-accelerator",
        "incident-to-improvement",
        "role-cto",
        "prompt-transformer",
        "service-health-doctor",
        "cognee",
        "demo-forge",
        "kwp-engineering-incident-response",
        "role-ceo",
        "role-sales"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "infra-runbook-generator",
          "tokens": 1623
        },
        {
          "skill_id": "helm-validator",
          "tokens": 1398
        },
        {
          "skill_id": "sre-devops-expert",
          "tokens": 1450
        },
        {
          "skill_id": "docs-tutor-setup",
          "tokens": 2288
        },
        {
          "skill_id": "kwp-engineering-documentation",
          "tokens": 654
        },
        {
          "skill_id": "infra-drift-detector",
          "tokens": 1315
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "k8s-manifest-validator",
          "tokens": 1431
        },
        {
          "skill_id": "terraform-reviewer",
          "tokens": 1457
        },
        {
          "skill_id": "ppt-template-engine",
          "tokens": 1101
        },
        {
          "skill_id": "docx-template-engine",
          "tokens": 1193
        },
        {
          "skill_id": "office-template-enforcer",
          "tokens": 1273
        },
        {
          "skill_id": "iac-review-agent",
          "tokens": 1312
        },
        {
          "skill_id": "skill-guide-generator",
          "tokens": 1858
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        },
        {
          "skill_id": "onboarding-accelerator",
          "tokens": 2419
        },
        {
          "skill_id": "incident-to-improvement",
          "tokens": 2624
        },
        {
          "skill_id": "role-cto",
          "tokens": 1413
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "service-health-doctor",
          "tokens": 1728
        },
        {
          "skill_id": "cognee",
          "tokens": 2070
        },
        {
          "skill_id": "demo-forge",
          "tokens": 2152
        },
        {
          "skill_id": "kwp-engineering-incident-response",
          "tokens": 645
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "role-sales",
          "tokens": 1221
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Infrastructure Runbook Generator. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"infra-runbook-generator\", \"helm-validator\", \"sre-devops-expert\", \"docs-tutor-setup\", \"kwp-engineering-documentation\", \"infra-drift-detector\", \"role-developer\", \"k8s-manifest-validator\", \"terraform-reviewer\", \"ppt-template-engine\", \"docx-template-engine\", \"office-template-enforcer\", \"iac-review-agent\", \"skill-guide-generator\", \"full-stack-planner\", \"onboarding-accelerator\", \"incident-to-improvement\", \"role-cto\", \"prompt-transformer\", \"service-health-doctor\", \"cognee\", \"demo-forge\", \"kwp-engineering-incident-response\", \"role-ceo\", \"role-sales\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0037",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AI Platform FE - FSD 개발 및 마이그레이션. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "fsd-development",
        "design-review",
        "figma-to-tds",
        "implement-screen",
        "fe-pipeline",
        "swagger-api-doc-generator",
        "github-workflow-automation",
        "overlay-layout-patterns",
        "screen-description",
        "notion-docs-sync",
        "anthropic-frontend-design",
        "design-architect",
        "tab-genai-features",
        "tab-llm-agents",
        "tab-screening",
        "planning-weekly-pulse",
        "dependency-radar",
        "lead-lifecycle-manager",
        "docs-freshness-guardian",
        "sales-weekly-dashboard",
        "sales-meeting-pipeline",
        "feature-intake-router",
        "proposal-factory",
        "planning-meeting-sentinel",
        "incident-lifecycle-manager"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "fsd-development",
          "tokens": 1234
        },
        {
          "skill_id": "design-review",
          "tokens": 947
        },
        {
          "skill_id": "figma-to-tds",
          "tokens": 987
        },
        {
          "skill_id": "implement-screen",
          "tokens": 996
        },
        {
          "skill_id": "fe-pipeline",
          "tokens": 2354
        },
        {
          "skill_id": "swagger-api-doc-generator",
          "tokens": 784
        },
        {
          "skill_id": "github-workflow-automation",
          "tokens": 971
        },
        {
          "skill_id": "overlay-layout-patterns",
          "tokens": 1047
        },
        {
          "skill_id": "screen-description",
          "tokens": 836
        },
        {
          "skill_id": "notion-docs-sync",
          "tokens": 857
        },
        {
          "skill_id": "anthropic-frontend-design",
          "tokens": 1312
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "tab-genai-features",
          "tokens": 354
        },
        {
          "skill_id": "tab-llm-agents",
          "tokens": 382
        },
        {
          "skill_id": "tab-screening",
          "tokens": 410
        },
        {
          "skill_id": "planning-weekly-pulse",
          "tokens": 528
        },
        {
          "skill_id": "dependency-radar",
          "tokens": 531
        },
        {
          "skill_id": "lead-lifecycle-manager",
          "tokens": 533
        },
        {
          "skill_id": "docs-freshness-guardian",
          "tokens": 557
        },
        {
          "skill_id": "sales-weekly-dashboard",
          "tokens": 558
        },
        {
          "skill_id": "sales-meeting-pipeline",
          "tokens": 559
        },
        {
          "skill_id": "feature-intake-router",
          "tokens": 568
        },
        {
          "skill_id": "proposal-factory",
          "tokens": 582
        },
        {
          "skill_id": "planning-meeting-sentinel",
          "tokens": 604
        },
        {
          "skill_id": "incident-lifecycle-manager",
          "tokens": 610
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AI Platform FE - FSD 개발 및 마이그레이션. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"fsd-development\", \"design-review\", \"figma-to-tds\", \"implement-screen\", \"fe-pipeline\", \"swagger-api-doc-generator\", \"github-workflow-automation\", \"overlay-layout-patterns\", \"screen-description\", \"notion-docs-sync\", \"anthropic-frontend-design\", \"design-architect\", \"tab-genai-features\", \"tab-llm-agents\", \"tab-screening\", \"planning-weekly-pulse\", \"dependency-radar\", \"lead-lifecycle-manager\", \"docs-freshness-guardian\", \"sales-weekly-dashboard\", \"sales-meeting-pipeline\", \"feature-intake-router\", \"proposal-factory\", \"planning-meeting-sentinel\", \"incident-lifecycle-manager\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0038",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Call Prep. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-common-room-call-prep",
        "kwp-common-room-account-research",
        "kwp-common-room-weekly-prep-brief",
        "kwp-common-room-prospect",
        "kwp-common-room-compose-outreach",
        "kwp-sales-call-prep",
        "kwp-common-room-contact-research",
        "kwp-legal-meeting-briefing",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-human-resources-employee-handbook",
        "kwp-operations-resource-planning",
        "kwp-sales-daily-briefing",
        "kwp-sales-account-research",
        "kwp-finance-journal-entry-prep",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-sales-draft-outreach",
        "kwp-operations-vendor-management",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-interview-prep",
        "kwp-apollo-enrich-lead",
        "kwp-apollo-prospect",
        "kwp-brand-voice-discover-brand",
        "kwp-customer-support-escalation",
        "kwp-customer-support-knowledge-management",
        "kwp-data-data-validation"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-common-room-call-prep",
          "tokens": 1729
        },
        {
          "skill_id": "kwp-common-room-account-research",
          "tokens": 1784
        },
        {
          "skill_id": "kwp-common-room-weekly-prep-brief",
          "tokens": 1607
        },
        {
          "skill_id": "kwp-common-room-prospect",
          "tokens": 1803
        },
        {
          "skill_id": "kwp-common-room-compose-outreach",
          "tokens": 1627
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-common-room-contact-research",
          "tokens": 1611
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-human-resources-employee-handbook",
          "tokens": 629
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "kwp-finance-journal-entry-prep",
          "tokens": 2215
        },
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "kwp-sales-draft-outreach",
          "tokens": 2782
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "kwp-apollo-enrich-lead",
          "tokens": 837
        },
        {
          "skill_id": "kwp-apollo-prospect",
          "tokens": 1006
        },
        {
          "skill_id": "kwp-brand-voice-discover-brand",
          "tokens": 1897
        },
        {
          "skill_id": "kwp-customer-support-escalation",
          "tokens": 2329
        },
        {
          "skill_id": "kwp-customer-support-knowledge-management",
          "tokens": 2624
        },
        {
          "skill_id": "kwp-data-data-validation",
          "tokens": 2650
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Call Prep. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-common-room-call-prep\", \"kwp-common-room-account-research\", \"kwp-common-room-weekly-prep-brief\", \"kwp-common-room-prospect\", \"kwp-common-room-compose-outreach\", \"kwp-sales-call-prep\", \"kwp-common-room-contact-research\", \"kwp-legal-meeting-briefing\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-human-resources-employee-handbook\", \"kwp-operations-resource-planning\", \"kwp-sales-daily-briefing\", \"kwp-sales-account-research\", \"kwp-finance-journal-entry-prep\", \"kwp-bio-research-single-cell-rna-qc\", \"kwp-sales-draft-outreach\", \"kwp-operations-vendor-management\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-interview-prep\", \"kwp-apollo-enrich-lead\", \"kwp-apollo-prospect\", \"kwp-brand-voice-discover-brand\", \"kwp-customer-support-escalation\", \"kwp-customer-support-knowledge-management\", \"kwp-data-data-validation\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0039",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Cultural Intelligence Strategist. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-cultural-intelligence-strategist",
        "agency-behavioral-nudge-engine",
        "agency-content-creator",
        "agency-visual-storyteller",
        "agency-tiktok-strategist",
        "agency-whimsy-injector",
        "agency-workflow-optimizer",
        "agency-trend-researcher",
        "agency-lsp-index-engineer",
        "agency-zhihu-strategist",
        "agency-api-tester",
        "agency-legal-compliance-checker",
        "agency-tool-evaluator",
        "agency-xr-interface-architect",
        "agency-xr-cockpit-interaction-specialist",
        "agency-data-consolidation-agent",
        "agency-growth-hacker",
        "agency-terminal-integration-specialist",
        "agency-instagram-curator",
        "agency-inclusive-visuals-specialist",
        "agency-social-media-strategist",
        "agency-reddit-community-builder",
        "agency-twitter-engager",
        "agency-autonomous-optimization-architect",
        "agency-xiaohongshu-specialist"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-cultural-intelligence-strategist",
          "tokens": 1892
        },
        {
          "skill_id": "agency-behavioral-nudge-engine",
          "tokens": 1553
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "agency-lsp-index-engineer",
          "tokens": 3006
        },
        {
          "skill_id": "agency-zhihu-strategist",
          "tokens": 3239
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-tool-evaluator",
          "tokens": 4549
        },
        {
          "skill_id": "agency-xr-interface-architect",
          "tokens": 657
        },
        {
          "skill_id": "agency-xr-cockpit-interaction-specialist",
          "tokens": 693
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-terminal-integration-specialist",
          "tokens": 1273
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-inclusive-visuals-specialist",
          "tokens": 1931
        },
        {
          "skill_id": "agency-social-media-strategist",
          "tokens": 2084
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-autonomous-optimization-architect",
          "tokens": 2220
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Cultural Intelligence Strategist. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-cultural-intelligence-strategist\", \"agency-behavioral-nudge-engine\", \"agency-content-creator\", \"agency-visual-storyteller\", \"agency-tiktok-strategist\", \"agency-whimsy-injector\", \"agency-workflow-optimizer\", \"agency-trend-researcher\", \"agency-lsp-index-engineer\", \"agency-zhihu-strategist\", \"agency-api-tester\", \"agency-legal-compliance-checker\", \"agency-tool-evaluator\", \"agency-xr-interface-architect\", \"agency-xr-cockpit-interaction-specialist\", \"agency-data-consolidation-agent\", \"agency-growth-hacker\", \"agency-terminal-integration-specialist\", \"agency-instagram-curator\", \"agency-inclusive-visuals-specialist\", \"agency-social-media-strategist\", \"agency-reddit-community-builder\", \"agency-twitter-engager\", \"agency-autonomous-optimization-architect\", \"agency-xiaohongshu-specialist\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0040",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Write internal communications using company formats. Use for 3P updates (Progress, Plans, Problems), company newsletters, FAQ responses, status reports, leadership updates, project updates, incident r…. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "anthropic-internal-comms",
        "agency-executive-summary-generator",
        "pm-go-to-market",
        "kwp-product-management-stakeholder-comms",
        "sales-meeting-pipeline",
        "pm-toolkit",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "agency-twitter-engager",
        "slack-agent",
        "agency-xiaohongshu-specialist",
        "agency-project-shepherd",
        "agency-developer-advocate",
        "kwp-engineering-testing-strategy",
        "gws-recipe-audit-sharing",
        "gws-docs",
        "kwp-productivity-task-management",
        "pm-marketing-growth",
        "anthropic-theme-factory",
        "weekly-status-report",
        "agency-instagram-curator",
        "kwp-product-management-roadmap-management",
        "agency-rapid-prototyper",
        "kwp-marketing-content-creation",
        "problem-definition"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "anthropic-internal-comms",
          "tokens": 522
        },
        {
          "skill_id": "agency-executive-summary-generator",
          "tokens": 2465
        },
        {
          "skill_id": "pm-go-to-market",
          "tokens": 992
        },
        {
          "skill_id": "kwp-product-management-stakeholder-comms",
          "tokens": 2952
        },
        {
          "skill_id": "sales-meeting-pipeline",
          "tokens": 559
        },
        {
          "skill_id": "pm-toolkit",
          "tokens": 935
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "slack-agent",
          "tokens": 2564
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "gws-recipe-audit-sharing",
          "tokens": 645
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "kwp-productivity-task-management",
          "tokens": 776
        },
        {
          "skill_id": "pm-marketing-growth",
          "tokens": 954
        },
        {
          "skill_id": "anthropic-theme-factory",
          "tokens": 990
        },
        {
          "skill_id": "weekly-status-report",
          "tokens": 1282
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "kwp-product-management-roadmap-management",
          "tokens": 2467
        },
        {
          "skill_id": "agency-rapid-prototyper",
          "tokens": 2496
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "problem-definition",
          "tokens": 2601
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Write internal communications using company formats. Use for 3P updates (Progress, Plans, Problems), company newsletters, FAQ responses, status reports, leadership updates, project updates, incident r…. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-internal-comms\", \"agency-executive-summary-generator\", \"pm-go-to-market\", \"kwp-product-management-stakeholder-comms\", \"sales-meeting-pipeline\", \"pm-toolkit\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"agency-twitter-engager\", \"slack-agent\", \"agency-xiaohongshu-specialist\", \"agency-project-shepherd\", \"agency-developer-advocate\", \"kwp-engineering-testing-strategy\", \"gws-recipe-audit-sharing\", \"gws-docs\", \"kwp-productivity-task-management\", \"pm-marketing-growth\", \"anthropic-theme-factory\", \"weekly-status-report\", \"agency-instagram-curator\", \"kwp-product-management-roadmap-management\", \"agency-rapid-prototyper\", \"kwp-marketing-content-creation\", \"problem-definition\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0041",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Contact Research. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-common-room-contact-research",
        "kwp-common-room-compose-outreach",
        "kwp-common-room-account-research",
        "kwp-common-room-prospect",
        "kwp-sales-account-research",
        "kwp-apollo-enrich-lead",
        "kwp-sales-draft-outreach",
        "kwp-common-room-weekly-prep-brief",
        "kwp-common-room-call-prep",
        "kwp-marketing-content-creation",
        "mission-control copy",
        "kwp-data-data-context-extractor",
        "kwp-sales-call-prep",
        "kwp-bio-research-single-cell-rna-qc",
        "mission-control",
        "kwp-legal-meeting-briefing",
        "kwp-customer-support-customer-research",
        "kwp-product-management-user-research-synthesis",
        "kwp-legal-canned-responses",
        "kwp-human-resources-people-analytics",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-operations-resource-planning",
        "kwp-sales-daily-briefing",
        "kwp-enterprise-search-search-strategy"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-common-room-contact-research",
          "tokens": 1611
        },
        {
          "skill_id": "kwp-common-room-compose-outreach",
          "tokens": 1627
        },
        {
          "skill_id": "kwp-common-room-account-research",
          "tokens": 1784
        },
        {
          "skill_id": "kwp-common-room-prospect",
          "tokens": 1803
        },
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "kwp-apollo-enrich-lead",
          "tokens": 837
        },
        {
          "skill_id": "kwp-sales-draft-outreach",
          "tokens": 2782
        },
        {
          "skill_id": "kwp-common-room-weekly-prep-brief",
          "tokens": 1607
        },
        {
          "skill_id": "kwp-common-room-call-prep",
          "tokens": 1729
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "kwp-data-data-context-extractor",
          "tokens": 1883
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-customer-support-customer-research",
          "tokens": 2529
        },
        {
          "skill_id": "kwp-product-management-user-research-synthesis",
          "tokens": 3071
        },
        {
          "skill_id": "kwp-legal-canned-responses",
          "tokens": 3164
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "kwp-enterprise-search-search-strategy",
          "tokens": 2009
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Contact Research. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-common-room-contact-research\", \"kwp-common-room-compose-outreach\", \"kwp-common-room-account-research\", \"kwp-common-room-prospect\", \"kwp-sales-account-research\", \"kwp-apollo-enrich-lead\", \"kwp-sales-draft-outreach\", \"kwp-common-room-weekly-prep-brief\", \"kwp-common-room-call-prep\", \"kwp-marketing-content-creation\", \"mission-control copy\", \"kwp-data-data-context-extractor\", \"kwp-sales-call-prep\", \"kwp-bio-research-single-cell-rna-qc\", \"mission-control\", \"kwp-legal-meeting-briefing\", \"kwp-customer-support-customer-research\", \"kwp-product-management-user-research-synthesis\", \"kwp-legal-canned-responses\", \"kwp-human-resources-people-analytics\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-operations-resource-planning\", \"kwp-sales-daily-briefing\", \"kwp-enterprise-search-search-strategy\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0042",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Memory Management. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-productivity-memory-management",
        "mission-control",
        "kwp-cowork-cowork-plugin-customizer",
        "kwp-customer-support-knowledge-management",
        "kwp-human-resources-people-analytics",
        "kwp-operations-change-management",
        "kwp-apollo-enrich-lead",
        "kwp-apollo-prospect",
        "mission-control copy",
        "kwp-enterprise-search-search-strategy",
        "context-engineer",
        "kwp-customer-support-escalation",
        "kwp-data-interactive-dashboard-builder",
        "kwp-product-management-roadmap-management",
        "kwp-customer-support-customer-research",
        "kwp-cowork-plugin-customizer",
        "kwp-customer-support-response-drafting",
        "kwp-cowork-create-cowork-plugin",
        "kwp-data-data-visualization",
        "kwp-legal-nda-triage",
        "kwp-legal-canned-responses",
        "kwp-legal-compliance",
        "kwp-marketing-performance-analytics",
        "kwp-finance-audit-support",
        "kwp-design-system"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-productivity-memory-management",
          "tokens": 2427
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "kwp-cowork-cowork-plugin-customizer",
          "tokens": 2437
        },
        {
          "skill_id": "kwp-customer-support-knowledge-management",
          "tokens": 2624
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "kwp-apollo-enrich-lead",
          "tokens": 837
        },
        {
          "skill_id": "kwp-apollo-prospect",
          "tokens": 1006
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "kwp-enterprise-search-search-strategy",
          "tokens": 2009
        },
        {
          "skill_id": "context-engineer",
          "tokens": 2324
        },
        {
          "skill_id": "kwp-customer-support-escalation",
          "tokens": 2329
        },
        {
          "skill_id": "kwp-data-interactive-dashboard-builder",
          "tokens": 2350
        },
        {
          "skill_id": "kwp-product-management-roadmap-management",
          "tokens": 2467
        },
        {
          "skill_id": "kwp-customer-support-customer-research",
          "tokens": 2529
        },
        {
          "skill_id": "kwp-cowork-plugin-customizer",
          "tokens": 2586
        },
        {
          "skill_id": "kwp-customer-support-response-drafting",
          "tokens": 2944
        },
        {
          "skill_id": "kwp-cowork-create-cowork-plugin",
          "tokens": 3017
        },
        {
          "skill_id": "kwp-data-data-visualization",
          "tokens": 3020
        },
        {
          "skill_id": "kwp-legal-nda-triage",
          "tokens": 3139
        },
        {
          "skill_id": "kwp-legal-canned-responses",
          "tokens": 3164
        },
        {
          "skill_id": "kwp-legal-compliance",
          "tokens": 3473
        },
        {
          "skill_id": "kwp-marketing-performance-analytics",
          "tokens": 3966
        },
        {
          "skill_id": "kwp-finance-audit-support",
          "tokens": 4426
        },
        {
          "skill_id": "kwp-design-system",
          "tokens": 4960
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Memory Management. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-productivity-memory-management\", \"mission-control\", \"kwp-cowork-cowork-plugin-customizer\", \"kwp-customer-support-knowledge-management\", \"kwp-human-resources-people-analytics\", \"kwp-operations-change-management\", \"kwp-apollo-enrich-lead\", \"kwp-apollo-prospect\", \"mission-control copy\", \"kwp-enterprise-search-search-strategy\", \"context-engineer\", \"kwp-customer-support-escalation\", \"kwp-data-interactive-dashboard-builder\", \"kwp-product-management-roadmap-management\", \"kwp-customer-support-customer-research\", \"kwp-cowork-plugin-customizer\", \"kwp-customer-support-response-drafting\", \"kwp-cowork-create-cowork-plugin\", \"kwp-data-data-visualization\", \"kwp-legal-nda-triage\", \"kwp-legal-canned-responses\", \"kwp-legal-compliance\", \"kwp-marketing-performance-analytics\", \"kwp-finance-audit-support\", \"kwp-design-system\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0043",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-technical-analysis. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "tab-technical-analysis",
        "tab-turtle-refresh",
        "tab-bollinger-refresh",
        "tab-market-breadth",
        "tab-strategy-comparison",
        "tab-glossary",
        "tab-stock-sync",
        "tab-genai-features",
        "tab-dualma-refresh",
        "tab-screening",
        "tab-dashboard-summary",
        "tab-news-fetch",
        "tab-market-environment",
        "tab-report-generate",
        "weekly-stock-update",
        "daily-stock-check",
        "morning-ship",
        "tab-event-detect",
        "tab-fundamental-sync",
        "tab-llm-agents",
        "tab-patterns",
        "daily-strategy-post",
        "tab-hot-stock-discovery",
        "tab-analysis-run",
        "email-research-dispatcher"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "tab-technical-analysis",
          "tokens": 414
        },
        {
          "skill_id": "tab-turtle-refresh",
          "tokens": 371
        },
        {
          "skill_id": "tab-bollinger-refresh",
          "tokens": 383
        },
        {
          "skill_id": "tab-market-breadth",
          "tokens": 504
        },
        {
          "skill_id": "tab-strategy-comparison",
          "tokens": 571
        },
        {
          "skill_id": "tab-glossary",
          "tokens": 538
        },
        {
          "skill_id": "tab-stock-sync",
          "tokens": 345
        },
        {
          "skill_id": "tab-genai-features",
          "tokens": 354
        },
        {
          "skill_id": "tab-dualma-refresh",
          "tokens": 391
        },
        {
          "skill_id": "tab-screening",
          "tokens": 410
        },
        {
          "skill_id": "tab-dashboard-summary",
          "tokens": 419
        },
        {
          "skill_id": "tab-news-fetch",
          "tokens": 474
        },
        {
          "skill_id": "tab-market-environment",
          "tokens": 566
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "morning-ship",
          "tokens": 2943
        },
        {
          "skill_id": "tab-event-detect",
          "tokens": 319
        },
        {
          "skill_id": "tab-fundamental-sync",
          "tokens": 331
        },
        {
          "skill_id": "tab-llm-agents",
          "tokens": 382
        },
        {
          "skill_id": "tab-patterns",
          "tokens": 509
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "tab-hot-stock-discovery",
          "tokens": 359
        },
        {
          "skill_id": "tab-analysis-run",
          "tokens": 404
        },
        {
          "skill_id": "email-research-dispatcher",
          "tokens": 703
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-technical-analysis. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"tab-technical-analysis\", \"tab-turtle-refresh\", \"tab-bollinger-refresh\", \"tab-market-breadth\", \"tab-strategy-comparison\", \"tab-glossary\", \"tab-stock-sync\", \"tab-genai-features\", \"tab-dualma-refresh\", \"tab-screening\", \"tab-dashboard-summary\", \"tab-news-fetch\", \"tab-market-environment\", \"tab-report-generate\", \"weekly-stock-update\", \"daily-stock-check\", \"morning-ship\", \"tab-event-detect\", \"tab-fundamental-sync\", \"tab-llm-agents\", \"tab-patterns\", \"daily-strategy-post\", \"tab-hot-stock-discovery\", \"tab-analysis-run\", \"email-research-dispatcher\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0044",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PM Toolkit. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "pm-toolkit",
        "pm-go-to-market",
        "i18n-sync",
        "pm-marketing-growth",
        "pm-data-analytics",
        "pr-review-captain",
        "proposal-factory",
        "pm-product-discovery",
        "compliance-governance",
        "prompt-transformer",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "docs-tutor-setup",
        "agency-xiaohongshu-specialist",
        "skill-optimizer",
        "agency-legal-compliance-checker",
        "anthropic-doc-coauthoring",
        "gws-recipe-audit-sharing",
        "sefo-governance",
        "pm-market-research",
        "gws-calendar",
        "ux-expert",
        "role-developer",
        "evals-skills",
        "skill-guide-generator"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "pm-toolkit",
          "tokens": 935
        },
        {
          "skill_id": "pm-go-to-market",
          "tokens": 992
        },
        {
          "skill_id": "i18n-sync",
          "tokens": 1282
        },
        {
          "skill_id": "pm-marketing-growth",
          "tokens": 954
        },
        {
          "skill_id": "pm-data-analytics",
          "tokens": 956
        },
        {
          "skill_id": "pr-review-captain",
          "tokens": 1496
        },
        {
          "skill_id": "proposal-factory",
          "tokens": 582
        },
        {
          "skill_id": "pm-product-discovery",
          "tokens": 1198
        },
        {
          "skill_id": "compliance-governance",
          "tokens": 1501
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "docs-tutor-setup",
          "tokens": 2288
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "skill-optimizer",
          "tokens": 2940
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "anthropic-doc-coauthoring",
          "tokens": 4138
        },
        {
          "skill_id": "gws-recipe-audit-sharing",
          "tokens": 645
        },
        {
          "skill_id": "sefo-governance",
          "tokens": 837
        },
        {
          "skill_id": "pm-market-research",
          "tokens": 966
        },
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "ux-expert",
          "tokens": 1210
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "evals-skills",
          "tokens": 1524
        },
        {
          "skill_id": "skill-guide-generator",
          "tokens": 1858
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PM Toolkit. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"pm-toolkit\", \"pm-go-to-market\", \"i18n-sync\", \"pm-marketing-growth\", \"pm-data-analytics\", \"pr-review-captain\", \"proposal-factory\", \"pm-product-discovery\", \"compliance-governance\", \"prompt-transformer\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"docs-tutor-setup\", \"agency-xiaohongshu-specialist\", \"skill-optimizer\", \"agency-legal-compliance-checker\", \"anthropic-doc-coauthoring\", \"gws-recipe-audit-sharing\", \"sefo-governance\", \"pm-market-research\", \"gws-calendar\", \"ux-expert\", \"role-developer\", \"evals-skills\", \"skill-guide-generator\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0045",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-strategy-comparison. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "tab-strategy-comparison",
        "tab-dualma-refresh",
        "tab-turtle-refresh",
        "tab-bollinger-refresh",
        "daily-stock-check",
        "tab-llm-agents",
        "tab-report-generate",
        "trading-backtest-expert",
        "tab-market-breadth",
        "alphaear-predictor",
        "tab-stock-sync",
        "tab-screening",
        "tab-technical-analysis",
        "tab-patterns",
        "tab-market-environment",
        "morning-ship",
        "tab-fundamental-sync",
        "tab-genai-features",
        "tab-dashboard-summary",
        "tab-kiwoom",
        "tab-news-fetch",
        "tab-glossary",
        "daily-strategy-post",
        "trading-trade-hypothesis-ideator",
        "weekly-stock-update"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "tab-strategy-comparison",
          "tokens": 571
        },
        {
          "skill_id": "tab-dualma-refresh",
          "tokens": 391
        },
        {
          "skill_id": "tab-turtle-refresh",
          "tokens": 371
        },
        {
          "skill_id": "tab-bollinger-refresh",
          "tokens": 383
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "tab-llm-agents",
          "tokens": 382
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "trading-backtest-expert",
          "tokens": 2754
        },
        {
          "skill_id": "tab-market-breadth",
          "tokens": 504
        },
        {
          "skill_id": "alphaear-predictor",
          "tokens": 803
        },
        {
          "skill_id": "tab-stock-sync",
          "tokens": 345
        },
        {
          "skill_id": "tab-screening",
          "tokens": 410
        },
        {
          "skill_id": "tab-technical-analysis",
          "tokens": 414
        },
        {
          "skill_id": "tab-patterns",
          "tokens": 509
        },
        {
          "skill_id": "tab-market-environment",
          "tokens": 566
        },
        {
          "skill_id": "morning-ship",
          "tokens": 2943
        },
        {
          "skill_id": "tab-fundamental-sync",
          "tokens": 331
        },
        {
          "skill_id": "tab-genai-features",
          "tokens": 354
        },
        {
          "skill_id": "tab-dashboard-summary",
          "tokens": 419
        },
        {
          "skill_id": "tab-kiwoom",
          "tokens": 452
        },
        {
          "skill_id": "tab-news-fetch",
          "tokens": 474
        },
        {
          "skill_id": "tab-glossary",
          "tokens": 538
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "trading-trade-hypothesis-ideator",
          "tokens": 909
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-strategy-comparison. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"tab-strategy-comparison\", \"tab-dualma-refresh\", \"tab-turtle-refresh\", \"tab-bollinger-refresh\", \"daily-stock-check\", \"tab-llm-agents\", \"tab-report-generate\", \"trading-backtest-expert\", \"tab-market-breadth\", \"alphaear-predictor\", \"tab-stock-sync\", \"tab-screening\", \"tab-technical-analysis\", \"tab-patterns\", \"tab-market-environment\", \"morning-ship\", \"tab-fundamental-sync\", \"tab-genai-features\", \"tab-dashboard-summary\", \"tab-kiwoom\", \"tab-news-fetch\", \"tab-glossary\", \"daily-strategy-post\", \"trading-trade-hypothesis-ideator\", \"weekly-stock-update\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0046",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Automation Strategist. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "automation-strategist",
        "pipeline-builder",
        "system-thinker",
        "qa-test-expert",
        "prompt-architect",
        "kwp-human-resources-org-planning",
        "kwp-engineering-testing-strategy",
        "kwp-engineering-tech-debt",
        "trading-trade-hypothesis-ideator",
        "agency-data-analytics-reporter",
        "trading-options-strategy-advisor",
        "codebase-archaeologist",
        "trading-position-sizer",
        "ai-workflow-integrator",
        "plans",
        "kwp-bio-research-scientific-problem-selection",
        "agency-analytics-reporter",
        "kwp-design-design-system-management",
        "daily-strategy-post",
        "pm-execution",
        "frontend-expert",
        "knowledge-daily-aggregator",
        "issue-to-code",
        "design-architect",
        "hf-trending-intelligence"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "automation-strategist",
          "tokens": 2482
        },
        {
          "skill_id": "pipeline-builder",
          "tokens": 2234
        },
        {
          "skill_id": "system-thinker",
          "tokens": 2684
        },
        {
          "skill_id": "qa-test-expert",
          "tokens": 1517
        },
        {
          "skill_id": "prompt-architect",
          "tokens": 1939
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "kwp-engineering-tech-debt",
          "tokens": 648
        },
        {
          "skill_id": "trading-trade-hypothesis-ideator",
          "tokens": 909
        },
        {
          "skill_id": "agency-data-analytics-reporter",
          "tokens": 1003
        },
        {
          "skill_id": "trading-options-strategy-advisor",
          "tokens": 1657
        },
        {
          "skill_id": "codebase-archaeologist",
          "tokens": 1977
        },
        {
          "skill_id": "trading-position-sizer",
          "tokens": 2013
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "plans",
          "tokens": 2297
        },
        {
          "skill_id": "kwp-bio-research-scientific-problem-selection",
          "tokens": 2990
        },
        {
          "skill_id": "agency-analytics-reporter",
          "tokens": 4157
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "pm-execution",
          "tokens": 1178
        },
        {
          "skill_id": "frontend-expert",
          "tokens": 1234
        },
        {
          "skill_id": "knowledge-daily-aggregator",
          "tokens": 1425
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "hf-trending-intelligence",
          "tokens": 1832
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Automation Strategist. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"automation-strategist\", \"pipeline-builder\", \"system-thinker\", \"qa-test-expert\", \"prompt-architect\", \"kwp-human-resources-org-planning\", \"kwp-engineering-testing-strategy\", \"kwp-engineering-tech-debt\", \"trading-trade-hypothesis-ideator\", \"agency-data-analytics-reporter\", \"trading-options-strategy-advisor\", \"codebase-archaeologist\", \"trading-position-sizer\", \"ai-workflow-integrator\", \"plans\", \"kwp-bio-research-scientific-problem-selection\", \"agency-analytics-reporter\", \"kwp-design-design-system-management\", \"daily-strategy-post\", \"pm-execution\", \"frontend-expert\", \"knowledge-daily-aggregator\", \"issue-to-code\", \"design-architect\", \"hf-trending-intelligence\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0047",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-market-breadth. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "tab-market-breadth",
        "tab-market-environment",
        "tab-turtle-refresh",
        "tab-bollinger-refresh",
        "tab-dualma-refresh",
        "tab-news-fetch",
        "tab-llm-agents",
        "tab-dashboard-summary",
        "tab-patterns",
        "tab-glossary",
        "tab-strategy-comparison",
        "tab-report-generate",
        "tab-technical-analysis",
        "daily-stock-check",
        "tab-fundamental-sync",
        "tab-stock-sync",
        "weekly-stock-update",
        "tab-hot-stock-discovery",
        "tab-kiwoom",
        "alphaear-news",
        "stock-csv-downloader",
        "trading-market-environment-analysis",
        "trading-market-breadth-analyzer",
        "trading-sector-analyst",
        "trading-technical-analyst"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "tab-market-breadth",
          "tokens": 504
        },
        {
          "skill_id": "tab-market-environment",
          "tokens": 566
        },
        {
          "skill_id": "tab-turtle-refresh",
          "tokens": 371
        },
        {
          "skill_id": "tab-bollinger-refresh",
          "tokens": 383
        },
        {
          "skill_id": "tab-dualma-refresh",
          "tokens": 391
        },
        {
          "skill_id": "tab-news-fetch",
          "tokens": 474
        },
        {
          "skill_id": "tab-llm-agents",
          "tokens": 382
        },
        {
          "skill_id": "tab-dashboard-summary",
          "tokens": 419
        },
        {
          "skill_id": "tab-patterns",
          "tokens": 509
        },
        {
          "skill_id": "tab-glossary",
          "tokens": 538
        },
        {
          "skill_id": "tab-strategy-comparison",
          "tokens": 571
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "tab-technical-analysis",
          "tokens": 414
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "tab-fundamental-sync",
          "tokens": 331
        },
        {
          "skill_id": "tab-stock-sync",
          "tokens": 345
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "tab-hot-stock-discovery",
          "tokens": 359
        },
        {
          "skill_id": "tab-kiwoom",
          "tokens": 452
        },
        {
          "skill_id": "alphaear-news",
          "tokens": 873
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "trading-market-breadth-analyzer",
          "tokens": 1830
        },
        {
          "skill_id": "trading-sector-analyst",
          "tokens": 2657
        },
        {
          "skill_id": "trading-technical-analyst",
          "tokens": 2868
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-market-breadth. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"tab-market-breadth\", \"tab-market-environment\", \"tab-turtle-refresh\", \"tab-bollinger-refresh\", \"tab-dualma-refresh\", \"tab-news-fetch\", \"tab-llm-agents\", \"tab-dashboard-summary\", \"tab-patterns\", \"tab-glossary\", \"tab-strategy-comparison\", \"tab-report-generate\", \"tab-technical-analysis\", \"daily-stock-check\", \"tab-fundamental-sync\", \"tab-stock-sync\", \"weekly-stock-update\", \"tab-hot-stock-discovery\", \"tab-kiwoom\", \"alphaear-news\", \"stock-csv-downloader\", \"trading-market-environment-analysis\", \"trading-market-breadth-analyzer\", \"trading-sector-analyst\", \"trading-technical-analyst\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0048",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Contact Research. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-common-room-contact-research",
        "kwp-common-room-compose-outreach",
        "kwp-common-room-account-research",
        "kwp-common-room-prospect",
        "kwp-sales-account-research",
        "kwp-apollo-enrich-lead",
        "kwp-sales-draft-outreach",
        "kwp-common-room-weekly-prep-brief",
        "kwp-common-room-call-prep",
        "kwp-marketing-content-creation",
        "mission-control copy",
        "kwp-data-data-context-extractor",
        "kwp-sales-call-prep",
        "kwp-bio-research-single-cell-rna-qc",
        "mission-control",
        "kwp-legal-meeting-briefing",
        "kwp-customer-support-customer-research",
        "kwp-product-management-user-research-synthesis",
        "kwp-legal-canned-responses",
        "kwp-human-resources-people-analytics",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-operations-resource-planning",
        "kwp-sales-daily-briefing",
        "kwp-enterprise-search-search-strategy"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-common-room-contact-research",
          "tokens": 1611
        },
        {
          "skill_id": "kwp-common-room-compose-outreach",
          "tokens": 1627
        },
        {
          "skill_id": "kwp-common-room-account-research",
          "tokens": 1784
        },
        {
          "skill_id": "kwp-common-room-prospect",
          "tokens": 1803
        },
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "kwp-apollo-enrich-lead",
          "tokens": 837
        },
        {
          "skill_id": "kwp-sales-draft-outreach",
          "tokens": 2782
        },
        {
          "skill_id": "kwp-common-room-weekly-prep-brief",
          "tokens": 1607
        },
        {
          "skill_id": "kwp-common-room-call-prep",
          "tokens": 1729
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "kwp-data-data-context-extractor",
          "tokens": 1883
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-customer-support-customer-research",
          "tokens": 2529
        },
        {
          "skill_id": "kwp-product-management-user-research-synthesis",
          "tokens": 3071
        },
        {
          "skill_id": "kwp-legal-canned-responses",
          "tokens": 3164
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "kwp-enterprise-search-search-strategy",
          "tokens": 2009
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Contact Research. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-common-room-contact-research\", \"kwp-common-room-compose-outreach\", \"kwp-common-room-account-research\", \"kwp-common-room-prospect\", \"kwp-sales-account-research\", \"kwp-apollo-enrich-lead\", \"kwp-sales-draft-outreach\", \"kwp-common-room-weekly-prep-brief\", \"kwp-common-room-call-prep\", \"kwp-marketing-content-creation\", \"mission-control copy\", \"kwp-data-data-context-extractor\", \"kwp-sales-call-prep\", \"kwp-bio-research-single-cell-rna-qc\", \"mission-control\", \"kwp-legal-meeting-briefing\", \"kwp-customer-support-customer-research\", \"kwp-product-management-user-research-synthesis\", \"kwp-legal-canned-responses\", \"kwp-human-resources-people-analytics\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-operations-resource-planning\", \"kwp-sales-daily-briefing\", \"kwp-enterprise-search-search-strategy\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0049",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Studio Operations Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-studio-operations",
        "agency-studio-producer",
        "agency-infrastructure-maintainer",
        "agency-project-shepherd",
        "agency-workflow-optimizer",
        "agency-legal-compliance-checker",
        "agency-tool-evaluator",
        "agency-ai-engineer",
        "agency-experiment-tracker",
        "agency-wechat-official-account-manager",
        "agency-developer-advocate",
        "agency-senior-project-manager",
        "agency-instagram-curator",
        "agency-visual-storyteller",
        "agency-tiktok-strategist",
        "agency-trend-researcher",
        "agency-api-tester",
        "agency-app-store-optimizer",
        "agency-xr-immersive-developer",
        "agency-growth-hacker",
        "agency-content-creator",
        "agency-reddit-community-builder",
        "agency-twitter-engager",
        "agency-whimsy-injector",
        "agency-xiaohongshu-specialist"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-studio-producer",
          "tokens": 2953
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-tool-evaluator",
          "tokens": 4549
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "agency-senior-project-manager",
          "tokens": 1555
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-app-store-optimizer",
          "tokens": 3512
        },
        {
          "skill_id": "agency-xr-immersive-developer",
          "tokens": 697
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Studio Operations Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-studio-operations\", \"agency-studio-producer\", \"agency-infrastructure-maintainer\", \"agency-project-shepherd\", \"agency-workflow-optimizer\", \"agency-legal-compliance-checker\", \"agency-tool-evaluator\", \"agency-ai-engineer\", \"agency-experiment-tracker\", \"agency-wechat-official-account-manager\", \"agency-developer-advocate\", \"agency-senior-project-manager\", \"agency-instagram-curator\", \"agency-visual-storyteller\", \"agency-tiktok-strategist\", \"agency-trend-researcher\", \"agency-api-tester\", \"agency-app-store-optimizer\", \"agency-xr-immersive-developer\", \"agency-growth-hacker\", \"agency-content-creator\", \"agency-reddit-community-builder\", \"agency-twitter-engager\", \"agency-whimsy-injector\", \"agency-xiaohongshu-specialist\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0050",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Decision Router. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "decision-router",
        "bespin-news-digest",
        "unified-intel-intake",
        "twitter-timeline-to-slack",
        "knowledge-daily-aggregator",
        "daily-am-orchestrator",
        "email-research-dispatcher",
        "ai-chief-of-staff",
        "iac-review-agent",
        "x-to-slack",
        "cursor-automations",
        "lead-lifecycle-manager",
        "planning-meeting-sentinel",
        "smart-meeting-scheduler",
        "github-sprint-digest",
        "daily-strategy-post",
        "proactive-meeting-scheduler",
        "hf-trending-intelligence",
        "morning",
        "full-stack-planner",
        "sentence-polisher",
        "setup-doctor",
        "meeting-digest",
        "sales-meeting-pipeline",
        "email-auto-reply"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "decision-router",
          "tokens": 1730
        },
        {
          "skill_id": "bespin-news-digest",
          "tokens": 3949
        },
        {
          "skill_id": "unified-intel-intake",
          "tokens": 1593
        },
        {
          "skill_id": "twitter-timeline-to-slack",
          "tokens": 6229
        },
        {
          "skill_id": "knowledge-daily-aggregator",
          "tokens": 1425
        },
        {
          "skill_id": "daily-am-orchestrator",
          "tokens": 3613
        },
        {
          "skill_id": "email-research-dispatcher",
          "tokens": 703
        },
        {
          "skill_id": "ai-chief-of-staff",
          "tokens": 1152
        },
        {
          "skill_id": "iac-review-agent",
          "tokens": 1312
        },
        {
          "skill_id": "x-to-slack",
          "tokens": 2001
        },
        {
          "skill_id": "cursor-automations",
          "tokens": 2911
        },
        {
          "skill_id": "lead-lifecycle-manager",
          "tokens": 533
        },
        {
          "skill_id": "planning-meeting-sentinel",
          "tokens": 604
        },
        {
          "skill_id": "smart-meeting-scheduler",
          "tokens": 639
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "proactive-meeting-scheduler",
          "tokens": 1444
        },
        {
          "skill_id": "hf-trending-intelligence",
          "tokens": 1832
        },
        {
          "skill_id": "morning",
          "tokens": 2088
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        },
        {
          "skill_id": "sentence-polisher",
          "tokens": 2167
        },
        {
          "skill_id": "setup-doctor",
          "tokens": 2578
        },
        {
          "skill_id": "meeting-digest",
          "tokens": 4491
        },
        {
          "skill_id": "sales-meeting-pipeline",
          "tokens": 559
        },
        {
          "skill_id": "email-auto-reply",
          "tokens": 638
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Decision Router. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"decision-router\", \"bespin-news-digest\", \"unified-intel-intake\", \"twitter-timeline-to-slack\", \"knowledge-daily-aggregator\", \"daily-am-orchestrator\", \"email-research-dispatcher\", \"ai-chief-of-staff\", \"iac-review-agent\", \"x-to-slack\", \"cursor-automations\", \"lead-lifecycle-manager\", \"planning-meeting-sentinel\", \"smart-meeting-scheduler\", \"github-sprint-digest\", \"daily-strategy-post\", \"proactive-meeting-scheduler\", \"hf-trending-intelligence\", \"morning\", \"full-stack-planner\", \"sentence-polisher\", \"setup-doctor\", \"meeting-digest\", \"sales-meeting-pipeline\", \"email-auto-reply\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0051",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AI Platform FE - FSD 개발 및 마이그레이션. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "fsd-development",
        "design-review",
        "figma-to-tds",
        "implement-screen",
        "fe-pipeline",
        "swagger-api-doc-generator",
        "github-workflow-automation",
        "overlay-layout-patterns",
        "screen-description",
        "notion-docs-sync",
        "anthropic-frontend-design",
        "design-architect",
        "tab-genai-features",
        "tab-llm-agents",
        "tab-screening",
        "planning-weekly-pulse",
        "dependency-radar",
        "lead-lifecycle-manager",
        "docs-freshness-guardian",
        "sales-weekly-dashboard",
        "sales-meeting-pipeline",
        "feature-intake-router",
        "proposal-factory",
        "planning-meeting-sentinel",
        "incident-lifecycle-manager"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "fsd-development",
          "tokens": 1234
        },
        {
          "skill_id": "design-review",
          "tokens": 947
        },
        {
          "skill_id": "figma-to-tds",
          "tokens": 987
        },
        {
          "skill_id": "implement-screen",
          "tokens": 996
        },
        {
          "skill_id": "fe-pipeline",
          "tokens": 2354
        },
        {
          "skill_id": "swagger-api-doc-generator",
          "tokens": 784
        },
        {
          "skill_id": "github-workflow-automation",
          "tokens": 971
        },
        {
          "skill_id": "overlay-layout-patterns",
          "tokens": 1047
        },
        {
          "skill_id": "screen-description",
          "tokens": 836
        },
        {
          "skill_id": "notion-docs-sync",
          "tokens": 857
        },
        {
          "skill_id": "anthropic-frontend-design",
          "tokens": 1312
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "tab-genai-features",
          "tokens": 354
        },
        {
          "skill_id": "tab-llm-agents",
          "tokens": 382
        },
        {
          "skill_id": "tab-screening",
          "tokens": 410
        },
        {
          "skill_id": "planning-weekly-pulse",
          "tokens": 528
        },
        {
          "skill_id": "dependency-radar",
          "tokens": 531
        },
        {
          "skill_id": "lead-lifecycle-manager",
          "tokens": 533
        },
        {
          "skill_id": "docs-freshness-guardian",
          "tokens": 557
        },
        {
          "skill_id": "sales-weekly-dashboard",
          "tokens": 558
        },
        {
          "skill_id": "sales-meeting-pipeline",
          "tokens": 559
        },
        {
          "skill_id": "feature-intake-router",
          "tokens": 568
        },
        {
          "skill_id": "proposal-factory",
          "tokens": 582
        },
        {
          "skill_id": "planning-meeting-sentinel",
          "tokens": 604
        },
        {
          "skill_id": "incident-lifecycle-manager",
          "tokens": 610
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AI Platform FE - FSD 개발 및 마이그레이션. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"fsd-development\", \"design-review\", \"figma-to-tds\", \"implement-screen\", \"fe-pipeline\", \"swagger-api-doc-generator\", \"github-workflow-automation\", \"overlay-layout-patterns\", \"screen-description\", \"notion-docs-sync\", \"anthropic-frontend-design\", \"design-architect\", \"tab-genai-features\", \"tab-llm-agents\", \"tab-screening\", \"planning-weekly-pulse\", \"dependency-radar\", \"lead-lifecycle-manager\", \"docs-freshness-guardian\", \"sales-weekly-dashboard\", \"sales-meeting-pipeline\", \"feature-intake-router\", \"proposal-factory\", \"planning-meeting-sentinel\", \"incident-lifecycle-manager\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0052",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Plans. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "plans",
        "full-stack-planner",
        "sefo-orchestrator",
        "test-suite",
        "rtk",
        "design-architect",
        "ralph-loop",
        "prompt-architect",
        "pipeline-builder",
        "release-commander",
        "mission-control",
        "problem-definition",
        "release-ship",
        "skill-autoimprove",
        "prd-research-factory",
        "pm-execution",
        "video-compress",
        "workflow-sequential",
        "ui-suite",
        "issue-to-code",
        "scrapling",
        "prompt-transformer",
        "air-gap-orchestrator",
        "mission-control copy",
        "trading-edge-candidate-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "plans",
          "tokens": 2297
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        },
        {
          "skill_id": "sefo-orchestrator",
          "tokens": 821
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "rtk",
          "tokens": 1214
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "ralph-loop",
          "tokens": 1925
        },
        {
          "skill_id": "prompt-architect",
          "tokens": 1939
        },
        {
          "skill_id": "pipeline-builder",
          "tokens": 2234
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "problem-definition",
          "tokens": 2601
        },
        {
          "skill_id": "release-ship",
          "tokens": 3056
        },
        {
          "skill_id": "skill-autoimprove",
          "tokens": 3666
        },
        {
          "skill_id": "prd-research-factory",
          "tokens": 648
        },
        {
          "skill_id": "pm-execution",
          "tokens": 1178
        },
        {
          "skill_id": "video-compress",
          "tokens": 1230
        },
        {
          "skill_id": "workflow-sequential",
          "tokens": 1485
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "scrapling",
          "tokens": 1675
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "air-gap-orchestrator",
          "tokens": 1788
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "trading-edge-candidate-agent",
          "tokens": 1971
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Plans. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"plans\", \"full-stack-planner\", \"sefo-orchestrator\", \"test-suite\", \"rtk\", \"design-architect\", \"ralph-loop\", \"prompt-architect\", \"pipeline-builder\", \"release-commander\", \"mission-control\", \"problem-definition\", \"release-ship\", \"skill-autoimprove\", \"prd-research-factory\", \"pm-execution\", \"video-compress\", \"workflow-sequential\", \"ui-suite\", \"issue-to-code\", \"scrapling\", \"prompt-transformer\", \"air-gap-orchestrator\", \"mission-control copy\", \"trading-edge-candidate-agent\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0053",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: NLM Slides: Local Document to Expert Slide Deck Pipeline. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "nlm-slides",
        "nlm-video",
        "nlm-arxiv-slides",
        "notebooklm-studio",
        "nlm-launch-kit",
        "nlm-strategy-hub",
        "paper-review",
        "nlm-discovery-lab",
        "nlm-deep-learn",
        "ppt-template-engine",
        "kwp-brand-voice-guideline-generation",
        "gws-docs",
        "gws-recipe-doc-from-template",
        "docx-template-engine",
        "office-template-enforcer",
        "hf-auto-finetune-planner",
        "auto-research",
        "pipeline-builder",
        "docs-tutor-setup",
        "prd-research-factory",
        "gws-recipe-save-email-to-doc",
        "knowledge-to-paper",
        "autoskill-judge",
        "commit-to-issue",
        "ui-suite"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "nlm-slides",
          "tokens": 1685
        },
        {
          "skill_id": "nlm-video",
          "tokens": 1763
        },
        {
          "skill_id": "nlm-arxiv-slides",
          "tokens": 3484
        },
        {
          "skill_id": "notebooklm-studio",
          "tokens": 1812
        },
        {
          "skill_id": "nlm-launch-kit",
          "tokens": 2368
        },
        {
          "skill_id": "nlm-strategy-hub",
          "tokens": 2207
        },
        {
          "skill_id": "paper-review",
          "tokens": 5815
        },
        {
          "skill_id": "nlm-discovery-lab",
          "tokens": 2808
        },
        {
          "skill_id": "nlm-deep-learn",
          "tokens": 3214
        },
        {
          "skill_id": "ppt-template-engine",
          "tokens": 1101
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "gws-recipe-doc-from-template",
          "tokens": 743
        },
        {
          "skill_id": "docx-template-engine",
          "tokens": 1193
        },
        {
          "skill_id": "office-template-enforcer",
          "tokens": 1273
        },
        {
          "skill_id": "hf-auto-finetune-planner",
          "tokens": 2091
        },
        {
          "skill_id": "auto-research",
          "tokens": 2139
        },
        {
          "skill_id": "pipeline-builder",
          "tokens": 2234
        },
        {
          "skill_id": "docs-tutor-setup",
          "tokens": 2288
        },
        {
          "skill_id": "prd-research-factory",
          "tokens": 648
        },
        {
          "skill_id": "gws-recipe-save-email-to-doc",
          "tokens": 676
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "autoskill-judge",
          "tokens": 1176
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: NLM Slides: Local Document to Expert Slide Deck Pipeline. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"nlm-slides\", \"nlm-video\", \"nlm-arxiv-slides\", \"notebooklm-studio\", \"nlm-launch-kit\", \"nlm-strategy-hub\", \"paper-review\", \"nlm-discovery-lab\", \"nlm-deep-learn\", \"ppt-template-engine\", \"kwp-brand-voice-guideline-generation\", \"gws-docs\", \"gws-recipe-doc-from-template\", \"docx-template-engine\", \"office-template-enforcer\", \"hf-auto-finetune-planner\", \"auto-research\", \"pipeline-builder\", \"docs-tutor-setup\", \"prd-research-factory\", \"gws-recipe-save-email-to-doc\", \"knowledge-to-paper\", \"autoskill-judge\", \"commit-to-issue\", \"ui-suite\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0054",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Test Suite. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "test-suite",
        "ui-suite",
        "deep-review",
        "simplify",
        "ship",
        "diagnose",
        "qa-test-expert",
        "release-commander",
        "ecc-verification-loop",
        "code-review-all",
        "e2e-overhaul",
        "skill-optimizer",
        "critical-review",
        "dependency-auditor",
        "incident-to-improvement",
        "ci-quality-gate",
        "i18n-sync",
        "compliance-gate",
        "evals-skills",
        "issue-to-code",
        "design-architect",
        "auto-research",
        "mission-control",
        "problem-definition",
        "release-ship"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "deep-review",
          "tokens": 1923
        },
        {
          "skill_id": "simplify",
          "tokens": 2596
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "diagnose",
          "tokens": 1433
        },
        {
          "skill_id": "qa-test-expert",
          "tokens": 1517
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "ecc-verification-loop",
          "tokens": 886
        },
        {
          "skill_id": "code-review-all",
          "tokens": 2524
        },
        {
          "skill_id": "e2e-overhaul",
          "tokens": 2425
        },
        {
          "skill_id": "skill-optimizer",
          "tokens": 2940
        },
        {
          "skill_id": "critical-review",
          "tokens": 3039
        },
        {
          "skill_id": "dependency-auditor",
          "tokens": 1458
        },
        {
          "skill_id": "incident-to-improvement",
          "tokens": 2624
        },
        {
          "skill_id": "ci-quality-gate",
          "tokens": 1054
        },
        {
          "skill_id": "i18n-sync",
          "tokens": 1282
        },
        {
          "skill_id": "compliance-gate",
          "tokens": 1376
        },
        {
          "skill_id": "evals-skills",
          "tokens": 1524
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "auto-research",
          "tokens": 2139
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "problem-definition",
          "tokens": 2601
        },
        {
          "skill_id": "release-ship",
          "tokens": 3056
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Test Suite. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"test-suite\", \"ui-suite\", \"deep-review\", \"simplify\", \"ship\", \"diagnose\", \"qa-test-expert\", \"release-commander\", \"ecc-verification-loop\", \"code-review-all\", \"e2e-overhaul\", \"skill-optimizer\", \"critical-review\", \"dependency-auditor\", \"incident-to-improvement\", \"ci-quality-gate\", \"i18n-sync\", \"compliance-gate\", \"evals-skills\", \"issue-to-code\", \"design-architect\", \"auto-research\", \"mission-control\", \"problem-definition\", \"release-ship\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0055",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PM Toolkit. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "pm-toolkit",
        "pm-go-to-market",
        "i18n-sync",
        "pm-marketing-growth",
        "pm-data-analytics",
        "pr-review-captain",
        "proposal-factory",
        "pm-product-discovery",
        "compliance-governance",
        "prompt-transformer",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "docs-tutor-setup",
        "agency-xiaohongshu-specialist",
        "skill-optimizer",
        "agency-legal-compliance-checker",
        "anthropic-doc-coauthoring",
        "gws-recipe-audit-sharing",
        "sefo-governance",
        "pm-market-research",
        "gws-calendar",
        "ux-expert",
        "role-developer",
        "evals-skills",
        "skill-guide-generator"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "pm-toolkit",
          "tokens": 935
        },
        {
          "skill_id": "pm-go-to-market",
          "tokens": 992
        },
        {
          "skill_id": "i18n-sync",
          "tokens": 1282
        },
        {
          "skill_id": "pm-marketing-growth",
          "tokens": 954
        },
        {
          "skill_id": "pm-data-analytics",
          "tokens": 956
        },
        {
          "skill_id": "pr-review-captain",
          "tokens": 1496
        },
        {
          "skill_id": "proposal-factory",
          "tokens": 582
        },
        {
          "skill_id": "pm-product-discovery",
          "tokens": 1198
        },
        {
          "skill_id": "compliance-governance",
          "tokens": 1501
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "docs-tutor-setup",
          "tokens": 2288
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "skill-optimizer",
          "tokens": 2940
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "anthropic-doc-coauthoring",
          "tokens": 4138
        },
        {
          "skill_id": "gws-recipe-audit-sharing",
          "tokens": 645
        },
        {
          "skill_id": "sefo-governance",
          "tokens": 837
        },
        {
          "skill_id": "pm-market-research",
          "tokens": 966
        },
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "ux-expert",
          "tokens": 1210
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "evals-skills",
          "tokens": 1524
        },
        {
          "skill_id": "skill-guide-generator",
          "tokens": 1858
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PM Toolkit. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"pm-toolkit\", \"pm-go-to-market\", \"i18n-sync\", \"pm-marketing-growth\", \"pm-data-analytics\", \"pr-review-captain\", \"proposal-factory\", \"pm-product-discovery\", \"compliance-governance\", \"prompt-transformer\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"docs-tutor-setup\", \"agency-xiaohongshu-specialist\", \"skill-optimizer\", \"agency-legal-compliance-checker\", \"anthropic-doc-coauthoring\", \"gws-recipe-audit-sharing\", \"sefo-governance\", \"pm-market-research\", \"gws-calendar\", \"ux-expert\", \"role-developer\", \"evals-skills\", \"skill-guide-generator\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0056",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Sentence Polisher. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "sentence-polisher",
        "workflow-sequential",
        "mission-control",
        "nlm-arxiv-slides",
        "pm-toolkit",
        "gws-calendar",
        "mission-control copy",
        "trading-data-quality-checker",
        "gws-email-reply",
        "notion-meeting-sync",
        "kwp-engineering-code-review",
        "github-sprint-digest",
        "autoskill-merger",
        "helm-validator",
        "commit-to-issue",
        "ui-suite",
        "ship",
        "decision-router",
        "gws-gmail",
        "test-suite",
        "prompt-architect",
        "full-stack-planner",
        "auto-research",
        "workflow-eval-opt",
        "ai-workflow-integrator"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "sentence-polisher",
          "tokens": 2167
        },
        {
          "skill_id": "workflow-sequential",
          "tokens": 1485
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "nlm-arxiv-slides",
          "tokens": 3484
        },
        {
          "skill_id": "pm-toolkit",
          "tokens": 935
        },
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "trading-data-quality-checker",
          "tokens": 1962
        },
        {
          "skill_id": "gws-email-reply",
          "tokens": 2222
        },
        {
          "skill_id": "notion-meeting-sync",
          "tokens": 4347
        },
        {
          "skill_id": "kwp-engineering-code-review",
          "tokens": 610
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "autoskill-merger",
          "tokens": 1142
        },
        {
          "skill_id": "helm-validator",
          "tokens": 1398
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "decision-router",
          "tokens": 1730
        },
        {
          "skill_id": "gws-gmail",
          "tokens": 1796
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "prompt-architect",
          "tokens": 1939
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        },
        {
          "skill_id": "auto-research",
          "tokens": 2139
        },
        {
          "skill_id": "workflow-eval-opt",
          "tokens": 2185
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Sentence Polisher. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"sentence-polisher\", \"workflow-sequential\", \"mission-control\", \"nlm-arxiv-slides\", \"pm-toolkit\", \"gws-calendar\", \"mission-control copy\", \"trading-data-quality-checker\", \"gws-email-reply\", \"notion-meeting-sync\", \"kwp-engineering-code-review\", \"github-sprint-digest\", \"autoskill-merger\", \"helm-validator\", \"commit-to-issue\", \"ui-suite\", \"ship\", \"decision-router\", \"gws-gmail\", \"test-suite\", \"prompt-architect\", \"full-stack-planner\", \"auto-research\", \"workflow-eval-opt\", \"ai-workflow-integrator\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0057",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AlphaXiv Paper Lookup. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "alphaxiv-paper-lookup",
        "paper-auto-classifier",
        "meeting-digest",
        "hf-papers",
        "hf-trending-intelligence",
        "defuddle",
        "docx-template-engine",
        "role-developer",
        "helm-validator",
        "refactor-simulator",
        "related-papers-scout",
        "notion-meeting-sync",
        "paper-review",
        "kwp-engineering-code-review",
        "email-auto-reply",
        "github-sprint-digest",
        "ppt-template-engine",
        "role-ceo",
        "role-ux-designer",
        "role-cso",
        "role-hr",
        "role-cto",
        "terraform-reviewer",
        "issue-to-code",
        "unified-intel-intake"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "alphaxiv-paper-lookup",
          "tokens": 1299
        },
        {
          "skill_id": "paper-auto-classifier",
          "tokens": 1262
        },
        {
          "skill_id": "meeting-digest",
          "tokens": 4491
        },
        {
          "skill_id": "hf-papers",
          "tokens": 817
        },
        {
          "skill_id": "hf-trending-intelligence",
          "tokens": 1832
        },
        {
          "skill_id": "defuddle",
          "tokens": 2409
        },
        {
          "skill_id": "docx-template-engine",
          "tokens": 1193
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "helm-validator",
          "tokens": 1398
        },
        {
          "skill_id": "refactor-simulator",
          "tokens": 2044
        },
        {
          "skill_id": "related-papers-scout",
          "tokens": 3560
        },
        {
          "skill_id": "notion-meeting-sync",
          "tokens": 4347
        },
        {
          "skill_id": "paper-review",
          "tokens": 5815
        },
        {
          "skill_id": "kwp-engineering-code-review",
          "tokens": 610
        },
        {
          "skill_id": "email-auto-reply",
          "tokens": 638
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "ppt-template-engine",
          "tokens": 1101
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "role-ux-designer",
          "tokens": 1268
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        },
        {
          "skill_id": "role-hr",
          "tokens": 1350
        },
        {
          "skill_id": "role-cto",
          "tokens": 1413
        },
        {
          "skill_id": "terraform-reviewer",
          "tokens": 1457
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "unified-intel-intake",
          "tokens": 1593
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AlphaXiv Paper Lookup. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"alphaxiv-paper-lookup\", \"paper-auto-classifier\", \"meeting-digest\", \"hf-papers\", \"hf-trending-intelligence\", \"defuddle\", \"docx-template-engine\", \"role-developer\", \"helm-validator\", \"refactor-simulator\", \"related-papers-scout\", \"notion-meeting-sync\", \"paper-review\", \"kwp-engineering-code-review\", \"email-auto-reply\", \"github-sprint-digest\", \"ppt-template-engine\", \"role-ceo\", \"role-ux-designer\", \"role-cso\", \"role-hr\", \"role-cto\", \"terraform-reviewer\", \"issue-to-code\", \"unified-intel-intake\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0058",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PR Review Captain. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "pr-review-captain",
        "technical-writer",
        "i18n-sync",
        "commit-to-issue",
        "docs-tutor-setup",
        "notion-meeting-sync",
        "domain-commit",
        "pm-toolkit",
        "critical-review",
        "anthropic-doc-coauthoring",
        "kwp-engineering-documentation",
        "github-sprint-digest",
        "kwp-operations-risk-assessment",
        "issue-to-code",
        "prompt-transformer",
        "release-commander",
        "visual-explainer",
        "kwp-operations-change-management",
        "alphaear-reporter",
        "role-developer",
        "sre-devops-expert",
        "qa-test-expert",
        "ship",
        "infra-runbook-generator",
        "skill-guide-generator"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "pr-review-captain",
          "tokens": 1496
        },
        {
          "skill_id": "technical-writer",
          "tokens": 1250
        },
        {
          "skill_id": "i18n-sync",
          "tokens": 1282
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "docs-tutor-setup",
          "tokens": 2288
        },
        {
          "skill_id": "notion-meeting-sync",
          "tokens": 4347
        },
        {
          "skill_id": "domain-commit",
          "tokens": 915
        },
        {
          "skill_id": "pm-toolkit",
          "tokens": 935
        },
        {
          "skill_id": "critical-review",
          "tokens": 3039
        },
        {
          "skill_id": "anthropic-doc-coauthoring",
          "tokens": 4138
        },
        {
          "skill_id": "kwp-engineering-documentation",
          "tokens": 654
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "kwp-operations-risk-assessment",
          "tokens": 808
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "visual-explainer",
          "tokens": 5408
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "alphaear-reporter",
          "tokens": 810
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "sre-devops-expert",
          "tokens": 1450
        },
        {
          "skill_id": "qa-test-expert",
          "tokens": 1517
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "infra-runbook-generator",
          "tokens": 1623
        },
        {
          "skill_id": "skill-guide-generator",
          "tokens": 1858
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PR Review Captain. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"pr-review-captain\", \"technical-writer\", \"i18n-sync\", \"commit-to-issue\", \"docs-tutor-setup\", \"notion-meeting-sync\", \"domain-commit\", \"pm-toolkit\", \"critical-review\", \"anthropic-doc-coauthoring\", \"kwp-engineering-documentation\", \"github-sprint-digest\", \"kwp-operations-risk-assessment\", \"issue-to-code\", \"prompt-transformer\", \"release-commander\", \"visual-explainer\", \"kwp-operations-change-management\", \"alphaear-reporter\", \"role-developer\", \"sre-devops-expert\", \"qa-test-expert\", \"ship\", \"infra-runbook-generator\", \"skill-guide-generator\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0059",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Proposal Factory. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "proposal-factory",
        "prd-research-factory",
        "lead-lifecycle-manager",
        "tech-support-brain",
        "sales-weekly-dashboard",
        "incident-lifecycle-manager",
        "pm-toolkit",
        "ship",
        "sales-meeting-pipeline",
        "feature-intake-router",
        "gws-recipe-doc-from-template",
        "docs-freshness-guardian",
        "tab-report-generate",
        "planning-meeting-sentinel",
        "commit-to-issue",
        "skill-seekers",
        "infra-runbook-generator",
        "test-suite",
        "full-stack-planner",
        "slack-agent",
        "kwp-legal-legal-risk-assessment",
        "anthropic-doc-coauthoring",
        "anthropic-algorithmic-art",
        "visual-explainer",
        "anthropic-skill-creator"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "proposal-factory",
          "tokens": 582
        },
        {
          "skill_id": "prd-research-factory",
          "tokens": 648
        },
        {
          "skill_id": "lead-lifecycle-manager",
          "tokens": 533
        },
        {
          "skill_id": "tech-support-brain",
          "tokens": 620
        },
        {
          "skill_id": "sales-weekly-dashboard",
          "tokens": 558
        },
        {
          "skill_id": "incident-lifecycle-manager",
          "tokens": 610
        },
        {
          "skill_id": "pm-toolkit",
          "tokens": 935
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "sales-meeting-pipeline",
          "tokens": 559
        },
        {
          "skill_id": "feature-intake-router",
          "tokens": 568
        },
        {
          "skill_id": "gws-recipe-doc-from-template",
          "tokens": 743
        },
        {
          "skill_id": "docs-freshness-guardian",
          "tokens": 557
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "planning-meeting-sentinel",
          "tokens": 604
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "skill-seekers",
          "tokens": 1568
        },
        {
          "skill_id": "infra-runbook-generator",
          "tokens": 1623
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        },
        {
          "skill_id": "slack-agent",
          "tokens": 2564
        },
        {
          "skill_id": "kwp-legal-legal-risk-assessment",
          "tokens": 3406
        },
        {
          "skill_id": "anthropic-doc-coauthoring",
          "tokens": 4138
        },
        {
          "skill_id": "anthropic-algorithmic-art",
          "tokens": 4530
        },
        {
          "skill_id": "visual-explainer",
          "tokens": 5408
        },
        {
          "skill_id": "anthropic-skill-creator",
          "tokens": 6862
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Proposal Factory. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"proposal-factory\", \"prd-research-factory\", \"lead-lifecycle-manager\", \"tech-support-brain\", \"sales-weekly-dashboard\", \"incident-lifecycle-manager\", \"pm-toolkit\", \"ship\", \"sales-meeting-pipeline\", \"feature-intake-router\", \"gws-recipe-doc-from-template\", \"docs-freshness-guardian\", \"tab-report-generate\", \"planning-meeting-sentinel\", \"commit-to-issue\", \"skill-seekers\", \"infra-runbook-generator\", \"test-suite\", \"full-stack-planner\", \"slack-agent\", \"kwp-legal-legal-risk-assessment\", \"anthropic-doc-coauthoring\", \"anthropic-algorithmic-art\", \"visual-explainer\", \"anthropic-skill-creator\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0060",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Workflow Optimizer Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-workflow-optimizer",
        "agency-studio-operations",
        "agency-infrastructure-maintainer",
        "agency-api-tester",
        "agency-tool-evaluator",
        "agency-app-store-optimizer",
        "agency-legal-compliance-checker",
        "agency-content-creator",
        "agency-instagram-curator",
        "agency-visual-storyteller",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "agency-twitter-engager",
        "agency-whimsy-injector",
        "agency-xiaohongshu-specialist",
        "agency-project-shepherd",
        "agency-zhihu-strategist",
        "agency-test-results-analyzer",
        "agency-growth-hacker",
        "agency-cultural-intelligence-strategist",
        "agency-inclusive-visuals-specialist",
        "agency-feedback-synthesizer",
        "agency-ai-engineer",
        "agency-experiment-tracker",
        "agency-trend-researcher"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-tool-evaluator",
          "tokens": 4549
        },
        {
          "skill_id": "agency-app-store-optimizer",
          "tokens": 3512
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-zhihu-strategist",
          "tokens": 3239
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-cultural-intelligence-strategist",
          "tokens": 1892
        },
        {
          "skill_id": "agency-inclusive-visuals-specialist",
          "tokens": 1931
        },
        {
          "skill_id": "agency-feedback-synthesizer",
          "tokens": 2041
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Workflow Optimizer Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-workflow-optimizer\", \"agency-studio-operations\", \"agency-infrastructure-maintainer\", \"agency-api-tester\", \"agency-tool-evaluator\", \"agency-app-store-optimizer\", \"agency-legal-compliance-checker\", \"agency-content-creator\", \"agency-instagram-curator\", \"agency-visual-storyteller\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"agency-twitter-engager\", \"agency-whimsy-injector\", \"agency-xiaohongshu-specialist\", \"agency-project-shepherd\", \"agency-zhihu-strategist\", \"agency-test-results-analyzer\", \"agency-growth-hacker\", \"agency-cultural-intelligence-strategist\", \"agency-inclusive-visuals-specialist\", \"agency-feedback-synthesizer\", \"agency-ai-engineer\", \"agency-experiment-tracker\", \"agency-trend-researcher\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0061",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Studio Operations Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-studio-operations",
        "agency-studio-producer",
        "agency-infrastructure-maintainer",
        "agency-project-shepherd",
        "agency-workflow-optimizer",
        "agency-legal-compliance-checker",
        "agency-tool-evaluator",
        "agency-ai-engineer",
        "agency-experiment-tracker",
        "agency-wechat-official-account-manager",
        "agency-developer-advocate",
        "agency-senior-project-manager",
        "agency-instagram-curator",
        "agency-visual-storyteller",
        "agency-tiktok-strategist",
        "agency-trend-researcher",
        "agency-api-tester",
        "agency-app-store-optimizer",
        "agency-xr-immersive-developer",
        "agency-growth-hacker",
        "agency-content-creator",
        "agency-reddit-community-builder",
        "agency-twitter-engager",
        "agency-whimsy-injector",
        "agency-xiaohongshu-specialist"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-studio-producer",
          "tokens": 2953
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-tool-evaluator",
          "tokens": 4549
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "agency-senior-project-manager",
          "tokens": 1555
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-app-store-optimizer",
          "tokens": 3512
        },
        {
          "skill_id": "agency-xr-immersive-developer",
          "tokens": 697
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Studio Operations Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-studio-operations\", \"agency-studio-producer\", \"agency-infrastructure-maintainer\", \"agency-project-shepherd\", \"agency-workflow-optimizer\", \"agency-legal-compliance-checker\", \"agency-tool-evaluator\", \"agency-ai-engineer\", \"agency-experiment-tracker\", \"agency-wechat-official-account-manager\", \"agency-developer-advocate\", \"agency-senior-project-manager\", \"agency-instagram-curator\", \"agency-visual-storyteller\", \"agency-tiktok-strategist\", \"agency-trend-researcher\", \"agency-api-tester\", \"agency-app-store-optimizer\", \"agency-xr-immersive-developer\", \"agency-growth-hacker\", \"agency-content-creator\", \"agency-reddit-community-builder\", \"agency-twitter-engager\", \"agency-whimsy-injector\", \"agency-xiaohongshu-specialist\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0062",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: UX Researcher Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-ux-researcher",
        "agency-trend-researcher",
        "agency-data-analytics-reporter",
        "agency-experiment-tracker",
        "agency-analytics-reporter",
        "agency-growth-hacker",
        "agency-feedback-synthesizer",
        "agency-sprint-prioritizer",
        "agency-finance-tracker",
        "agency-ui-designer",
        "agency-developer-advocate",
        "agency-visual-storyteller",
        "agency-ai-engineer",
        "agency-test-results-analyzer",
        "agency-data-engineer",
        "agency-data-consolidation-agent",
        "agency-reddit-community-builder",
        "agency-twitter-engager",
        "agency-whimsy-injector",
        "agency-backend-architect",
        "agency-xiaohongshu-specialist",
        "agency-project-shepherd",
        "agency-studio-operations",
        "agency-wechat-official-account-manager",
        "agency-image-prompt-engineer"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-ux-researcher",
          "tokens": 3473
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "agency-data-analytics-reporter",
          "tokens": 1003
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-analytics-reporter",
          "tokens": 4157
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-feedback-synthesizer",
          "tokens": 2041
        },
        {
          "skill_id": "agency-sprint-prioritizer",
          "tokens": 2579
        },
        {
          "skill_id": "agency-finance-tracker",
          "tokens": 3177
        },
        {
          "skill_id": "agency-ui-designer",
          "tokens": 3538
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "agency-data-engineer",
          "tokens": 3843
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-backend-architect",
          "tokens": 2600
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-image-prompt-engineer",
          "tokens": 3027
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: UX Researcher Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-ux-researcher\", \"agency-trend-researcher\", \"agency-data-analytics-reporter\", \"agency-experiment-tracker\", \"agency-analytics-reporter\", \"agency-growth-hacker\", \"agency-feedback-synthesizer\", \"agency-sprint-prioritizer\", \"agency-finance-tracker\", \"agency-ui-designer\", \"agency-developer-advocate\", \"agency-visual-storyteller\", \"agency-ai-engineer\", \"agency-test-results-analyzer\", \"agency-data-engineer\", \"agency-data-consolidation-agent\", \"agency-reddit-community-builder\", \"agency-twitter-engager\", \"agency-whimsy-injector\", \"agency-backend-architect\", \"agency-xiaohongshu-specialist\", \"agency-project-shepherd\", \"agency-studio-operations\", \"agency-wechat-official-account-manager\", \"agency-image-prompt-engineer\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0063",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PM Product Discovery. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "pm-product-discovery",
        "pm-market-research",
        "pm-marketing-growth",
        "pm-go-to-market",
        "pm-execution",
        "pm-product-strategy",
        "pm-data-analytics",
        "nlm-discovery-lab",
        "pm-toolkit",
        "role-cso",
        "role-pm",
        "evals-skills",
        "prompt-architect",
        "agency-sprint-prioritizer",
        "kwp-product-management-user-research-synthesis",
        "role-ux-designer",
        "role-hr",
        "qa-test-expert",
        "ai-workflow-integrator",
        "mission-control",
        "agency-executive-summary-generator",
        "agency-experiment-tracker",
        "kwp-cowork-create-cowork-plugin",
        "agency-ux-researcher",
        "email-research-dispatcher"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "pm-product-discovery",
          "tokens": 1198
        },
        {
          "skill_id": "pm-market-research",
          "tokens": 966
        },
        {
          "skill_id": "pm-marketing-growth",
          "tokens": 954
        },
        {
          "skill_id": "pm-go-to-market",
          "tokens": 992
        },
        {
          "skill_id": "pm-execution",
          "tokens": 1178
        },
        {
          "skill_id": "pm-product-strategy",
          "tokens": 1271
        },
        {
          "skill_id": "pm-data-analytics",
          "tokens": 956
        },
        {
          "skill_id": "nlm-discovery-lab",
          "tokens": 2808
        },
        {
          "skill_id": "pm-toolkit",
          "tokens": 935
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        },
        {
          "skill_id": "role-pm",
          "tokens": 1304
        },
        {
          "skill_id": "evals-skills",
          "tokens": 1524
        },
        {
          "skill_id": "prompt-architect",
          "tokens": 1939
        },
        {
          "skill_id": "agency-sprint-prioritizer",
          "tokens": 2579
        },
        {
          "skill_id": "kwp-product-management-user-research-synthesis",
          "tokens": 3071
        },
        {
          "skill_id": "role-ux-designer",
          "tokens": 1268
        },
        {
          "skill_id": "role-hr",
          "tokens": 1350
        },
        {
          "skill_id": "qa-test-expert",
          "tokens": 1517
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "agency-executive-summary-generator",
          "tokens": 2465
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "kwp-cowork-create-cowork-plugin",
          "tokens": 3017
        },
        {
          "skill_id": "agency-ux-researcher",
          "tokens": 3473
        },
        {
          "skill_id": "email-research-dispatcher",
          "tokens": 703
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PM Product Discovery. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"pm-product-discovery\", \"pm-market-research\", \"pm-marketing-growth\", \"pm-go-to-market\", \"pm-execution\", \"pm-product-strategy\", \"pm-data-analytics\", \"nlm-discovery-lab\", \"pm-toolkit\", \"role-cso\", \"role-pm\", \"evals-skills\", \"prompt-architect\", \"agency-sprint-prioritizer\", \"kwp-product-management-user-research-synthesis\", \"role-ux-designer\", \"role-hr\", \"qa-test-expert\", \"ai-workflow-integrator\", \"mission-control\", \"agency-executive-summary-generator\", \"agency-experiment-tracker\", \"kwp-cowork-create-cowork-plugin\", \"agency-ux-researcher\", \"email-research-dispatcher\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0064",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Call Prep. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-common-room-call-prep",
        "kwp-common-room-account-research",
        "kwp-common-room-weekly-prep-brief",
        "kwp-common-room-prospect",
        "kwp-common-room-compose-outreach",
        "kwp-sales-call-prep",
        "kwp-common-room-contact-research",
        "kwp-legal-meeting-briefing",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-human-resources-employee-handbook",
        "kwp-operations-resource-planning",
        "kwp-sales-daily-briefing",
        "kwp-sales-account-research",
        "kwp-finance-journal-entry-prep",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-sales-draft-outreach",
        "kwp-operations-vendor-management",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-interview-prep",
        "kwp-apollo-enrich-lead",
        "kwp-apollo-prospect",
        "kwp-brand-voice-discover-brand",
        "kwp-customer-support-escalation",
        "kwp-customer-support-knowledge-management",
        "kwp-data-data-validation"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-common-room-call-prep",
          "tokens": 1729
        },
        {
          "skill_id": "kwp-common-room-account-research",
          "tokens": 1784
        },
        {
          "skill_id": "kwp-common-room-weekly-prep-brief",
          "tokens": 1607
        },
        {
          "skill_id": "kwp-common-room-prospect",
          "tokens": 1803
        },
        {
          "skill_id": "kwp-common-room-compose-outreach",
          "tokens": 1627
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-common-room-contact-research",
          "tokens": 1611
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-human-resources-employee-handbook",
          "tokens": 629
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "kwp-finance-journal-entry-prep",
          "tokens": 2215
        },
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "kwp-sales-draft-outreach",
          "tokens": 2782
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "kwp-apollo-enrich-lead",
          "tokens": 837
        },
        {
          "skill_id": "kwp-apollo-prospect",
          "tokens": 1006
        },
        {
          "skill_id": "kwp-brand-voice-discover-brand",
          "tokens": 1897
        },
        {
          "skill_id": "kwp-customer-support-escalation",
          "tokens": 2329
        },
        {
          "skill_id": "kwp-customer-support-knowledge-management",
          "tokens": 2624
        },
        {
          "skill_id": "kwp-data-data-validation",
          "tokens": 2650
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Call Prep. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-common-room-call-prep\", \"kwp-common-room-account-research\", \"kwp-common-room-weekly-prep-brief\", \"kwp-common-room-prospect\", \"kwp-common-room-compose-outreach\", \"kwp-sales-call-prep\", \"kwp-common-room-contact-research\", \"kwp-legal-meeting-briefing\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-human-resources-employee-handbook\", \"kwp-operations-resource-planning\", \"kwp-sales-daily-briefing\", \"kwp-sales-account-research\", \"kwp-finance-journal-entry-prep\", \"kwp-bio-research-single-cell-rna-qc\", \"kwp-sales-draft-outreach\", \"kwp-operations-vendor-management\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-interview-prep\", \"kwp-apollo-enrich-lead\", \"kwp-apollo-prospect\", \"kwp-brand-voice-discover-brand\", \"kwp-customer-support-escalation\", \"kwp-customer-support-knowledge-management\", \"kwp-data-data-validation\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0065",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Mission Control. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "mission-control",
        "mission-control copy",
        "release-commander",
        "sefo-orchestrator",
        "workflow-parallel",
        "full-stack-planner",
        "workflow-sequential",
        "ui-suite",
        "deep-review",
        "agency-data-consolidation-agent",
        "ai-chief-of-staff",
        "iac-review-agent",
        "commit-to-issue",
        "test-suite",
        "agency-visual-storyteller",
        "ai-workflow-integrator",
        "agency-workflow-optimizer",
        "agency-studio-producer",
        "agency-support-responder",
        "agency-test-results-analyzer",
        "agency-agents-orchestrator",
        "agency-agentic-identity-trust-architect",
        "github-sprint-digest",
        "gws-workflows",
        "agency-behavioral-nudge-engine"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "sefo-orchestrator",
          "tokens": 821
        },
        {
          "skill_id": "workflow-parallel",
          "tokens": 1734
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        },
        {
          "skill_id": "workflow-sequential",
          "tokens": 1485
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "deep-review",
          "tokens": 1923
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        },
        {
          "skill_id": "ai-chief-of-staff",
          "tokens": 1152
        },
        {
          "skill_id": "iac-review-agent",
          "tokens": 1312
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-studio-producer",
          "tokens": 2953
        },
        {
          "skill_id": "agency-support-responder",
          "tokens": 3296
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "agency-agents-orchestrator",
          "tokens": 4128
        },
        {
          "skill_id": "agency-agentic-identity-trust-architect",
          "tokens": 4533
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "gws-workflows",
          "tokens": 952
        },
        {
          "skill_id": "agency-behavioral-nudge-engine",
          "tokens": 1553
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Mission Control. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"mission-control\", \"mission-control copy\", \"release-commander\", \"sefo-orchestrator\", \"workflow-parallel\", \"full-stack-planner\", \"workflow-sequential\", \"ui-suite\", \"deep-review\", \"agency-data-consolidation-agent\", \"ai-chief-of-staff\", \"iac-review-agent\", \"commit-to-issue\", \"test-suite\", \"agency-visual-storyteller\", \"ai-workflow-integrator\", \"agency-workflow-optimizer\", \"agency-studio-producer\", \"agency-support-responder\", \"agency-test-results-analyzer\", \"agency-agents-orchestrator\", \"agency-agentic-identity-trust-architect\", \"github-sprint-digest\", \"gws-workflows\", \"agency-behavioral-nudge-engine\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0066",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Helm Validator. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "helm-validator",
        "k8s-manifest-validator",
        "terraform-reviewer",
        "iac-review-agent",
        "infra-drift-detector",
        "ci-quality-gate",
        "compliance-gate",
        "infra-runbook-generator",
        "ecc-verification-loop",
        "semantic-guard",
        "sefo-governance",
        "role-cto",
        "test-suite",
        "prompt-architect",
        "intent-alignment-tracker",
        "alphaxiv-paper-lookup",
        "dependency-auditor",
        "ui-suite",
        "deep-review",
        "refactor-simulator",
        "ai-quality-evaluator",
        "release-commander",
        "e2e-overhaul",
        "code-review-all",
        "setup-doctor"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "helm-validator",
          "tokens": 1398
        },
        {
          "skill_id": "k8s-manifest-validator",
          "tokens": 1431
        },
        {
          "skill_id": "terraform-reviewer",
          "tokens": 1457
        },
        {
          "skill_id": "iac-review-agent",
          "tokens": 1312
        },
        {
          "skill_id": "infra-drift-detector",
          "tokens": 1315
        },
        {
          "skill_id": "ci-quality-gate",
          "tokens": 1054
        },
        {
          "skill_id": "compliance-gate",
          "tokens": 1376
        },
        {
          "skill_id": "infra-runbook-generator",
          "tokens": 1623
        },
        {
          "skill_id": "ecc-verification-loop",
          "tokens": 886
        },
        {
          "skill_id": "semantic-guard",
          "tokens": 571
        },
        {
          "skill_id": "sefo-governance",
          "tokens": 837
        },
        {
          "skill_id": "role-cto",
          "tokens": 1413
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "prompt-architect",
          "tokens": 1939
        },
        {
          "skill_id": "intent-alignment-tracker",
          "tokens": 558
        },
        {
          "skill_id": "alphaxiv-paper-lookup",
          "tokens": 1299
        },
        {
          "skill_id": "dependency-auditor",
          "tokens": 1458
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "deep-review",
          "tokens": 1923
        },
        {
          "skill_id": "refactor-simulator",
          "tokens": 2044
        },
        {
          "skill_id": "ai-quality-evaluator",
          "tokens": 2228
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "e2e-overhaul",
          "tokens": 2425
        },
        {
          "skill_id": "code-review-all",
          "tokens": 2524
        },
        {
          "skill_id": "setup-doctor",
          "tokens": 2578
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Helm Validator. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"helm-validator\", \"k8s-manifest-validator\", \"terraform-reviewer\", \"iac-review-agent\", \"infra-drift-detector\", \"ci-quality-gate\", \"compliance-gate\", \"infra-runbook-generator\", \"ecc-verification-loop\", \"semantic-guard\", \"sefo-governance\", \"role-cto\", \"test-suite\", \"prompt-architect\", \"intent-alignment-tracker\", \"alphaxiv-paper-lookup\", \"dependency-auditor\", \"ui-suite\", \"deep-review\", \"refactor-simulator\", \"ai-quality-evaluator\", \"release-commander\", \"e2e-overhaul\", \"code-review-all\", \"setup-doctor\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0067",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Project Manager Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-senior-project-manager",
        "agency-project-shepherd",
        "agency-studio-operations",
        "agency-experiment-tracker",
        "agency-studio-producer",
        "agency-instagram-curator",
        "agency-visual-storyteller",
        "agency-ai-engineer",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "agency-twitter-engager",
        "agency-whimsy-injector",
        "agency-xiaohongshu-specialist",
        "agency-wechat-official-account-manager",
        "agency-workflow-optimizer",
        "agency-trend-researcher",
        "agency-zhihu-strategist",
        "agency-infrastructure-maintainer",
        "agency-api-tester",
        "agency-app-store-optimizer",
        "agency-test-results-analyzer",
        "agency-tool-evaluator",
        "agency-xr-interface-architect",
        "agency-xr-cockpit-interaction-specialist",
        "agency-xr-immersive-developer"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-senior-project-manager",
          "tokens": 1555
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-studio-producer",
          "tokens": 2953
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "agency-zhihu-strategist",
          "tokens": 3239
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-app-store-optimizer",
          "tokens": 3512
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "agency-tool-evaluator",
          "tokens": 4549
        },
        {
          "skill_id": "agency-xr-interface-architect",
          "tokens": 657
        },
        {
          "skill_id": "agency-xr-cockpit-interaction-specialist",
          "tokens": 693
        },
        {
          "skill_id": "agency-xr-immersive-developer",
          "tokens": 697
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Project Manager Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-senior-project-manager\", \"agency-project-shepherd\", \"agency-studio-operations\", \"agency-experiment-tracker\", \"agency-studio-producer\", \"agency-instagram-curator\", \"agency-visual-storyteller\", \"agency-ai-engineer\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"agency-twitter-engager\", \"agency-whimsy-injector\", \"agency-xiaohongshu-specialist\", \"agency-wechat-official-account-manager\", \"agency-workflow-optimizer\", \"agency-trend-researcher\", \"agency-zhihu-strategist\", \"agency-infrastructure-maintainer\", \"agency-api-tester\", \"agency-app-store-optimizer\", \"agency-test-results-analyzer\", \"agency-tool-evaluator\", \"agency-xr-interface-architect\", \"agency-xr-cockpit-interaction-specialist\", \"agency-xr-immersive-developer\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0068",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Close Management. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-finance-close-management",
        "kwp-finance-journal-entry-prep",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-sales-daily-briefing",
        "kwp-marketing-campaign-planning",
        "kwp-operations-vendor-management",
        "kwp-human-resources-org-planning",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-operations-resource-planning",
        "kwp-finance-reconciliation",
        "kwp-legal-meeting-briefing",
        "kwp-finance-financial-statements",
        "kwp-finance-variance-analysis",
        "kwp-product-management-metrics-tracking",
        "kwp-marketing-performance-analytics",
        "kwp-finance-audit-support",
        "kwp-human-resources-people-analytics",
        "kwp-human-resources-interview-prep",
        "gws-calendar",
        "mission-control copy",
        "kwp-sales-account-research",
        "kwp-sales-call-prep",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-customer-support-escalation"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-finance-close-management",
          "tokens": 2782
        },
        {
          "skill_id": "kwp-finance-journal-entry-prep",
          "tokens": 2215
        },
        {
          "skill_id": "kwp-human-resources-recruiting-pipeline",
          "tokens": 578
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "kwp-marketing-campaign-planning",
          "tokens": 2803
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-finance-reconciliation",
          "tokens": 2287
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-finance-financial-statements",
          "tokens": 2768
        },
        {
          "skill_id": "kwp-finance-variance-analysis",
          "tokens": 3208
        },
        {
          "skill_id": "kwp-product-management-metrics-tracking",
          "tokens": 3606
        },
        {
          "skill_id": "kwp-marketing-performance-analytics",
          "tokens": 3966
        },
        {
          "skill_id": "kwp-finance-audit-support",
          "tokens": 4426
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "kwp-customer-support-escalation",
          "tokens": 2329
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Close Management. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-finance-close-management\", \"kwp-finance-journal-entry-prep\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-sales-daily-briefing\", \"kwp-marketing-campaign-planning\", \"kwp-operations-vendor-management\", \"kwp-human-resources-org-planning\", \"kwp-operations-process-optimization\", \"kwp-operations-change-management\", \"kwp-operations-resource-planning\", \"kwp-finance-reconciliation\", \"kwp-legal-meeting-briefing\", \"kwp-finance-financial-statements\", \"kwp-finance-variance-analysis\", \"kwp-product-management-metrics-tracking\", \"kwp-marketing-performance-analytics\", \"kwp-finance-audit-support\", \"kwp-human-resources-people-analytics\", \"kwp-human-resources-interview-prep\", \"gws-calendar\", \"mission-control copy\", \"kwp-sales-account-research\", \"kwp-sales-call-prep\", \"kwp-bio-research-single-cell-rna-qc\", \"kwp-customer-support-escalation\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0069",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PM Product Strategy. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "pm-product-strategy",
        "pm-market-research",
        "pm-marketing-growth",
        "pm-go-to-market",
        "pm-product-discovery",
        "pm-execution",
        "nlm-strategy-hub",
        "pm-data-analytics",
        "role-cso",
        "evals-skills",
        "trading-options-strategy-advisor",
        "agency-executive-summary-generator",
        "pm-toolkit",
        "ai-workflow-integrator",
        "agency-sprint-prioritizer",
        "nlm-discovery-lab",
        "agency-trend-researcher",
        "daily-strategy-post",
        "agency-data-analytics-reporter",
        "hf-hub",
        "role-ceo",
        "role-pm",
        "e2e-testing",
        "mission-control copy",
        "deep-research-pipeline"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "pm-product-strategy",
          "tokens": 1271
        },
        {
          "skill_id": "pm-market-research",
          "tokens": 966
        },
        {
          "skill_id": "pm-marketing-growth",
          "tokens": 954
        },
        {
          "skill_id": "pm-go-to-market",
          "tokens": 992
        },
        {
          "skill_id": "pm-product-discovery",
          "tokens": 1198
        },
        {
          "skill_id": "pm-execution",
          "tokens": 1178
        },
        {
          "skill_id": "nlm-strategy-hub",
          "tokens": 2207
        },
        {
          "skill_id": "pm-data-analytics",
          "tokens": 956
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        },
        {
          "skill_id": "evals-skills",
          "tokens": 1524
        },
        {
          "skill_id": "trading-options-strategy-advisor",
          "tokens": 1657
        },
        {
          "skill_id": "agency-executive-summary-generator",
          "tokens": 2465
        },
        {
          "skill_id": "pm-toolkit",
          "tokens": 935
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "agency-sprint-prioritizer",
          "tokens": 2579
        },
        {
          "skill_id": "nlm-discovery-lab",
          "tokens": 2808
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "agency-data-analytics-reporter",
          "tokens": 1003
        },
        {
          "skill_id": "hf-hub",
          "tokens": 1151
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "role-pm",
          "tokens": 1304
        },
        {
          "skill_id": "e2e-testing",
          "tokens": 1484
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PM Product Strategy. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"pm-product-strategy\", \"pm-market-research\", \"pm-marketing-growth\", \"pm-go-to-market\", \"pm-product-discovery\", \"pm-execution\", \"nlm-strategy-hub\", \"pm-data-analytics\", \"role-cso\", \"evals-skills\", \"trading-options-strategy-advisor\", \"agency-executive-summary-generator\", \"pm-toolkit\", \"ai-workflow-integrator\", \"agency-sprint-prioritizer\", \"nlm-discovery-lab\", \"agency-trend-researcher\", \"daily-strategy-post\", \"agency-data-analytics-reporter\", \"hf-hub\", \"role-ceo\", \"role-pm\", \"e2e-testing\", \"mission-control copy\", \"deep-research-pipeline\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0070",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Running Workloads on Hugging Face Jobs. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "hf-jobs",
        "hf-model-trainer",
        "hf-cli",
        "hf-evaluation",
        "hf-datasets",
        "weekly-stock-update",
        "hf-spaces",
        "hf-endpoints",
        "hf-dataset-viewer",
        "daily-stock-check",
        "alphaear-reporter",
        "stock-csv-downloader",
        "hf-trackio",
        "tab-llm-agents",
        "tab-kiwoom",
        "alphaear-search",
        "alphaear-deepear-lite",
        "alphaear-predictor",
        "alphaear-stock",
        "alphaear-sentiment",
        "hf-models",
        "mirofish-financial-sim",
        "sre-devops-expert",
        "trading-market-environment-analysis",
        "ai-workflow-integrator"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "hf-jobs",
          "tokens": 4467
        },
        {
          "skill_id": "hf-model-trainer",
          "tokens": 5150
        },
        {
          "skill_id": "hf-cli",
          "tokens": 2275
        },
        {
          "skill_id": "hf-evaluation",
          "tokens": 4549
        },
        {
          "skill_id": "hf-datasets",
          "tokens": 3241
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "hf-spaces",
          "tokens": 930
        },
        {
          "skill_id": "hf-endpoints",
          "tokens": 1177
        },
        {
          "skill_id": "hf-dataset-viewer",
          "tokens": 1387
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "alphaear-reporter",
          "tokens": 810
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "hf-trackio",
          "tokens": 1424
        },
        {
          "skill_id": "tab-llm-agents",
          "tokens": 382
        },
        {
          "skill_id": "tab-kiwoom",
          "tokens": 452
        },
        {
          "skill_id": "alphaear-search",
          "tokens": 672
        },
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        },
        {
          "skill_id": "alphaear-predictor",
          "tokens": 803
        },
        {
          "skill_id": "alphaear-stock",
          "tokens": 820
        },
        {
          "skill_id": "alphaear-sentiment",
          "tokens": 928
        },
        {
          "skill_id": "hf-models",
          "tokens": 1212
        },
        {
          "skill_id": "mirofish-financial-sim",
          "tokens": 1330
        },
        {
          "skill_id": "sre-devops-expert",
          "tokens": 1450
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Running Workloads on Hugging Face Jobs. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"hf-jobs\", \"hf-model-trainer\", \"hf-cli\", \"hf-evaluation\", \"hf-datasets\", \"weekly-stock-update\", \"hf-spaces\", \"hf-endpoints\", \"hf-dataset-viewer\", \"daily-stock-check\", \"alphaear-reporter\", \"stock-csv-downloader\", \"hf-trackio\", \"tab-llm-agents\", \"tab-kiwoom\", \"alphaear-search\", \"alphaear-deepear-lite\", \"alphaear-predictor\", \"alphaear-stock\", \"alphaear-sentiment\", \"hf-models\", \"mirofish-financial-sim\", \"sre-devops-expert\", \"trading-market-environment-analysis\", \"ai-workflow-integrator\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0071",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Problem Definition. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "problem-definition",
        "diagnose",
        "role-developer",
        "ship",
        "test-suite",
        "deep-review",
        "codebase-archaeologist",
        "incident-to-improvement",
        "kwp-engineering-code-review",
        "kwp-engineering-tech-debt",
        "kwp-design-design-system-management",
        "kwp-design-design-critique",
        "kwp-design-accessibility-review",
        "kwp-design-user-research",
        "ecc-verification-loop",
        "role-ceo",
        "role-pm",
        "role-cso",
        "role-cto",
        "ui-suite",
        "refactor-simulator",
        "agency-visual-storyteller",
        "agency-ai-engineer",
        "agency-twitter-engager",
        "code-review-all"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "problem-definition",
          "tokens": 2601
        },
        {
          "skill_id": "diagnose",
          "tokens": 1433
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "deep-review",
          "tokens": 1923
        },
        {
          "skill_id": "codebase-archaeologist",
          "tokens": 1977
        },
        {
          "skill_id": "incident-to-improvement",
          "tokens": 2624
        },
        {
          "skill_id": "kwp-engineering-code-review",
          "tokens": 610
        },
        {
          "skill_id": "kwp-engineering-tech-debt",
          "tokens": 648
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "kwp-design-design-critique",
          "tokens": 671
        },
        {
          "skill_id": "kwp-design-accessibility-review",
          "tokens": 679
        },
        {
          "skill_id": "kwp-design-user-research",
          "tokens": 718
        },
        {
          "skill_id": "ecc-verification-loop",
          "tokens": 886
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "role-pm",
          "tokens": 1304
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        },
        {
          "skill_id": "role-cto",
          "tokens": 1413
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "refactor-simulator",
          "tokens": 2044
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "code-review-all",
          "tokens": 2524
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Problem Definition. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"problem-definition\", \"diagnose\", \"role-developer\", \"ship\", \"test-suite\", \"deep-review\", \"codebase-archaeologist\", \"incident-to-improvement\", \"kwp-engineering-code-review\", \"kwp-engineering-tech-debt\", \"kwp-design-design-system-management\", \"kwp-design-design-critique\", \"kwp-design-accessibility-review\", \"kwp-design-user-research\", \"ecc-verification-loop\", \"role-ceo\", \"role-pm\", \"role-cso\", \"role-cto\", \"ui-suite\", \"refactor-simulator\", \"agency-visual-storyteller\", \"agency-ai-engineer\", \"agency-twitter-engager\", \"code-review-all\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0072",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: SEFO Orchestrator. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "sefo-orchestrator",
        "sefo-federation",
        "sefo-governance",
        "mission-control copy",
        "mission-control",
        "scrapling",
        "ai-workflow-integrator",
        "ralph-loop",
        "plans",
        "agency-studio-producer",
        "ecc-agentic-engineering",
        "ai-chief-of-staff",
        "workflow-sequential",
        "issue-to-code",
        "paperclip-setup",
        "prompt-transformer",
        "paperclip-agents",
        "hf-auto-finetune-planner",
        "full-stack-planner",
        "ecc-continuous-learning",
        "autoskill-merger",
        "paper-auto-classifier",
        "paperclip-control",
        "knowledge-daily-aggregator",
        "air-gap-orchestrator"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "sefo-orchestrator",
          "tokens": 821
        },
        {
          "skill_id": "sefo-federation",
          "tokens": 756
        },
        {
          "skill_id": "sefo-governance",
          "tokens": 837
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "scrapling",
          "tokens": 1675
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "ralph-loop",
          "tokens": 1925
        },
        {
          "skill_id": "plans",
          "tokens": 2297
        },
        {
          "skill_id": "agency-studio-producer",
          "tokens": 2953
        },
        {
          "skill_id": "ecc-agentic-engineering",
          "tokens": 690
        },
        {
          "skill_id": "ai-chief-of-staff",
          "tokens": 1152
        },
        {
          "skill_id": "workflow-sequential",
          "tokens": 1485
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "paperclip-setup",
          "tokens": 1708
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "paperclip-agents",
          "tokens": 1887
        },
        {
          "skill_id": "hf-auto-finetune-planner",
          "tokens": 2091
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        },
        {
          "skill_id": "ecc-continuous-learning",
          "tokens": 3352
        },
        {
          "skill_id": "autoskill-merger",
          "tokens": 1142
        },
        {
          "skill_id": "paper-auto-classifier",
          "tokens": 1262
        },
        {
          "skill_id": "paperclip-control",
          "tokens": 1407
        },
        {
          "skill_id": "knowledge-daily-aggregator",
          "tokens": 1425
        },
        {
          "skill_id": "air-gap-orchestrator",
          "tokens": 1788
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: SEFO Orchestrator. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"sefo-orchestrator\", \"sefo-federation\", \"sefo-governance\", \"mission-control copy\", \"mission-control\", \"scrapling\", \"ai-workflow-integrator\", \"ralph-loop\", \"plans\", \"agency-studio-producer\", \"ecc-agentic-engineering\", \"ai-chief-of-staff\", \"workflow-sequential\", \"issue-to-code\", \"paperclip-setup\", \"prompt-transformer\", \"paperclip-agents\", \"hf-auto-finetune-planner\", \"full-stack-planner\", \"ecc-continuous-learning\", \"autoskill-merger\", \"paper-auto-classifier\", \"paperclip-control\", \"knowledge-daily-aggregator\", \"air-gap-orchestrator\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0073",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Workflow Sequential. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "workflow-sequential",
        "workflow-parallel",
        "ai-workflow-integrator",
        "mission-control",
        "ralph-loop",
        "workflow-eval-opt",
        "compliance-gate",
        "test-suite",
        "sefo-orchestrator",
        "mission-control copy",
        "auto-research",
        "sentence-polisher",
        "deep-research-pipeline",
        "code-review-all",
        "critical-review",
        "evals-skills",
        "ui-suite",
        "air-gap-orchestrator",
        "prompt-architect",
        "plans",
        "kwp-operations-process-optimization",
        "email-research-dispatcher",
        "ci-quality-gate",
        "ppt-template-engine",
        "docx-template-engine"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "workflow-sequential",
          "tokens": 1485
        },
        {
          "skill_id": "workflow-parallel",
          "tokens": 1734
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "ralph-loop",
          "tokens": 1925
        },
        {
          "skill_id": "workflow-eval-opt",
          "tokens": 2185
        },
        {
          "skill_id": "compliance-gate",
          "tokens": 1376
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "sefo-orchestrator",
          "tokens": 821
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "auto-research",
          "tokens": 2139
        },
        {
          "skill_id": "sentence-polisher",
          "tokens": 2167
        },
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        },
        {
          "skill_id": "code-review-all",
          "tokens": 2524
        },
        {
          "skill_id": "critical-review",
          "tokens": 3039
        },
        {
          "skill_id": "evals-skills",
          "tokens": 1524
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "air-gap-orchestrator",
          "tokens": 1788
        },
        {
          "skill_id": "prompt-architect",
          "tokens": 1939
        },
        {
          "skill_id": "plans",
          "tokens": 2297
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "email-research-dispatcher",
          "tokens": 703
        },
        {
          "skill_id": "ci-quality-gate",
          "tokens": 1054
        },
        {
          "skill_id": "ppt-template-engine",
          "tokens": 1101
        },
        {
          "skill_id": "docx-template-engine",
          "tokens": 1193
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Workflow Sequential. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"workflow-sequential\", \"workflow-parallel\", \"ai-workflow-integrator\", \"mission-control\", \"ralph-loop\", \"workflow-eval-opt\", \"compliance-gate\", \"test-suite\", \"sefo-orchestrator\", \"mission-control copy\", \"auto-research\", \"sentence-polisher\", \"deep-research-pipeline\", \"code-review-all\", \"critical-review\", \"evals-skills\", \"ui-suite\", \"air-gap-orchestrator\", \"prompt-architect\", \"plans\", \"kwp-operations-process-optimization\", \"email-research-dispatcher\", \"ci-quality-gate\", \"ppt-template-engine\", \"docx-template-engine\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0074",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: System Design. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-engineering-system-design",
        "kwp-engineering-testing-strategy",
        "kwp-design-design-system-management",
        "kwp-engineering-code-review",
        "kwp-engineering-tech-debt",
        "kwp-engineering-documentation",
        "kwp-design-design-critique",
        "kwp-design-ux-writing",
        "kwp-design-user-research",
        "kwp-operations-compliance-tracking",
        "kwp-operations-risk-assessment",
        "agency-backend-architect",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-engineering-incident-response",
        "kwp-design-accessibility-review",
        "backend-expert",
        "db-expert",
        "agency-ai-engineer",
        "agency-ux-architect",
        "agency-data-engineer",
        "kwp-operations-resource-planning",
        "role-ux-designer",
        "agency-visual-storyteller",
        "agency-experiment-tracker",
        "agency-security-engineer"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-engineering-system-design",
          "tokens": 756
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "kwp-engineering-code-review",
          "tokens": 610
        },
        {
          "skill_id": "kwp-engineering-tech-debt",
          "tokens": 648
        },
        {
          "skill_id": "kwp-engineering-documentation",
          "tokens": 654
        },
        {
          "skill_id": "kwp-design-design-critique",
          "tokens": 671
        },
        {
          "skill_id": "kwp-design-ux-writing",
          "tokens": 675
        },
        {
          "skill_id": "kwp-design-user-research",
          "tokens": 718
        },
        {
          "skill_id": "kwp-operations-compliance-tracking",
          "tokens": 686
        },
        {
          "skill_id": "kwp-operations-risk-assessment",
          "tokens": 808
        },
        {
          "skill_id": "agency-backend-architect",
          "tokens": 2600
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-engineering-incident-response",
          "tokens": 645
        },
        {
          "skill_id": "kwp-design-accessibility-review",
          "tokens": 679
        },
        {
          "skill_id": "backend-expert",
          "tokens": 1132
        },
        {
          "skill_id": "db-expert",
          "tokens": 1523
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-ux-architect",
          "tokens": 3119
        },
        {
          "skill_id": "agency-data-engineer",
          "tokens": 3843
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "role-ux-designer",
          "tokens": 1268
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-security-engineer",
          "tokens": 3168
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: System Design. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-engineering-system-design\", \"kwp-engineering-testing-strategy\", \"kwp-design-design-system-management\", \"kwp-engineering-code-review\", \"kwp-engineering-tech-debt\", \"kwp-engineering-documentation\", \"kwp-design-design-critique\", \"kwp-design-ux-writing\", \"kwp-design-user-research\", \"kwp-operations-compliance-tracking\", \"kwp-operations-risk-assessment\", \"agency-backend-architect\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-engineering-incident-response\", \"kwp-design-accessibility-review\", \"backend-expert\", \"db-expert\", \"agency-ai-engineer\", \"agency-ux-architect\", \"agency-data-engineer\", \"kwp-operations-resource-planning\", \"role-ux-designer\", \"agency-visual-storyteller\", \"agency-experiment-tracker\", \"agency-security-engineer\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0075",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Lead Lifecycle Manager. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "lead-lifecycle-manager",
        "sales-weekly-dashboard",
        "feature-intake-router",
        "planning-meeting-sentinel",
        "docs-freshness-guardian",
        "sales-meeting-pipeline",
        "prd-research-factory",
        "proposal-factory",
        "incident-lifecycle-manager",
        "tech-support-brain",
        "planning-weekly-pulse",
        "ui-suite",
        "kwp-sales-call-prep",
        "ship",
        "deep-research-pipeline",
        "dependency-radar",
        "smart-meeting-scheduler",
        "knowledge-to-paper",
        "role-sales",
        "ecc-strategic-compact",
        "knowledge-daily-aggregator",
        "proactive-meeting-scheduler",
        "unified-intel-intake",
        "ecc-search-first",
        "decision-router"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "lead-lifecycle-manager",
          "tokens": 533
        },
        {
          "skill_id": "sales-weekly-dashboard",
          "tokens": 558
        },
        {
          "skill_id": "feature-intake-router",
          "tokens": 568
        },
        {
          "skill_id": "planning-meeting-sentinel",
          "tokens": 604
        },
        {
          "skill_id": "docs-freshness-guardian",
          "tokens": 557
        },
        {
          "skill_id": "sales-meeting-pipeline",
          "tokens": 559
        },
        {
          "skill_id": "prd-research-factory",
          "tokens": 648
        },
        {
          "skill_id": "proposal-factory",
          "tokens": 582
        },
        {
          "skill_id": "incident-lifecycle-manager",
          "tokens": 610
        },
        {
          "skill_id": "tech-support-brain",
          "tokens": 620
        },
        {
          "skill_id": "planning-weekly-pulse",
          "tokens": 528
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        },
        {
          "skill_id": "dependency-radar",
          "tokens": 531
        },
        {
          "skill_id": "smart-meeting-scheduler",
          "tokens": 639
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "role-sales",
          "tokens": 1221
        },
        {
          "skill_id": "ecc-strategic-compact",
          "tokens": 1252
        },
        {
          "skill_id": "knowledge-daily-aggregator",
          "tokens": 1425
        },
        {
          "skill_id": "proactive-meeting-scheduler",
          "tokens": 1444
        },
        {
          "skill_id": "unified-intel-intake",
          "tokens": 1593
        },
        {
          "skill_id": "ecc-search-first",
          "tokens": 1597
        },
        {
          "skill_id": "decision-router",
          "tokens": 1730
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Lead Lifecycle Manager. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"lead-lifecycle-manager\", \"sales-weekly-dashboard\", \"feature-intake-router\", \"planning-meeting-sentinel\", \"docs-freshness-guardian\", \"sales-meeting-pipeline\", \"prd-research-factory\", \"proposal-factory\", \"incident-lifecycle-manager\", \"tech-support-brain\", \"planning-weekly-pulse\", \"ui-suite\", \"kwp-sales-call-prep\", \"ship\", \"deep-research-pipeline\", \"dependency-radar\", \"smart-meeting-scheduler\", \"knowledge-to-paper\", \"role-sales\", \"ecc-strategic-compact\", \"knowledge-daily-aggregator\", \"proactive-meeting-scheduler\", \"unified-intel-intake\", \"ecc-search-first\", \"decision-router\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0076",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: US Stock Analysis. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "trading-us-stock-analysis",
        "trading-technical-analyst",
        "alphaear-deepear-lite",
        "trading-scenario-analyzer",
        "trading-market-environment-analysis",
        "daily-stock-check",
        "trading-stanley-druckenmiller-investment",
        "trading-sector-analyst",
        "alphaear-stock",
        "stock-csv-downloader",
        "weekly-stock-update",
        "trading-backtest-expert",
        "trading-us-market-bubble-detector",
        "trading-finviz-screener",
        "trading-breadth-chart-analyst",
        "alphaear-reporter",
        "alphaear-sentiment",
        "terraform-reviewer",
        "ai-workflow-integrator",
        "ai-quality-evaluator",
        "mirofish-financial-sim",
        "role-trading-expert",
        "trading-market-breadth-analyzer",
        "trading-data-quality-checker",
        "trading-edge-candidate-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "trading-us-stock-analysis",
          "tokens": 2926
        },
        {
          "skill_id": "trading-technical-analyst",
          "tokens": 2868
        },
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        },
        {
          "skill_id": "trading-scenario-analyzer",
          "tokens": 2068
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "trading-stanley-druckenmiller-investment",
          "tokens": 2317
        },
        {
          "skill_id": "trading-sector-analyst",
          "tokens": 2657
        },
        {
          "skill_id": "alphaear-stock",
          "tokens": 820
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "trading-backtest-expert",
          "tokens": 2754
        },
        {
          "skill_id": "trading-us-market-bubble-detector",
          "tokens": 3062
        },
        {
          "skill_id": "trading-finviz-screener",
          "tokens": 3681
        },
        {
          "skill_id": "trading-breadth-chart-analyst",
          "tokens": 3912
        },
        {
          "skill_id": "alphaear-reporter",
          "tokens": 810
        },
        {
          "skill_id": "alphaear-sentiment",
          "tokens": 928
        },
        {
          "skill_id": "terraform-reviewer",
          "tokens": 1457
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "ai-quality-evaluator",
          "tokens": 2228
        },
        {
          "skill_id": "mirofish-financial-sim",
          "tokens": 1330
        },
        {
          "skill_id": "role-trading-expert",
          "tokens": 1791
        },
        {
          "skill_id": "trading-market-breadth-analyzer",
          "tokens": 1830
        },
        {
          "skill_id": "trading-data-quality-checker",
          "tokens": 1962
        },
        {
          "skill_id": "trading-edge-candidate-agent",
          "tokens": 1971
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: US Stock Analysis. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"trading-us-stock-analysis\", \"trading-technical-analyst\", \"alphaear-deepear-lite\", \"trading-scenario-analyzer\", \"trading-market-environment-analysis\", \"daily-stock-check\", \"trading-stanley-druckenmiller-investment\", \"trading-sector-analyst\", \"alphaear-stock\", \"stock-csv-downloader\", \"weekly-stock-update\", \"trading-backtest-expert\", \"trading-us-market-bubble-detector\", \"trading-finviz-screener\", \"trading-breadth-chart-analyst\", \"alphaear-reporter\", \"alphaear-sentiment\", \"terraform-reviewer\", \"ai-workflow-integrator\", \"ai-quality-evaluator\", \"mirofish-financial-sim\", \"role-trading-expert\", \"trading-market-breadth-analyzer\", \"trading-data-quality-checker\", \"trading-edge-candidate-agent\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0077",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AlphaEar Signal Tracker. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "alphaear-signal-tracker",
        "alphaear-news",
        "alphaear-deepear-lite",
        "trading-market-environment-analysis",
        "alphaear-sentiment",
        "tab-report-generate",
        "alphaear-search",
        "alphaear-predictor",
        "alphaear-reporter",
        "hf-evaluation",
        "stock-csv-downloader",
        "daily-stock-check",
        "trading-scenario-analyzer",
        "alphaear-stock",
        "weekly-stock-update",
        "trading-sector-analyst",
        "trading-technical-analyst",
        "trading-us-stock-analysis",
        "trading-breadth-chart-analyst",
        "hf-model-trainer",
        "tab-kiwoom",
        "tab-news-fetch",
        "tab-market-environment",
        "alphaear-logic-visualizer",
        "hf-trackio"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "alphaear-signal-tracker",
          "tokens": 854
        },
        {
          "skill_id": "alphaear-news",
          "tokens": 873
        },
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "alphaear-sentiment",
          "tokens": 928
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "alphaear-search",
          "tokens": 672
        },
        {
          "skill_id": "alphaear-predictor",
          "tokens": 803
        },
        {
          "skill_id": "alphaear-reporter",
          "tokens": 810
        },
        {
          "skill_id": "hf-evaluation",
          "tokens": 4549
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "trading-scenario-analyzer",
          "tokens": 2068
        },
        {
          "skill_id": "alphaear-stock",
          "tokens": 820
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "trading-sector-analyst",
          "tokens": 2657
        },
        {
          "skill_id": "trading-technical-analyst",
          "tokens": 2868
        },
        {
          "skill_id": "trading-us-stock-analysis",
          "tokens": 2926
        },
        {
          "skill_id": "trading-breadth-chart-analyst",
          "tokens": 3912
        },
        {
          "skill_id": "hf-model-trainer",
          "tokens": 5150
        },
        {
          "skill_id": "tab-kiwoom",
          "tokens": 452
        },
        {
          "skill_id": "tab-news-fetch",
          "tokens": 474
        },
        {
          "skill_id": "tab-market-environment",
          "tokens": 566
        },
        {
          "skill_id": "alphaear-logic-visualizer",
          "tokens": 740
        },
        {
          "skill_id": "hf-trackio",
          "tokens": 1424
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AlphaEar Signal Tracker. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-signal-tracker\", \"alphaear-news\", \"alphaear-deepear-lite\", \"trading-market-environment-analysis\", \"alphaear-sentiment\", \"tab-report-generate\", \"alphaear-search\", \"alphaear-predictor\", \"alphaear-reporter\", \"hf-evaluation\", \"stock-csv-downloader\", \"daily-stock-check\", \"trading-scenario-analyzer\", \"alphaear-stock\", \"weekly-stock-update\", \"trading-sector-analyst\", \"trading-technical-analyst\", \"trading-us-stock-analysis\", \"trading-breadth-chart-analyst\", \"hf-model-trainer\", \"tab-kiwoom\", \"tab-news-fetch\", \"tab-market-environment\", \"alphaear-logic-visualizer\", \"hf-trackio\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0078",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Hugging Face Inference Endpoints. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "hf-endpoints",
        "hf-buckets",
        "hf-spaces",
        "hf-models",
        "hf-collections",
        "hf-discussions",
        "hf-repos",
        "hf-cli",
        "hf-papers",
        "hf-hub",
        "gws-chat",
        "gws-drive",
        "hf-dataset-viewer",
        "hf-trackio",
        "hf-model-arena",
        "hf-model-trainer",
        "gws-docs",
        "gws-sheets",
        "gws-calendar",
        "gws-gmail",
        "hf-jobs",
        "hf-evaluation",
        "gws-workflows",
        "backend-expert",
        "performance-profiler"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "hf-endpoints",
          "tokens": 1177
        },
        {
          "skill_id": "hf-buckets",
          "tokens": 1334
        },
        {
          "skill_id": "hf-spaces",
          "tokens": 930
        },
        {
          "skill_id": "hf-models",
          "tokens": 1212
        },
        {
          "skill_id": "hf-collections",
          "tokens": 1239
        },
        {
          "skill_id": "hf-discussions",
          "tokens": 1118
        },
        {
          "skill_id": "hf-repos",
          "tokens": 1106
        },
        {
          "skill_id": "hf-cli",
          "tokens": 2275
        },
        {
          "skill_id": "hf-papers",
          "tokens": 817
        },
        {
          "skill_id": "hf-hub",
          "tokens": 1151
        },
        {
          "skill_id": "gws-chat",
          "tokens": 751
        },
        {
          "skill_id": "gws-drive",
          "tokens": 997
        },
        {
          "skill_id": "hf-dataset-viewer",
          "tokens": 1387
        },
        {
          "skill_id": "hf-trackio",
          "tokens": 1424
        },
        {
          "skill_id": "hf-model-arena",
          "tokens": 2232
        },
        {
          "skill_id": "hf-model-trainer",
          "tokens": 5150
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "gws-sheets",
          "tokens": 918
        },
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "gws-gmail",
          "tokens": 1796
        },
        {
          "skill_id": "hf-jobs",
          "tokens": 4467
        },
        {
          "skill_id": "hf-evaluation",
          "tokens": 4549
        },
        {
          "skill_id": "gws-workflows",
          "tokens": 952
        },
        {
          "skill_id": "backend-expert",
          "tokens": 1132
        },
        {
          "skill_id": "performance-profiler",
          "tokens": 1738
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Hugging Face Inference Endpoints. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"hf-endpoints\", \"hf-buckets\", \"hf-spaces\", \"hf-models\", \"hf-collections\", \"hf-discussions\", \"hf-repos\", \"hf-cli\", \"hf-papers\", \"hf-hub\", \"gws-chat\", \"gws-drive\", \"hf-dataset-viewer\", \"hf-trackio\", \"hf-model-arena\", \"hf-model-trainer\", \"gws-docs\", \"gws-sheets\", \"gws-calendar\", \"gws-gmail\", \"hf-jobs\", \"hf-evaluation\", \"gws-workflows\", \"backend-expert\", \"performance-profiler\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0079",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Contract Review Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-legal-contract-review",
        "kwp-legal-compliance",
        "kwp-legal-meeting-briefing",
        "kwp-customer-support-customer-research",
        "kwp-marketing-content-creation",
        "kwp-legal-nda-triage",
        "kwp-operations-vendor-management",
        "kwp-sales-daily-briefing",
        "kwp-sales-account-research",
        "kwp-sales-call-prep",
        "kwp-data-data-validation",
        "kwp-customer-support-response-drafting",
        "kwp-legal-canned-responses",
        "kwp-legal-legal-risk-assessment",
        "kwp-marketing-performance-analytics",
        "kwp-human-resources-people-analytics",
        "kwp-apollo-prospect",
        "mission-control copy",
        "kwp-finance-journal-entry-prep",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-customer-support-escalation",
        "kwp-data-interactive-dashboard-builder",
        "kwp-data-data-exploration",
        "mission-control",
        "kwp-marketing-brand-voice"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-legal-contract-review",
          "tokens": 3137
        },
        {
          "skill_id": "kwp-legal-compliance",
          "tokens": 3473
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-customer-support-customer-research",
          "tokens": 2529
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "kwp-legal-nda-triage",
          "tokens": 3139
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-data-data-validation",
          "tokens": 2650
        },
        {
          "skill_id": "kwp-customer-support-response-drafting",
          "tokens": 2944
        },
        {
          "skill_id": "kwp-legal-canned-responses",
          "tokens": 3164
        },
        {
          "skill_id": "kwp-legal-legal-risk-assessment",
          "tokens": 3406
        },
        {
          "skill_id": "kwp-marketing-performance-analytics",
          "tokens": 3966
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-apollo-prospect",
          "tokens": 1006
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "kwp-finance-journal-entry-prep",
          "tokens": 2215
        },
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "kwp-customer-support-escalation",
          "tokens": 2329
        },
        {
          "skill_id": "kwp-data-interactive-dashboard-builder",
          "tokens": 2350
        },
        {
          "skill_id": "kwp-data-data-exploration",
          "tokens": 2358
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "kwp-marketing-brand-voice",
          "tokens": 2392
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Contract Review Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-legal-contract-review\", \"kwp-legal-compliance\", \"kwp-legal-meeting-briefing\", \"kwp-customer-support-customer-research\", \"kwp-marketing-content-creation\", \"kwp-legal-nda-triage\", \"kwp-operations-vendor-management\", \"kwp-sales-daily-briefing\", \"kwp-sales-account-research\", \"kwp-sales-call-prep\", \"kwp-data-data-validation\", \"kwp-customer-support-response-drafting\", \"kwp-legal-canned-responses\", \"kwp-legal-legal-risk-assessment\", \"kwp-marketing-performance-analytics\", \"kwp-human-resources-people-analytics\", \"kwp-apollo-prospect\", \"mission-control copy\", \"kwp-finance-journal-entry-prep\", \"kwp-bio-research-single-cell-rna-qc\", \"kwp-customer-support-escalation\", \"kwp-data-interactive-dashboard-builder\", \"kwp-data-data-exploration\", \"mission-control\", \"kwp-marketing-brand-voice\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0080",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Overview. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "hf-evaluation",
        "hf-trackio",
        "hf-model-trainer",
        "hf-cli",
        "hf-jobs",
        "ecc-eval-harness",
        "hf-datasets",
        "hf-dataset-viewer",
        "alphaear-deepear-lite",
        "alphaear-signal-tracker",
        "alphaear-sentiment",
        "hf-models",
        "alphaear-predictor",
        "alphaear-news",
        "hf-endpoints",
        "weekly-stock-update",
        "ai-quality-evaluator",
        "trading-us-stock-analysis",
        "agency-test-results-analyzer",
        "anthropic-skill-creator",
        "tab-kiwoom",
        "tab-report-generate",
        "alphaear-reporter",
        "alphaear-stock",
        "stock-csv-downloader"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "hf-evaluation",
          "tokens": 4549
        },
        {
          "skill_id": "hf-trackio",
          "tokens": 1424
        },
        {
          "skill_id": "hf-model-trainer",
          "tokens": 5150
        },
        {
          "skill_id": "hf-cli",
          "tokens": 2275
        },
        {
          "skill_id": "hf-jobs",
          "tokens": 4467
        },
        {
          "skill_id": "ecc-eval-harness",
          "tokens": 1810
        },
        {
          "skill_id": "hf-datasets",
          "tokens": 3241
        },
        {
          "skill_id": "hf-dataset-viewer",
          "tokens": 1387
        },
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        },
        {
          "skill_id": "alphaear-signal-tracker",
          "tokens": 854
        },
        {
          "skill_id": "alphaear-sentiment",
          "tokens": 928
        },
        {
          "skill_id": "hf-models",
          "tokens": 1212
        },
        {
          "skill_id": "alphaear-predictor",
          "tokens": 803
        },
        {
          "skill_id": "alphaear-news",
          "tokens": 873
        },
        {
          "skill_id": "hf-endpoints",
          "tokens": 1177
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "ai-quality-evaluator",
          "tokens": 2228
        },
        {
          "skill_id": "trading-us-stock-analysis",
          "tokens": 2926
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "anthropic-skill-creator",
          "tokens": 6862
        },
        {
          "skill_id": "tab-kiwoom",
          "tokens": 452
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "alphaear-reporter",
          "tokens": 810
        },
        {
          "skill_id": "alphaear-stock",
          "tokens": 820
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Overview. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"hf-evaluation\", \"hf-trackio\", \"hf-model-trainer\", \"hf-cli\", \"hf-jobs\", \"ecc-eval-harness\", \"hf-datasets\", \"hf-dataset-viewer\", \"alphaear-deepear-lite\", \"alphaear-signal-tracker\", \"alphaear-sentiment\", \"hf-models\", \"alphaear-predictor\", \"alphaear-news\", \"hf-endpoints\", \"weekly-stock-update\", \"ai-quality-evaluator\", \"trading-us-stock-analysis\", \"agency-test-results-analyzer\", \"anthropic-skill-creator\", \"tab-kiwoom\", \"tab-report-generate\", \"alphaear-reporter\", \"alphaear-stock\", \"stock-csv-downloader\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0081",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Finance Tracker Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-finance-tracker",
        "agency-growth-hacker",
        "agency-infrastructure-maintainer",
        "agency-trend-researcher",
        "agency-studio-producer",
        "agency-feedback-synthesizer",
        "agency-experiment-tracker",
        "agency-project-shepherd",
        "agency-brand-guardian",
        "agency-content-creator",
        "agency-ai-engineer",
        "agency-tiktok-strategist",
        "agency-twitter-engager",
        "agency-autonomous-optimization-architect",
        "agency-xiaohongshu-specialist",
        "agency-studio-operations",
        "agency-wechat-official-account-manager",
        "agency-workflow-optimizer",
        "agency-api-tester",
        "agency-legal-compliance-checker",
        "agency-data-engineer",
        "agency-developer-advocate",
        "agency-tool-evaluator",
        "agency-xr-immersive-developer",
        "agency-data-analytics-reporter"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-finance-tracker",
          "tokens": 3177
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "agency-studio-producer",
          "tokens": 2953
        },
        {
          "skill_id": "agency-feedback-synthesizer",
          "tokens": 2041
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-brand-guardian",
          "tokens": 3135
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-autonomous-optimization-architect",
          "tokens": 2220
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-data-engineer",
          "tokens": 3843
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "agency-tool-evaluator",
          "tokens": 4549
        },
        {
          "skill_id": "agency-xr-immersive-developer",
          "tokens": 697
        },
        {
          "skill_id": "agency-data-analytics-reporter",
          "tokens": 1003
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Finance Tracker Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-finance-tracker\", \"agency-growth-hacker\", \"agency-infrastructure-maintainer\", \"agency-trend-researcher\", \"agency-studio-producer\", \"agency-feedback-synthesizer\", \"agency-experiment-tracker\", \"agency-project-shepherd\", \"agency-brand-guardian\", \"agency-content-creator\", \"agency-ai-engineer\", \"agency-tiktok-strategist\", \"agency-twitter-engager\", \"agency-autonomous-optimization-architect\", \"agency-xiaohongshu-specialist\", \"agency-studio-operations\", \"agency-wechat-official-account-manager\", \"agency-workflow-optimizer\", \"agency-api-tester\", \"agency-legal-compliance-checker\", \"agency-data-engineer\", \"agency-developer-advocate\", \"agency-tool-evaluator\", \"agency-xr-immersive-developer\", \"agency-data-analytics-reporter\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0082",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Sales Perspective Analyzer. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "role-sales",
        "role-ux-designer",
        "role-hr",
        "role-cso",
        "role-finance",
        "role-developer",
        "role-ceo",
        "role-pm",
        "role-security-engineer",
        "role-cto",
        "role-data-scientist",
        "role-trading-expert",
        "deep-research-pipeline",
        "executive-briefing",
        "paper-review",
        "full-stack-planner",
        "role-dispatcher",
        "prd-research-factory",
        "pm-market-research",
        "pm-go-to-market",
        "kwp-brand-voice-guideline-generation",
        "nlm-launch-kit",
        "meeting-digest",
        "kwp-engineering-documentation",
        "agency-data-consolidation-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "role-sales",
          "tokens": 1221
        },
        {
          "skill_id": "role-ux-designer",
          "tokens": 1268
        },
        {
          "skill_id": "role-hr",
          "tokens": 1350
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        },
        {
          "skill_id": "role-finance",
          "tokens": 1353
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "role-pm",
          "tokens": 1304
        },
        {
          "skill_id": "role-security-engineer",
          "tokens": 1406
        },
        {
          "skill_id": "role-cto",
          "tokens": 1413
        },
        {
          "skill_id": "role-data-scientist",
          "tokens": 1695
        },
        {
          "skill_id": "role-trading-expert",
          "tokens": 1791
        },
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        },
        {
          "skill_id": "executive-briefing",
          "tokens": 1419
        },
        {
          "skill_id": "paper-review",
          "tokens": 5815
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        },
        {
          "skill_id": "role-dispatcher",
          "tokens": 2075
        },
        {
          "skill_id": "prd-research-factory",
          "tokens": 648
        },
        {
          "skill_id": "pm-market-research",
          "tokens": 966
        },
        {
          "skill_id": "pm-go-to-market",
          "tokens": 992
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "nlm-launch-kit",
          "tokens": 2368
        },
        {
          "skill_id": "meeting-digest",
          "tokens": 4491
        },
        {
          "skill_id": "kwp-engineering-documentation",
          "tokens": 654
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Sales Perspective Analyzer. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"role-sales\", \"role-ux-designer\", \"role-hr\", \"role-cso\", \"role-finance\", \"role-developer\", \"role-ceo\", \"role-pm\", \"role-security-engineer\", \"role-cto\", \"role-data-scientist\", \"role-trading-expert\", \"deep-research-pipeline\", \"executive-briefing\", \"paper-review\", \"full-stack-planner\", \"role-dispatcher\", \"prd-research-factory\", \"pm-market-research\", \"pm-go-to-market\", \"kwp-brand-voice-guideline-generation\", \"nlm-launch-kit\", \"meeting-digest\", \"kwp-engineering-documentation\", \"agency-data-consolidation-agent\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0083",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Building LLM-Powered Applications with Claude. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "anthropic-claude-api",
        "agency-mobile-app-builder",
        "agency-ai-engineer",
        "agency-rapid-prototyper",
        "agency-backend-architect",
        "anthropic-mcp-builder",
        "agency-xr-immersive-developer",
        "agency-data-consolidation-agent",
        "hf-spaces",
        "gws-workspace",
        "ui-suite",
        "agency-instagram-curator",
        "agency-visual-storyteller",
        "agency-reddit-community-builder",
        "slack-agent",
        "agency-studio-operations",
        "agency-studio-producer",
        "agency-lsp-index-engineer",
        "agency-ux-architect",
        "agency-brand-guardian",
        "agency-infrastructure-maintainer",
        "agency-api-tester",
        "agency-legal-compliance-checker",
        "agency-developer-advocate",
        "agency-agents-orchestrator"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "anthropic-claude-api",
          "tokens": 4709
        },
        {
          "skill_id": "agency-mobile-app-builder",
          "tokens": 2595
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-rapid-prototyper",
          "tokens": 2496
        },
        {
          "skill_id": "agency-backend-architect",
          "tokens": 2600
        },
        {
          "skill_id": "anthropic-mcp-builder",
          "tokens": 2825
        },
        {
          "skill_id": "agency-xr-immersive-developer",
          "tokens": 697
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        },
        {
          "skill_id": "hf-spaces",
          "tokens": 930
        },
        {
          "skill_id": "gws-workspace",
          "tokens": 1186
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "slack-agent",
          "tokens": 2564
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-studio-producer",
          "tokens": 2953
        },
        {
          "skill_id": "agency-lsp-index-engineer",
          "tokens": 3006
        },
        {
          "skill_id": "agency-ux-architect",
          "tokens": 3119
        },
        {
          "skill_id": "agency-brand-guardian",
          "tokens": 3135
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "agency-agents-orchestrator",
          "tokens": 4128
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Building LLM-Powered Applications with Claude. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-claude-api\", \"agency-mobile-app-builder\", \"agency-ai-engineer\", \"agency-rapid-prototyper\", \"agency-backend-architect\", \"anthropic-mcp-builder\", \"agency-xr-immersive-developer\", \"agency-data-consolidation-agent\", \"hf-spaces\", \"gws-workspace\", \"ui-suite\", \"agency-instagram-curator\", \"agency-visual-storyteller\", \"agency-reddit-community-builder\", \"slack-agent\", \"agency-studio-operations\", \"agency-studio-producer\", \"agency-lsp-index-engineer\", \"agency-ux-architect\", \"agency-brand-guardian\", \"agency-infrastructure-maintainer\", \"agency-api-tester\", \"agency-legal-compliance-checker\", \"agency-developer-advocate\", \"agency-agents-orchestrator\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0084",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Breadth Chart Analyst. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "trading-breadth-chart-analyst",
        "trading-market-environment-analysis",
        "trading-uptrend-analyzer",
        "trading-sector-analyst",
        "trading-stanley-druckenmiller-investment",
        "alphaear-deepear-lite",
        "trading-market-breadth-analyzer",
        "trading-market-top-detector",
        "trading-technical-analyst",
        "trading-us-stock-analysis",
        "trading-theme-detector",
        "daily-strategy-post",
        "trading-trade-hypothesis-ideator",
        "alphaear-sentiment",
        "role-ceo",
        "trading-edge-candidate-agent",
        "stock-csv-downloader",
        "mirofish-financial-sim",
        "trading-market-news-analyst",
        "weekly-stock-update",
        "daily-stock-check",
        "role-trading-expert",
        "trading-scenario-analyzer",
        "trading-finviz-screener",
        "alphaear-search"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "trading-breadth-chart-analyst",
          "tokens": 3912
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "trading-uptrend-analyzer",
          "tokens": 1780
        },
        {
          "skill_id": "trading-sector-analyst",
          "tokens": 2657
        },
        {
          "skill_id": "trading-stanley-druckenmiller-investment",
          "tokens": 2317
        },
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        },
        {
          "skill_id": "trading-market-breadth-analyzer",
          "tokens": 1830
        },
        {
          "skill_id": "trading-market-top-detector",
          "tokens": 2295
        },
        {
          "skill_id": "trading-technical-analyst",
          "tokens": 2868
        },
        {
          "skill_id": "trading-us-stock-analysis",
          "tokens": 2926
        },
        {
          "skill_id": "trading-theme-detector",
          "tokens": 3487
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "trading-trade-hypothesis-ideator",
          "tokens": 909
        },
        {
          "skill_id": "alphaear-sentiment",
          "tokens": 928
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "trading-edge-candidate-agent",
          "tokens": 1971
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "mirofish-financial-sim",
          "tokens": 1330
        },
        {
          "skill_id": "trading-market-news-analyst",
          "tokens": 1467
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "role-trading-expert",
          "tokens": 1791
        },
        {
          "skill_id": "trading-scenario-analyzer",
          "tokens": 2068
        },
        {
          "skill_id": "trading-finviz-screener",
          "tokens": 3681
        },
        {
          "skill_id": "alphaear-search",
          "tokens": 672
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Breadth Chart Analyst. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"trading-breadth-chart-analyst\", \"trading-market-environment-analysis\", \"trading-uptrend-analyzer\", \"trading-sector-analyst\", \"trading-stanley-druckenmiller-investment\", \"alphaear-deepear-lite\", \"trading-market-breadth-analyzer\", \"trading-market-top-detector\", \"trading-technical-analyst\", \"trading-us-stock-analysis\", \"trading-theme-detector\", \"daily-strategy-post\", \"trading-trade-hypothesis-ideator\", \"alphaear-sentiment\", \"role-ceo\", \"trading-edge-candidate-agent\", \"stock-csv-downloader\", \"mirofish-financial-sim\", \"trading-market-news-analyst\", \"weekly-stock-update\", \"daily-stock-check\", \"role-trading-expert\", \"trading-scenario-analyzer\", \"trading-finviz-screener\", \"alphaear-search\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0085",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Google Calendar. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "gws-calendar",
        "smart-meeting-scheduler",
        "gws-gmail",
        "gws-workflows",
        "gws-chat",
        "proactive-meeting-scheduler",
        "calendar-daily-briefing",
        "feedback-meeting-scheduler",
        "gws-drive",
        "gws-workspace",
        "gws-recipe-personalized-emails",
        "gws-recipe-events-from-sheet",
        "gws-docs",
        "gws-sheets",
        "ai-chief-of-staff",
        "gws-email-reply",
        "email-auto-reply",
        "gws-recipe-audit-sharing",
        "hf-discussions",
        "hf-models",
        "hf-collections",
        "hf-buckets",
        "gmail-daily-triage",
        "kwp-sales-call-prep",
        "cursor-automations"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "smart-meeting-scheduler",
          "tokens": 639
        },
        {
          "skill_id": "gws-gmail",
          "tokens": 1796
        },
        {
          "skill_id": "gws-workflows",
          "tokens": 952
        },
        {
          "skill_id": "gws-chat",
          "tokens": 751
        },
        {
          "skill_id": "proactive-meeting-scheduler",
          "tokens": 1444
        },
        {
          "skill_id": "calendar-daily-briefing",
          "tokens": 1042
        },
        {
          "skill_id": "feedback-meeting-scheduler",
          "tokens": 1238
        },
        {
          "skill_id": "gws-drive",
          "tokens": 997
        },
        {
          "skill_id": "gws-workspace",
          "tokens": 1186
        },
        {
          "skill_id": "gws-recipe-personalized-emails",
          "tokens": 641
        },
        {
          "skill_id": "gws-recipe-events-from-sheet",
          "tokens": 674
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "gws-sheets",
          "tokens": 918
        },
        {
          "skill_id": "ai-chief-of-staff",
          "tokens": 1152
        },
        {
          "skill_id": "gws-email-reply",
          "tokens": 2222
        },
        {
          "skill_id": "email-auto-reply",
          "tokens": 638
        },
        {
          "skill_id": "gws-recipe-audit-sharing",
          "tokens": 645
        },
        {
          "skill_id": "hf-discussions",
          "tokens": 1118
        },
        {
          "skill_id": "hf-models",
          "tokens": 1212
        },
        {
          "skill_id": "hf-collections",
          "tokens": 1239
        },
        {
          "skill_id": "hf-buckets",
          "tokens": 1334
        },
        {
          "skill_id": "gmail-daily-triage",
          "tokens": 2229
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "cursor-automations",
          "tokens": 2911
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Google Calendar. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"gws-calendar\", \"smart-meeting-scheduler\", \"gws-gmail\", \"gws-workflows\", \"gws-chat\", \"proactive-meeting-scheduler\", \"calendar-daily-briefing\", \"feedback-meeting-scheduler\", \"gws-drive\", \"gws-workspace\", \"gws-recipe-personalized-emails\", \"gws-recipe-events-from-sheet\", \"gws-docs\", \"gws-sheets\", \"ai-chief-of-staff\", \"gws-email-reply\", \"email-auto-reply\", \"gws-recipe-audit-sharing\", \"hf-discussions\", \"hf-models\", \"hf-collections\", \"hf-buckets\", \"gmail-daily-triage\", \"kwp-sales-call-prep\", \"cursor-automations\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0086",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Performance Analytics Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-marketing-performance-analytics",
        "kwp-marketing-campaign-planning",
        "kwp-marketing-content-creation",
        "kwp-marketing-competitive-analysis",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-people-analytics",
        "kwp-data-statistical-analysis",
        "kwp-operations-resource-planning",
        "kwp-sales-daily-briefing",
        "kwp-data-interactive-dashboard-builder",
        "kwp-data-data-exploration",
        "kwp-data-data-validation",
        "kwp-finance-financial-statements",
        "kwp-product-management-competitive-analysis",
        "kwp-finance-variance-analysis",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-operations-process-optimization",
        "kwp-sales-call-prep",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-customer-support-escalation",
        "kwp-marketing-brand-voice",
        "kwp-product-management-feature-spec",
        "kwp-finance-close-management",
        "kwp-customer-support-response-drafting"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-marketing-performance-analytics",
          "tokens": 3966
        },
        {
          "skill_id": "kwp-marketing-campaign-planning",
          "tokens": 2803
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "kwp-marketing-competitive-analysis",
          "tokens": 2723
        },
        {
          "skill_id": "kwp-product-management-metrics-tracking",
          "tokens": 3606
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-data-statistical-analysis",
          "tokens": 2865
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "kwp-data-interactive-dashboard-builder",
          "tokens": 2350
        },
        {
          "skill_id": "kwp-data-data-exploration",
          "tokens": 2358
        },
        {
          "skill_id": "kwp-data-data-validation",
          "tokens": 2650
        },
        {
          "skill_id": "kwp-finance-financial-statements",
          "tokens": 2768
        },
        {
          "skill_id": "kwp-product-management-competitive-analysis",
          "tokens": 2899
        },
        {
          "skill_id": "kwp-finance-variance-analysis",
          "tokens": 3208
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "kwp-customer-support-escalation",
          "tokens": 2329
        },
        {
          "skill_id": "kwp-marketing-brand-voice",
          "tokens": 2392
        },
        {
          "skill_id": "kwp-product-management-feature-spec",
          "tokens": 2614
        },
        {
          "skill_id": "kwp-finance-close-management",
          "tokens": 2782
        },
        {
          "skill_id": "kwp-customer-support-response-drafting",
          "tokens": 2944
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Performance Analytics Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-marketing-performance-analytics\", \"kwp-marketing-campaign-planning\", \"kwp-marketing-content-creation\", \"kwp-marketing-competitive-analysis\", \"kwp-product-management-metrics-tracking\", \"kwp-human-resources-people-analytics\", \"kwp-data-statistical-analysis\", \"kwp-operations-resource-planning\", \"kwp-sales-daily-briefing\", \"kwp-data-interactive-dashboard-builder\", \"kwp-data-data-exploration\", \"kwp-data-data-validation\", \"kwp-finance-financial-statements\", \"kwp-product-management-competitive-analysis\", \"kwp-finance-variance-analysis\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-operations-process-optimization\", \"kwp-sales-call-prep\", \"kwp-bio-research-single-cell-rna-qc\", \"kwp-customer-support-escalation\", \"kwp-marketing-brand-voice\", \"kwp-product-management-feature-spec\", \"kwp-finance-close-management\", \"kwp-customer-support-response-drafting\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0087",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-genai-features. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "tab-genai-features",
        "tab-llm-agents",
        "tab-report-generate",
        "tab-stock-sync",
        "tab-analysis-run",
        "tab-screening",
        "tab-event-detect",
        "tab-dashboard-summary",
        "tab-patterns",
        "tab-strategy-comparison",
        "mirofish-financial-sim",
        "ai-workflow-integrator",
        "morning-ship",
        "mirofish",
        "tab-dualma-refresh",
        "tab-technical-analysis",
        "i18n-sync",
        "weekly-stock-update",
        "daily-stock-check",
        "ai-quality-evaluator",
        "bespin-news-digest",
        "tab-fundamental-sync",
        "tab-hot-stock-discovery",
        "tab-turtle-refresh",
        "tab-bollinger-refresh"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "tab-genai-features",
          "tokens": 354
        },
        {
          "skill_id": "tab-llm-agents",
          "tokens": 382
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "tab-stock-sync",
          "tokens": 345
        },
        {
          "skill_id": "tab-analysis-run",
          "tokens": 404
        },
        {
          "skill_id": "tab-screening",
          "tokens": 410
        },
        {
          "skill_id": "tab-event-detect",
          "tokens": 319
        },
        {
          "skill_id": "tab-dashboard-summary",
          "tokens": 419
        },
        {
          "skill_id": "tab-patterns",
          "tokens": 509
        },
        {
          "skill_id": "tab-strategy-comparison",
          "tokens": 571
        },
        {
          "skill_id": "mirofish-financial-sim",
          "tokens": 1330
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "morning-ship",
          "tokens": 2943
        },
        {
          "skill_id": "mirofish",
          "tokens": 3657
        },
        {
          "skill_id": "tab-dualma-refresh",
          "tokens": 391
        },
        {
          "skill_id": "tab-technical-analysis",
          "tokens": 414
        },
        {
          "skill_id": "i18n-sync",
          "tokens": 1282
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "ai-quality-evaluator",
          "tokens": 2228
        },
        {
          "skill_id": "bespin-news-digest",
          "tokens": 3949
        },
        {
          "skill_id": "tab-fundamental-sync",
          "tokens": 331
        },
        {
          "skill_id": "tab-hot-stock-discovery",
          "tokens": 359
        },
        {
          "skill_id": "tab-turtle-refresh",
          "tokens": 371
        },
        {
          "skill_id": "tab-bollinger-refresh",
          "tokens": 383
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-genai-features. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"tab-genai-features\", \"tab-llm-agents\", \"tab-report-generate\", \"tab-stock-sync\", \"tab-analysis-run\", \"tab-screening\", \"tab-event-detect\", \"tab-dashboard-summary\", \"tab-patterns\", \"tab-strategy-comparison\", \"mirofish-financial-sim\", \"ai-workflow-integrator\", \"morning-ship\", \"mirofish\", \"tab-dualma-refresh\", \"tab-technical-analysis\", \"i18n-sync\", \"weekly-stock-update\", \"daily-stock-check\", \"ai-quality-evaluator\", \"bespin-news-digest\", \"tab-fundamental-sync\", \"tab-hot-stock-discovery\", \"tab-turtle-refresh\", \"tab-bollinger-refresh\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0088",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: NotebookLM Studio: Content Generation. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "notebooklm-studio",
        "nlm-deep-learn",
        "notebooklm",
        "nlm-strategy-hub",
        "nlm-slides",
        "nlm-video",
        "nlm-launch-kit",
        "notebooklm-research",
        "nlm-discovery-lab",
        "gws-recipe-doc-from-template",
        "knowledge-to-paper",
        "gws-drive",
        "nlm-arxiv-slides",
        "notion-meeting-sync",
        "gws-docs",
        "gws-sheets",
        "office-template-enforcer",
        "bespin-news-digest",
        "gws-recipe-events-from-sheet",
        "gws-recipe-save-email-to-doc",
        "gws-workspace",
        "hf-buckets",
        "kwp-brand-voice-brand-voice-enforcement",
        "kwp-brand-voice-guideline-generation",
        "cognee"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "notebooklm-studio",
          "tokens": 1812
        },
        {
          "skill_id": "nlm-deep-learn",
          "tokens": 3214
        },
        {
          "skill_id": "notebooklm",
          "tokens": 1844
        },
        {
          "skill_id": "nlm-strategy-hub",
          "tokens": 2207
        },
        {
          "skill_id": "nlm-slides",
          "tokens": 1685
        },
        {
          "skill_id": "nlm-video",
          "tokens": 1763
        },
        {
          "skill_id": "nlm-launch-kit",
          "tokens": 2368
        },
        {
          "skill_id": "notebooklm-research",
          "tokens": 1362
        },
        {
          "skill_id": "nlm-discovery-lab",
          "tokens": 2808
        },
        {
          "skill_id": "gws-recipe-doc-from-template",
          "tokens": 743
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "gws-drive",
          "tokens": 997
        },
        {
          "skill_id": "nlm-arxiv-slides",
          "tokens": 3484
        },
        {
          "skill_id": "notion-meeting-sync",
          "tokens": 4347
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "gws-sheets",
          "tokens": 918
        },
        {
          "skill_id": "office-template-enforcer",
          "tokens": 1273
        },
        {
          "skill_id": "bespin-news-digest",
          "tokens": 3949
        },
        {
          "skill_id": "gws-recipe-events-from-sheet",
          "tokens": 674
        },
        {
          "skill_id": "gws-recipe-save-email-to-doc",
          "tokens": 676
        },
        {
          "skill_id": "gws-workspace",
          "tokens": 1186
        },
        {
          "skill_id": "hf-buckets",
          "tokens": 1334
        },
        {
          "skill_id": "kwp-brand-voice-brand-voice-enforcement",
          "tokens": 1634
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "cognee",
          "tokens": 2070
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: NotebookLM Studio: Content Generation. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"notebooklm-studio\", \"nlm-deep-learn\", \"notebooklm\", \"nlm-strategy-hub\", \"nlm-slides\", \"nlm-video\", \"nlm-launch-kit\", \"notebooklm-research\", \"nlm-discovery-lab\", \"gws-recipe-doc-from-template\", \"knowledge-to-paper\", \"gws-drive\", \"nlm-arxiv-slides\", \"notion-meeting-sync\", \"gws-docs\", \"gws-sheets\", \"office-template-enforcer\", \"bespin-news-digest\", \"gws-recipe-events-from-sheet\", \"gws-recipe-save-email-to-doc\", \"gws-workspace\", \"hf-buckets\", \"kwp-brand-voice-brand-voice-enforcement\", \"kwp-brand-voice-guideline-generation\", \"cognee\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0089",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-bollinger-refresh. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "tab-bollinger-refresh",
        "tab-turtle-refresh",
        "tab-dualma-refresh",
        "tab-strategy-comparison",
        "tab-market-breadth",
        "daily-stock-check",
        "tab-llm-agents",
        "tab-technical-analysis",
        "tab-market-environment",
        "tab-report-generate",
        "tab-fundamental-sync",
        "tab-stock-sync",
        "tab-hot-stock-discovery",
        "tab-screening",
        "tab-news-fetch",
        "trading-backtest-expert",
        "tab-genai-features",
        "tab-patterns",
        "tab-glossary",
        "alphaear-predictor",
        "trading-trade-hypothesis-ideator",
        "stock-csv-downloader",
        "weekly-stock-update",
        "trading-us-stock-analysis",
        "tab-dashboard-summary"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "tab-bollinger-refresh",
          "tokens": 383
        },
        {
          "skill_id": "tab-turtle-refresh",
          "tokens": 371
        },
        {
          "skill_id": "tab-dualma-refresh",
          "tokens": 391
        },
        {
          "skill_id": "tab-strategy-comparison",
          "tokens": 571
        },
        {
          "skill_id": "tab-market-breadth",
          "tokens": 504
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "tab-llm-agents",
          "tokens": 382
        },
        {
          "skill_id": "tab-technical-analysis",
          "tokens": 414
        },
        {
          "skill_id": "tab-market-environment",
          "tokens": 566
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "tab-fundamental-sync",
          "tokens": 331
        },
        {
          "skill_id": "tab-stock-sync",
          "tokens": 345
        },
        {
          "skill_id": "tab-hot-stock-discovery",
          "tokens": 359
        },
        {
          "skill_id": "tab-screening",
          "tokens": 410
        },
        {
          "skill_id": "tab-news-fetch",
          "tokens": 474
        },
        {
          "skill_id": "trading-backtest-expert",
          "tokens": 2754
        },
        {
          "skill_id": "tab-genai-features",
          "tokens": 354
        },
        {
          "skill_id": "tab-patterns",
          "tokens": 509
        },
        {
          "skill_id": "tab-glossary",
          "tokens": 538
        },
        {
          "skill_id": "alphaear-predictor",
          "tokens": 803
        },
        {
          "skill_id": "trading-trade-hypothesis-ideator",
          "tokens": 909
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "trading-us-stock-analysis",
          "tokens": 2926
        },
        {
          "skill_id": "tab-dashboard-summary",
          "tokens": 419
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-bollinger-refresh. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"tab-bollinger-refresh\", \"tab-turtle-refresh\", \"tab-dualma-refresh\", \"tab-strategy-comparison\", \"tab-market-breadth\", \"daily-stock-check\", \"tab-llm-agents\", \"tab-technical-analysis\", \"tab-market-environment\", \"tab-report-generate\", \"tab-fundamental-sync\", \"tab-stock-sync\", \"tab-hot-stock-discovery\", \"tab-screening\", \"tab-news-fetch\", \"trading-backtest-expert\", \"tab-genai-features\", \"tab-patterns\", \"tab-glossary\", \"alphaear-predictor\", \"trading-trade-hypothesis-ideator\", \"stock-csv-downloader\", \"weekly-stock-update\", \"trading-us-stock-analysis\", \"tab-dashboard-summary\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0090",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AlphaEar News. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "alphaear-news",
        "alphaear-search",
        "alphaear-deepear-lite",
        "alphaear-signal-tracker",
        "alphaear-sentiment",
        "stock-csv-downloader",
        "weekly-stock-update",
        "alphaear-predictor",
        "daily-stock-check",
        "tab-news-fetch",
        "alphaear-stock",
        "hf-trackio",
        "trading-market-news-analyst",
        "trading-market-environment-analysis",
        "trading-technical-analyst",
        "mirofish-financial-sim",
        "portfolio-report-generator",
        "trading-sector-analyst",
        "hf-datasets",
        "hf-evaluation",
        "alphaear-reporter",
        "weekly-status-report",
        "service-health-doctor",
        "deep-review",
        "trading-data-quality-checker"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "alphaear-news",
          "tokens": 873
        },
        {
          "skill_id": "alphaear-search",
          "tokens": 672
        },
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        },
        {
          "skill_id": "alphaear-signal-tracker",
          "tokens": 854
        },
        {
          "skill_id": "alphaear-sentiment",
          "tokens": 928
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "alphaear-predictor",
          "tokens": 803
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "tab-news-fetch",
          "tokens": 474
        },
        {
          "skill_id": "alphaear-stock",
          "tokens": 820
        },
        {
          "skill_id": "hf-trackio",
          "tokens": 1424
        },
        {
          "skill_id": "trading-market-news-analyst",
          "tokens": 1467
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "trading-technical-analyst",
          "tokens": 2868
        },
        {
          "skill_id": "mirofish-financial-sim",
          "tokens": 1330
        },
        {
          "skill_id": "portfolio-report-generator",
          "tokens": 1406
        },
        {
          "skill_id": "trading-sector-analyst",
          "tokens": 2657
        },
        {
          "skill_id": "hf-datasets",
          "tokens": 3241
        },
        {
          "skill_id": "hf-evaluation",
          "tokens": 4549
        },
        {
          "skill_id": "alphaear-reporter",
          "tokens": 810
        },
        {
          "skill_id": "weekly-status-report",
          "tokens": 1282
        },
        {
          "skill_id": "service-health-doctor",
          "tokens": 1728
        },
        {
          "skill_id": "deep-review",
          "tokens": 1923
        },
        {
          "skill_id": "trading-data-quality-checker",
          "tokens": 1962
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AlphaEar News. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-news\", \"alphaear-search\", \"alphaear-deepear-lite\", \"alphaear-signal-tracker\", \"alphaear-sentiment\", \"stock-csv-downloader\", \"weekly-stock-update\", \"alphaear-predictor\", \"daily-stock-check\", \"tab-news-fetch\", \"alphaear-stock\", \"hf-trackio\", \"trading-market-news-analyst\", \"trading-market-environment-analysis\", \"trading-technical-analyst\", \"mirofish-financial-sim\", \"portfolio-report-generator\", \"trading-sector-analyst\", \"hf-datasets\", \"hf-evaluation\", \"alphaear-reporter\", \"weekly-status-report\", \"service-health-doctor\", \"deep-review\", \"trading-data-quality-checker\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0091",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AlphaXiv Paper Lookup. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "alphaxiv-paper-lookup",
        "paper-auto-classifier",
        "meeting-digest",
        "hf-papers",
        "hf-trending-intelligence",
        "defuddle",
        "docx-template-engine",
        "role-developer",
        "helm-validator",
        "refactor-simulator",
        "related-papers-scout",
        "notion-meeting-sync",
        "paper-review",
        "kwp-engineering-code-review",
        "email-auto-reply",
        "github-sprint-digest",
        "ppt-template-engine",
        "role-ceo",
        "role-ux-designer",
        "role-cso",
        "role-hr",
        "role-cto",
        "terraform-reviewer",
        "issue-to-code",
        "unified-intel-intake"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "alphaxiv-paper-lookup",
          "tokens": 1299
        },
        {
          "skill_id": "paper-auto-classifier",
          "tokens": 1262
        },
        {
          "skill_id": "meeting-digest",
          "tokens": 4491
        },
        {
          "skill_id": "hf-papers",
          "tokens": 817
        },
        {
          "skill_id": "hf-trending-intelligence",
          "tokens": 1832
        },
        {
          "skill_id": "defuddle",
          "tokens": 2409
        },
        {
          "skill_id": "docx-template-engine",
          "tokens": 1193
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "helm-validator",
          "tokens": 1398
        },
        {
          "skill_id": "refactor-simulator",
          "tokens": 2044
        },
        {
          "skill_id": "related-papers-scout",
          "tokens": 3560
        },
        {
          "skill_id": "notion-meeting-sync",
          "tokens": 4347
        },
        {
          "skill_id": "paper-review",
          "tokens": 5815
        },
        {
          "skill_id": "kwp-engineering-code-review",
          "tokens": 610
        },
        {
          "skill_id": "email-auto-reply",
          "tokens": 638
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "ppt-template-engine",
          "tokens": 1101
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "role-ux-designer",
          "tokens": 1268
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        },
        {
          "skill_id": "role-hr",
          "tokens": 1350
        },
        {
          "skill_id": "role-cto",
          "tokens": 1413
        },
        {
          "skill_id": "terraform-reviewer",
          "tokens": 1457
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "unified-intel-intake",
          "tokens": 1593
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AlphaXiv Paper Lookup. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"alphaxiv-paper-lookup\", \"paper-auto-classifier\", \"meeting-digest\", \"hf-papers\", \"hf-trending-intelligence\", \"defuddle\", \"docx-template-engine\", \"role-developer\", \"helm-validator\", \"refactor-simulator\", \"related-papers-scout\", \"notion-meeting-sync\", \"paper-review\", \"kwp-engineering-code-review\", \"email-auto-reply\", \"github-sprint-digest\", \"ppt-template-engine\", \"role-ceo\", \"role-ux-designer\", \"role-cso\", \"role-hr\", \"role-cto\", \"terraform-reviewer\", \"issue-to-code\", \"unified-intel-intake\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0092",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-event-detect. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "tab-event-detect",
        "tab-patterns",
        "tab-genai-features",
        "tab-analysis-run",
        "tab-fundamental-sync",
        "tab-screening",
        "tab-strategy-comparison",
        "cursor-automations",
        "tab-stock-sync",
        "tab-hot-stock-discovery",
        "tab-turtle-refresh",
        "tab-llm-agents",
        "tab-dualma-refresh",
        "tab-technical-analysis",
        "tab-dashboard-summary",
        "tab-news-fetch",
        "tab-market-breadth",
        "tab-glossary",
        "tab-report-generate",
        "gws-recipe-events-from-sheet",
        "email-research-dispatcher",
        "sefo-orchestrator",
        "hf-dataset-viewer",
        "skill-seekers",
        "daily-stock-check"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "tab-event-detect",
          "tokens": 319
        },
        {
          "skill_id": "tab-patterns",
          "tokens": 509
        },
        {
          "skill_id": "tab-genai-features",
          "tokens": 354
        },
        {
          "skill_id": "tab-analysis-run",
          "tokens": 404
        },
        {
          "skill_id": "tab-fundamental-sync",
          "tokens": 331
        },
        {
          "skill_id": "tab-screening",
          "tokens": 410
        },
        {
          "skill_id": "tab-strategy-comparison",
          "tokens": 571
        },
        {
          "skill_id": "cursor-automations",
          "tokens": 2911
        },
        {
          "skill_id": "tab-stock-sync",
          "tokens": 345
        },
        {
          "skill_id": "tab-hot-stock-discovery",
          "tokens": 359
        },
        {
          "skill_id": "tab-turtle-refresh",
          "tokens": 371
        },
        {
          "skill_id": "tab-llm-agents",
          "tokens": 382
        },
        {
          "skill_id": "tab-dualma-refresh",
          "tokens": 391
        },
        {
          "skill_id": "tab-technical-analysis",
          "tokens": 414
        },
        {
          "skill_id": "tab-dashboard-summary",
          "tokens": 419
        },
        {
          "skill_id": "tab-news-fetch",
          "tokens": 474
        },
        {
          "skill_id": "tab-market-breadth",
          "tokens": 504
        },
        {
          "skill_id": "tab-glossary",
          "tokens": 538
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "gws-recipe-events-from-sheet",
          "tokens": 674
        },
        {
          "skill_id": "email-research-dispatcher",
          "tokens": 703
        },
        {
          "skill_id": "sefo-orchestrator",
          "tokens": 821
        },
        {
          "skill_id": "hf-dataset-viewer",
          "tokens": 1387
        },
        {
          "skill_id": "skill-seekers",
          "tokens": 1568
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-event-detect. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"tab-event-detect\", \"tab-patterns\", \"tab-genai-features\", \"tab-analysis-run\", \"tab-fundamental-sync\", \"tab-screening\", \"tab-strategy-comparison\", \"cursor-automations\", \"tab-stock-sync\", \"tab-hot-stock-discovery\", \"tab-turtle-refresh\", \"tab-llm-agents\", \"tab-dualma-refresh\", \"tab-technical-analysis\", \"tab-dashboard-summary\", \"tab-news-fetch\", \"tab-market-breadth\", \"tab-glossary\", \"tab-report-generate\", \"gws-recipe-events-from-sheet\", \"email-research-dispatcher\", \"sefo-orchestrator\", \"hf-dataset-viewer\", \"skill-seekers\", \"daily-stock-check\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0093",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AlphaEar DeepEar Lite. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "alphaear-deepear-lite",
        "alphaear-news",
        "alphaear-sentiment",
        "trading-market-environment-analysis",
        "trading-us-stock-analysis",
        "alphaear-predictor",
        "alphaear-stock",
        "stock-csv-downloader",
        "daily-stock-check",
        "alphaear-search",
        "weekly-stock-update",
        "trading-sector-analyst",
        "trading-technical-analyst",
        "alphaear-signal-tracker",
        "trading-scenario-analyzer",
        "mirofish-financial-sim",
        "trading-market-news-analyst",
        "trading-market-breadth-analyzer",
        "trading-us-market-bubble-detector",
        "trading-breadth-chart-analyst",
        "alphaear-reporter",
        "trading-uptrend-analyzer",
        "ai-workflow-integrator",
        "release-commander",
        "mission-control"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        },
        {
          "skill_id": "alphaear-news",
          "tokens": 873
        },
        {
          "skill_id": "alphaear-sentiment",
          "tokens": 928
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "trading-us-stock-analysis",
          "tokens": 2926
        },
        {
          "skill_id": "alphaear-predictor",
          "tokens": 803
        },
        {
          "skill_id": "alphaear-stock",
          "tokens": 820
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "alphaear-search",
          "tokens": 672
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "trading-sector-analyst",
          "tokens": 2657
        },
        {
          "skill_id": "trading-technical-analyst",
          "tokens": 2868
        },
        {
          "skill_id": "alphaear-signal-tracker",
          "tokens": 854
        },
        {
          "skill_id": "trading-scenario-analyzer",
          "tokens": 2068
        },
        {
          "skill_id": "mirofish-financial-sim",
          "tokens": 1330
        },
        {
          "skill_id": "trading-market-news-analyst",
          "tokens": 1467
        },
        {
          "skill_id": "trading-market-breadth-analyzer",
          "tokens": 1830
        },
        {
          "skill_id": "trading-us-market-bubble-detector",
          "tokens": 3062
        },
        {
          "skill_id": "trading-breadth-chart-analyst",
          "tokens": 3912
        },
        {
          "skill_id": "alphaear-reporter",
          "tokens": 810
        },
        {
          "skill_id": "trading-uptrend-analyzer",
          "tokens": 1780
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AlphaEar DeepEar Lite. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-deepear-lite\", \"alphaear-news\", \"alphaear-sentiment\", \"trading-market-environment-analysis\", \"trading-us-stock-analysis\", \"alphaear-predictor\", \"alphaear-stock\", \"stock-csv-downloader\", \"daily-stock-check\", \"alphaear-search\", \"weekly-stock-update\", \"trading-sector-analyst\", \"trading-technical-analyst\", \"alphaear-signal-tracker\", \"trading-scenario-analyzer\", \"mirofish-financial-sim\", \"trading-market-news-analyst\", \"trading-market-breadth-analyzer\", \"trading-us-market-bubble-detector\", \"trading-breadth-chart-analyst\", \"alphaear-reporter\", \"trading-uptrend-analyzer\", \"ai-workflow-integrator\", \"release-commander\", \"mission-control\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0094",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Write internal communications using company formats. Use for 3P updates (Progress, Plans, Problems), company newsletters, FAQ responses, status reports, leadership updates, project updates, incident r…. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "anthropic-internal-comms",
        "agency-executive-summary-generator",
        "pm-go-to-market",
        "kwp-product-management-stakeholder-comms",
        "sales-meeting-pipeline",
        "pm-toolkit",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "agency-twitter-engager",
        "slack-agent",
        "agency-xiaohongshu-specialist",
        "agency-project-shepherd",
        "agency-developer-advocate",
        "kwp-engineering-testing-strategy",
        "gws-recipe-audit-sharing",
        "gws-docs",
        "kwp-productivity-task-management",
        "pm-marketing-growth",
        "anthropic-theme-factory",
        "weekly-status-report",
        "agency-instagram-curator",
        "kwp-product-management-roadmap-management",
        "agency-rapid-prototyper",
        "kwp-marketing-content-creation",
        "problem-definition"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "anthropic-internal-comms",
          "tokens": 522
        },
        {
          "skill_id": "agency-executive-summary-generator",
          "tokens": 2465
        },
        {
          "skill_id": "pm-go-to-market",
          "tokens": 992
        },
        {
          "skill_id": "kwp-product-management-stakeholder-comms",
          "tokens": 2952
        },
        {
          "skill_id": "sales-meeting-pipeline",
          "tokens": 559
        },
        {
          "skill_id": "pm-toolkit",
          "tokens": 935
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "slack-agent",
          "tokens": 2564
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "gws-recipe-audit-sharing",
          "tokens": 645
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "kwp-productivity-task-management",
          "tokens": 776
        },
        {
          "skill_id": "pm-marketing-growth",
          "tokens": 954
        },
        {
          "skill_id": "anthropic-theme-factory",
          "tokens": 990
        },
        {
          "skill_id": "weekly-status-report",
          "tokens": 1282
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "kwp-product-management-roadmap-management",
          "tokens": 2467
        },
        {
          "skill_id": "agency-rapid-prototyper",
          "tokens": 2496
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "problem-definition",
          "tokens": 2601
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Write internal communications using company formats. Use for 3P updates (Progress, Plans, Problems), company newsletters, FAQ responses, status reports, leadership updates, project updates, incident r…. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-internal-comms\", \"agency-executive-summary-generator\", \"pm-go-to-market\", \"kwp-product-management-stakeholder-comms\", \"sales-meeting-pipeline\", \"pm-toolkit\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"agency-twitter-engager\", \"slack-agent\", \"agency-xiaohongshu-specialist\", \"agency-project-shepherd\", \"agency-developer-advocate\", \"kwp-engineering-testing-strategy\", \"gws-recipe-audit-sharing\", \"gws-docs\", \"kwp-productivity-task-management\", \"pm-marketing-growth\", \"anthropic-theme-factory\", \"weekly-status-report\", \"agency-instagram-curator\", \"kwp-product-management-roadmap-management\", \"agency-rapid-prototyper\", \"kwp-marketing-content-creation\", \"problem-definition\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0095",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Sales Perspective Analyzer. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "role-sales",
        "role-ux-designer",
        "role-hr",
        "role-cso",
        "role-finance",
        "role-developer",
        "role-ceo",
        "role-pm",
        "role-security-engineer",
        "role-cto",
        "role-data-scientist",
        "role-trading-expert",
        "deep-research-pipeline",
        "executive-briefing",
        "paper-review",
        "full-stack-planner",
        "role-dispatcher",
        "prd-research-factory",
        "pm-market-research",
        "pm-go-to-market",
        "kwp-brand-voice-guideline-generation",
        "nlm-launch-kit",
        "meeting-digest",
        "kwp-engineering-documentation",
        "agency-data-consolidation-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "role-sales",
          "tokens": 1221
        },
        {
          "skill_id": "role-ux-designer",
          "tokens": 1268
        },
        {
          "skill_id": "role-hr",
          "tokens": 1350
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        },
        {
          "skill_id": "role-finance",
          "tokens": 1353
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "role-pm",
          "tokens": 1304
        },
        {
          "skill_id": "role-security-engineer",
          "tokens": 1406
        },
        {
          "skill_id": "role-cto",
          "tokens": 1413
        },
        {
          "skill_id": "role-data-scientist",
          "tokens": 1695
        },
        {
          "skill_id": "role-trading-expert",
          "tokens": 1791
        },
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        },
        {
          "skill_id": "executive-briefing",
          "tokens": 1419
        },
        {
          "skill_id": "paper-review",
          "tokens": 5815
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        },
        {
          "skill_id": "role-dispatcher",
          "tokens": 2075
        },
        {
          "skill_id": "prd-research-factory",
          "tokens": 648
        },
        {
          "skill_id": "pm-market-research",
          "tokens": 966
        },
        {
          "skill_id": "pm-go-to-market",
          "tokens": 992
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "nlm-launch-kit",
          "tokens": 2368
        },
        {
          "skill_id": "meeting-digest",
          "tokens": 4491
        },
        {
          "skill_id": "kwp-engineering-documentation",
          "tokens": 654
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Sales Perspective Analyzer. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"role-sales\", \"role-ux-designer\", \"role-hr\", \"role-cso\", \"role-finance\", \"role-developer\", \"role-ceo\", \"role-pm\", \"role-security-engineer\", \"role-cto\", \"role-data-scientist\", \"role-trading-expert\", \"deep-research-pipeline\", \"executive-briefing\", \"paper-review\", \"full-stack-planner\", \"role-dispatcher\", \"prd-research-factory\", \"pm-market-research\", \"pm-go-to-market\", \"kwp-brand-voice-guideline-generation\", \"nlm-launch-kit\", \"meeting-digest\", \"kwp-engineering-documentation\", \"agency-data-consolidation-agent\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0096",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PM Toolkit. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "pm-toolkit",
        "pm-go-to-market",
        "i18n-sync",
        "pm-marketing-growth",
        "pm-data-analytics",
        "pr-review-captain",
        "proposal-factory",
        "pm-product-discovery",
        "compliance-governance",
        "prompt-transformer",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "docs-tutor-setup",
        "agency-xiaohongshu-specialist",
        "skill-optimizer",
        "agency-legal-compliance-checker",
        "anthropic-doc-coauthoring",
        "gws-recipe-audit-sharing",
        "sefo-governance",
        "pm-market-research",
        "gws-calendar",
        "ux-expert",
        "role-developer",
        "evals-skills",
        "skill-guide-generator"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "pm-toolkit",
          "tokens": 935
        },
        {
          "skill_id": "pm-go-to-market",
          "tokens": 992
        },
        {
          "skill_id": "i18n-sync",
          "tokens": 1282
        },
        {
          "skill_id": "pm-marketing-growth",
          "tokens": 954
        },
        {
          "skill_id": "pm-data-analytics",
          "tokens": 956
        },
        {
          "skill_id": "pr-review-captain",
          "tokens": 1496
        },
        {
          "skill_id": "proposal-factory",
          "tokens": 582
        },
        {
          "skill_id": "pm-product-discovery",
          "tokens": 1198
        },
        {
          "skill_id": "compliance-governance",
          "tokens": 1501
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "docs-tutor-setup",
          "tokens": 2288
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "skill-optimizer",
          "tokens": 2940
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "anthropic-doc-coauthoring",
          "tokens": 4138
        },
        {
          "skill_id": "gws-recipe-audit-sharing",
          "tokens": 645
        },
        {
          "skill_id": "sefo-governance",
          "tokens": 837
        },
        {
          "skill_id": "pm-market-research",
          "tokens": 966
        },
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "ux-expert",
          "tokens": 1210
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "evals-skills",
          "tokens": 1524
        },
        {
          "skill_id": "skill-guide-generator",
          "tokens": 1858
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PM Toolkit. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"pm-toolkit\", \"pm-go-to-market\", \"i18n-sync\", \"pm-marketing-growth\", \"pm-data-analytics\", \"pr-review-captain\", \"proposal-factory\", \"pm-product-discovery\", \"compliance-governance\", \"prompt-transformer\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"docs-tutor-setup\", \"agency-xiaohongshu-specialist\", \"skill-optimizer\", \"agency-legal-compliance-checker\", \"anthropic-doc-coauthoring\", \"gws-recipe-audit-sharing\", \"sefo-governance\", \"pm-market-research\", \"gws-calendar\", \"ux-expert\", \"role-developer\", \"evals-skills\", \"skill-guide-generator\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0097",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Inclusive Visuals Specialist. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-inclusive-visuals-specialist",
        "agency-instagram-curator",
        "agency-visual-storyteller",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "agency-twitter-engager",
        "agency-whimsy-injector",
        "agency-xiaohongshu-specialist",
        "agency-workflow-optimizer",
        "agency-ux-architect",
        "agency-zhihu-strategist",
        "agency-infrastructure-maintainer",
        "agency-support-responder",
        "agency-api-tester",
        "agency-app-store-optimizer",
        "agency-legal-compliance-checker",
        "agency-test-results-analyzer",
        "agency-tool-evaluator",
        "agency-xr-cockpit-interaction-specialist",
        "agency-xr-immersive-developer",
        "agency-growth-hacker",
        "agency-content-creator",
        "agency-terminal-integration-specialist",
        "agency-behavioral-nudge-engine",
        "agency-cultural-intelligence-strategist"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-inclusive-visuals-specialist",
          "tokens": 1931
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-ux-architect",
          "tokens": 3119
        },
        {
          "skill_id": "agency-zhihu-strategist",
          "tokens": 3239
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-support-responder",
          "tokens": 3296
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-app-store-optimizer",
          "tokens": 3512
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "agency-tool-evaluator",
          "tokens": 4549
        },
        {
          "skill_id": "agency-xr-cockpit-interaction-specialist",
          "tokens": 693
        },
        {
          "skill_id": "agency-xr-immersive-developer",
          "tokens": 697
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "agency-terminal-integration-specialist",
          "tokens": 1273
        },
        {
          "skill_id": "agency-behavioral-nudge-engine",
          "tokens": 1553
        },
        {
          "skill_id": "agency-cultural-intelligence-strategist",
          "tokens": 1892
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Inclusive Visuals Specialist. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-inclusive-visuals-specialist\", \"agency-instagram-curator\", \"agency-visual-storyteller\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"agency-twitter-engager\", \"agency-whimsy-injector\", \"agency-xiaohongshu-specialist\", \"agency-workflow-optimizer\", \"agency-ux-architect\", \"agency-zhihu-strategist\", \"agency-infrastructure-maintainer\", \"agency-support-responder\", \"agency-api-tester\", \"agency-app-store-optimizer\", \"agency-legal-compliance-checker\", \"agency-test-results-analyzer\", \"agency-tool-evaluator\", \"agency-xr-cockpit-interaction-specialist\", \"agency-xr-immersive-developer\", \"agency-growth-hacker\", \"agency-content-creator\", \"agency-terminal-integration-specialist\", \"agency-behavioral-nudge-engine\", \"agency-cultural-intelligence-strategist\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0098",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Jira/Linear Adapter. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "jira-linear-adapter",
        "pr-to-issue-linker",
        "sprint-orchestrator",
        "github-sprint-digest",
        "portfolio-report-generator",
        "sefo-federation",
        "commit-to-issue",
        "feedback-meeting-scheduler",
        "sod-ship",
        "weekly-status-report",
        "standup-digest",
        "proactive-meeting-scheduler",
        "agency-project-shepherd",
        "paperclip-tasks",
        "issue-to-code",
        "ship",
        "gmail-daily-triage",
        "release-commander",
        "eod-ship",
        "simplify",
        "cursor-automations",
        "critical-review",
        "kwp-design-design-system-management",
        "domain-commit",
        "ppt-template-engine"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "jira-linear-adapter",
          "tokens": 1365
        },
        {
          "skill_id": "pr-to-issue-linker",
          "tokens": 1223
        },
        {
          "skill_id": "sprint-orchestrator",
          "tokens": 1320
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "portfolio-report-generator",
          "tokens": 1406
        },
        {
          "skill_id": "sefo-federation",
          "tokens": 756
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "feedback-meeting-scheduler",
          "tokens": 1238
        },
        {
          "skill_id": "sod-ship",
          "tokens": 4151
        },
        {
          "skill_id": "weekly-status-report",
          "tokens": 1282
        },
        {
          "skill_id": "standup-digest",
          "tokens": 1330
        },
        {
          "skill_id": "proactive-meeting-scheduler",
          "tokens": 1444
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "paperclip-tasks",
          "tokens": 1534
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "gmail-daily-triage",
          "tokens": 2229
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "eod-ship",
          "tokens": 2420
        },
        {
          "skill_id": "simplify",
          "tokens": 2596
        },
        {
          "skill_id": "cursor-automations",
          "tokens": 2911
        },
        {
          "skill_id": "critical-review",
          "tokens": 3039
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "domain-commit",
          "tokens": 915
        },
        {
          "skill_id": "ppt-template-engine",
          "tokens": 1101
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Jira/Linear Adapter. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"jira-linear-adapter\", \"pr-to-issue-linker\", \"sprint-orchestrator\", \"github-sprint-digest\", \"portfolio-report-generator\", \"sefo-federation\", \"commit-to-issue\", \"feedback-meeting-scheduler\", \"sod-ship\", \"weekly-status-report\", \"standup-digest\", \"proactive-meeting-scheduler\", \"agency-project-shepherd\", \"paperclip-tasks\", \"issue-to-code\", \"ship\", \"gmail-daily-triage\", \"release-commander\", \"eod-ship\", \"simplify\", \"cursor-automations\", \"critical-review\", \"kwp-design-design-system-management\", \"domain-commit\", \"ppt-template-engine\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0099",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: EOD Ship. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "eod-ship",
        "sod-ship",
        "morning-ship",
        "morning",
        "portfolio-report-generator",
        "cursor-sync",
        "daily-am-orchestrator",
        "daily-strategy-post",
        "release-commander",
        "release-ship",
        "planning-weekly-pulse",
        "github-sprint-digest",
        "knowledge-daily-aggregator",
        "critical-review",
        "bespin-news-digest",
        "standup-digest",
        "jira-linear-adapter",
        "weekly-stock-update",
        "ui-suite",
        "unified-intel-intake",
        "ship",
        "test-suite",
        "setup-doctor",
        "simplify",
        "cursor-automations"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "eod-ship",
          "tokens": 2420
        },
        {
          "skill_id": "sod-ship",
          "tokens": 4151
        },
        {
          "skill_id": "morning-ship",
          "tokens": 2943
        },
        {
          "skill_id": "morning",
          "tokens": 2088
        },
        {
          "skill_id": "portfolio-report-generator",
          "tokens": 1406
        },
        {
          "skill_id": "cursor-sync",
          "tokens": 3267
        },
        {
          "skill_id": "daily-am-orchestrator",
          "tokens": 3613
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "release-ship",
          "tokens": 3056
        },
        {
          "skill_id": "planning-weekly-pulse",
          "tokens": 528
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "knowledge-daily-aggregator",
          "tokens": 1425
        },
        {
          "skill_id": "critical-review",
          "tokens": 3039
        },
        {
          "skill_id": "bespin-news-digest",
          "tokens": 3949
        },
        {
          "skill_id": "standup-digest",
          "tokens": 1330
        },
        {
          "skill_id": "jira-linear-adapter",
          "tokens": 1365
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "unified-intel-intake",
          "tokens": 1593
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "setup-doctor",
          "tokens": 2578
        },
        {
          "skill_id": "simplify",
          "tokens": 2596
        },
        {
          "skill_id": "cursor-automations",
          "tokens": 2911
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: EOD Ship. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"eod-ship\", \"sod-ship\", \"morning-ship\", \"morning\", \"portfolio-report-generator\", \"cursor-sync\", \"daily-am-orchestrator\", \"daily-strategy-post\", \"release-commander\", \"release-ship\", \"planning-weekly-pulse\", \"github-sprint-digest\", \"knowledge-daily-aggregator\", \"critical-review\", \"bespin-news-digest\", \"standup-digest\", \"jira-linear-adapter\", \"weekly-stock-update\", \"ui-suite\", \"unified-intel-intake\", \"ship\", \"test-suite\", \"setup-doctor\", \"simplify\", \"cursor-automations\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0100",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Evals Skills. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "evals-skills",
        "ai-workflow-integrator",
        "ai-quality-evaluator",
        "semantic-guard",
        "pm-data-analytics",
        "intent-alignment-tracker",
        "terraform-reviewer",
        "ui-suite",
        "performance-profiler",
        "test-suite",
        "trading-data-quality-checker",
        "skill-optimizer",
        "backend-expert",
        "autoskill-judge",
        "ux-expert",
        "i18n-sync",
        "e2e-testing",
        "compliance-governance",
        "deep-review",
        "agency-ai-engineer",
        "pipeline-builder",
        "release-commander",
        "kwp-engineering-tech-debt",
        "ci-quality-gate",
        "frontend-expert"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "evals-skills",
          "tokens": 1524
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "ai-quality-evaluator",
          "tokens": 2228
        },
        {
          "skill_id": "semantic-guard",
          "tokens": 571
        },
        {
          "skill_id": "pm-data-analytics",
          "tokens": 956
        },
        {
          "skill_id": "intent-alignment-tracker",
          "tokens": 558
        },
        {
          "skill_id": "terraform-reviewer",
          "tokens": 1457
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "performance-profiler",
          "tokens": 1738
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "trading-data-quality-checker",
          "tokens": 1962
        },
        {
          "skill_id": "skill-optimizer",
          "tokens": 2940
        },
        {
          "skill_id": "backend-expert",
          "tokens": 1132
        },
        {
          "skill_id": "autoskill-judge",
          "tokens": 1176
        },
        {
          "skill_id": "ux-expert",
          "tokens": 1210
        },
        {
          "skill_id": "i18n-sync",
          "tokens": 1282
        },
        {
          "skill_id": "e2e-testing",
          "tokens": 1484
        },
        {
          "skill_id": "compliance-governance",
          "tokens": 1501
        },
        {
          "skill_id": "deep-review",
          "tokens": 1923
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "pipeline-builder",
          "tokens": 2234
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "kwp-engineering-tech-debt",
          "tokens": 648
        },
        {
          "skill_id": "ci-quality-gate",
          "tokens": 1054
        },
        {
          "skill_id": "frontend-expert",
          "tokens": 1234
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Evals Skills. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"evals-skills\", \"ai-workflow-integrator\", \"ai-quality-evaluator\", \"semantic-guard\", \"pm-data-analytics\", \"intent-alignment-tracker\", \"terraform-reviewer\", \"ui-suite\", \"performance-profiler\", \"test-suite\", \"trading-data-quality-checker\", \"skill-optimizer\", \"backend-expert\", \"autoskill-judge\", \"ux-expert\", \"i18n-sync\", \"e2e-testing\", \"compliance-governance\", \"deep-review\", \"agency-ai-engineer\", \"pipeline-builder\", \"release-commander\", \"kwp-engineering-tech-debt\", \"ci-quality-gate\", \"frontend-expert\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0101",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Prompt Architect. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "prompt-architect",
        "helm-validator",
        "prompt-transformer",
        "plans",
        "automation-strategist",
        "kwp-design-user-research",
        "pm-product-discovery",
        "terraform-reviewer",
        "workflow-sequential",
        "workflow-eval-opt",
        "gws-email-reply",
        "defuddle",
        "agency-ux-researcher",
        "notion-meeting-sync",
        "kwp-engineering-tech-debt",
        "kwp-design-design-system-management",
        "kwp-design-design-critique",
        "knowledge-to-paper",
        "mirofish-graph-explorer",
        "role-ceo",
        "docx-template-engine",
        "paper-auto-classifier",
        "role-ux-designer",
        "alphaxiv-paper-lookup",
        "role-cso"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "prompt-architect",
          "tokens": 1939
        },
        {
          "skill_id": "helm-validator",
          "tokens": 1398
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "plans",
          "tokens": 2297
        },
        {
          "skill_id": "automation-strategist",
          "tokens": 2482
        },
        {
          "skill_id": "kwp-design-user-research",
          "tokens": 718
        },
        {
          "skill_id": "pm-product-discovery",
          "tokens": 1198
        },
        {
          "skill_id": "terraform-reviewer",
          "tokens": 1457
        },
        {
          "skill_id": "workflow-sequential",
          "tokens": 1485
        },
        {
          "skill_id": "workflow-eval-opt",
          "tokens": 2185
        },
        {
          "skill_id": "gws-email-reply",
          "tokens": 2222
        },
        {
          "skill_id": "defuddle",
          "tokens": 2409
        },
        {
          "skill_id": "agency-ux-researcher",
          "tokens": 3473
        },
        {
          "skill_id": "notion-meeting-sync",
          "tokens": 4347
        },
        {
          "skill_id": "kwp-engineering-tech-debt",
          "tokens": 648
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "kwp-design-design-critique",
          "tokens": 671
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "mirofish-graph-explorer",
          "tokens": 1175
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "docx-template-engine",
          "tokens": 1193
        },
        {
          "skill_id": "paper-auto-classifier",
          "tokens": 1262
        },
        {
          "skill_id": "role-ux-designer",
          "tokens": 1268
        },
        {
          "skill_id": "alphaxiv-paper-lookup",
          "tokens": 1299
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Prompt Architect. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"prompt-architect\", \"helm-validator\", \"prompt-transformer\", \"plans\", \"automation-strategist\", \"kwp-design-user-research\", \"pm-product-discovery\", \"terraform-reviewer\", \"workflow-sequential\", \"workflow-eval-opt\", \"gws-email-reply\", \"defuddle\", \"agency-ux-researcher\", \"notion-meeting-sync\", \"kwp-engineering-tech-debt\", \"kwp-design-design-system-management\", \"kwp-design-design-critique\", \"knowledge-to-paper\", \"mirofish-graph-explorer\", \"role-ceo\", \"docx-template-engine\", \"paper-auto-classifier\", \"role-ux-designer\", \"alphaxiv-paper-lookup\", \"role-cso\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0102",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Sales Perspective Analyzer. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "role-sales",
        "role-ux-designer",
        "role-hr",
        "role-cso",
        "role-finance",
        "role-developer",
        "role-ceo",
        "role-pm",
        "role-security-engineer",
        "role-cto",
        "role-data-scientist",
        "role-trading-expert",
        "deep-research-pipeline",
        "executive-briefing",
        "paper-review",
        "full-stack-planner",
        "role-dispatcher",
        "prd-research-factory",
        "pm-market-research",
        "pm-go-to-market",
        "kwp-brand-voice-guideline-generation",
        "nlm-launch-kit",
        "meeting-digest",
        "kwp-engineering-documentation",
        "agency-data-consolidation-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "role-sales",
          "tokens": 1221
        },
        {
          "skill_id": "role-ux-designer",
          "tokens": 1268
        },
        {
          "skill_id": "role-hr",
          "tokens": 1350
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        },
        {
          "skill_id": "role-finance",
          "tokens": 1353
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "role-pm",
          "tokens": 1304
        },
        {
          "skill_id": "role-security-engineer",
          "tokens": 1406
        },
        {
          "skill_id": "role-cto",
          "tokens": 1413
        },
        {
          "skill_id": "role-data-scientist",
          "tokens": 1695
        },
        {
          "skill_id": "role-trading-expert",
          "tokens": 1791
        },
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        },
        {
          "skill_id": "executive-briefing",
          "tokens": 1419
        },
        {
          "skill_id": "paper-review",
          "tokens": 5815
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        },
        {
          "skill_id": "role-dispatcher",
          "tokens": 2075
        },
        {
          "skill_id": "prd-research-factory",
          "tokens": 648
        },
        {
          "skill_id": "pm-market-research",
          "tokens": 966
        },
        {
          "skill_id": "pm-go-to-market",
          "tokens": 992
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "nlm-launch-kit",
          "tokens": 2368
        },
        {
          "skill_id": "meeting-digest",
          "tokens": 4491
        },
        {
          "skill_id": "kwp-engineering-documentation",
          "tokens": 654
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Sales Perspective Analyzer. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"role-sales\", \"role-ux-designer\", \"role-hr\", \"role-cso\", \"role-finance\", \"role-developer\", \"role-ceo\", \"role-pm\", \"role-security-engineer\", \"role-cto\", \"role-data-scientist\", \"role-trading-expert\", \"deep-research-pipeline\", \"executive-briefing\", \"paper-review\", \"full-stack-planner\", \"role-dispatcher\", \"prd-research-factory\", \"pm-market-research\", \"pm-go-to-market\", \"kwp-brand-voice-guideline-generation\", \"nlm-launch-kit\", \"meeting-digest\", \"kwp-engineering-documentation\", \"agency-data-consolidation-agent\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0103",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Video Compress. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "video-compress",
        "transcribee",
        "design-architect",
        "test-suite",
        "anthropic-pdf",
        "plans",
        "kwp-design-accessibility-review",
        "kwp-productivity-task-management",
        "hf-repos",
        "hf-buckets",
        "prompt-transformer",
        "agency-wechat-official-account-manager",
        "skill-optimizer",
        "kwp-human-resources-org-planning",
        "kwp-engineering-code-review",
        "gws-recipe-audit-sharing",
        "kwp-engineering-tech-debt",
        "kwp-design-design-system-management",
        "kwp-design-design-critique",
        "gws-docs",
        "domain-commit",
        "gws-sheets",
        "gws-workflows",
        "role-ceo",
        "role-sales"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "video-compress",
          "tokens": 1230
        },
        {
          "skill_id": "transcribee",
          "tokens": 1255
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "anthropic-pdf",
          "tokens": 2219
        },
        {
          "skill_id": "plans",
          "tokens": 2297
        },
        {
          "skill_id": "kwp-design-accessibility-review",
          "tokens": 679
        },
        {
          "skill_id": "kwp-productivity-task-management",
          "tokens": 776
        },
        {
          "skill_id": "hf-repos",
          "tokens": 1106
        },
        {
          "skill_id": "hf-buckets",
          "tokens": 1334
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "skill-optimizer",
          "tokens": 2940
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-engineering-code-review",
          "tokens": 610
        },
        {
          "skill_id": "gws-recipe-audit-sharing",
          "tokens": 645
        },
        {
          "skill_id": "kwp-engineering-tech-debt",
          "tokens": 648
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "kwp-design-design-critique",
          "tokens": 671
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "domain-commit",
          "tokens": 915
        },
        {
          "skill_id": "gws-sheets",
          "tokens": 918
        },
        {
          "skill_id": "gws-workflows",
          "tokens": 952
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "role-sales",
          "tokens": 1221
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Video Compress. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"video-compress\", \"transcribee\", \"design-architect\", \"test-suite\", \"anthropic-pdf\", \"plans\", \"kwp-design-accessibility-review\", \"kwp-productivity-task-management\", \"hf-repos\", \"hf-buckets\", \"prompt-transformer\", \"agency-wechat-official-account-manager\", \"skill-optimizer\", \"kwp-human-resources-org-planning\", \"kwp-engineering-code-review\", \"gws-recipe-audit-sharing\", \"kwp-engineering-tech-debt\", \"kwp-design-design-system-management\", \"kwp-design-design-critique\", \"gws-docs\", \"domain-commit\", \"gws-sheets\", \"gws-workflows\", \"role-ceo\", \"role-sales\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0104",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Guideline Generation. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-brand-voice-guideline-generation",
        "kwp-brand-voice-discover-brand",
        "kwp-brand-voice-brand-voice-enforcement",
        "executive-briefing",
        "cognee",
        "issue-to-code",
        "nlm-slides",
        "nlm-video",
        "agency-feedback-synthesizer",
        "pipeline-builder",
        "kwp-cowork-create-cowork-plugin",
        "kwp-human-resources-people-analytics",
        "daily-strategy-post",
        "role-sales",
        "skill-guide-generator",
        "nlm-strategy-hub",
        "docs-tutor-setup",
        "nlm-launch-kit",
        "kwp-marketing-brand-voice",
        "kwp-cowork-create-plugin",
        "nlm-deep-learn",
        "kwp-design-design-system-management",
        "knowledge-to-paper",
        "agency-data-consolidation-agent",
        "agency-data-analytics-reporter"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "kwp-brand-voice-discover-brand",
          "tokens": 1897
        },
        {
          "skill_id": "kwp-brand-voice-brand-voice-enforcement",
          "tokens": 1634
        },
        {
          "skill_id": "executive-briefing",
          "tokens": 1419
        },
        {
          "skill_id": "cognee",
          "tokens": 2070
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "nlm-slides",
          "tokens": 1685
        },
        {
          "skill_id": "nlm-video",
          "tokens": 1763
        },
        {
          "skill_id": "agency-feedback-synthesizer",
          "tokens": 2041
        },
        {
          "skill_id": "pipeline-builder",
          "tokens": 2234
        },
        {
          "skill_id": "kwp-cowork-create-cowork-plugin",
          "tokens": 3017
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "role-sales",
          "tokens": 1221
        },
        {
          "skill_id": "skill-guide-generator",
          "tokens": 1858
        },
        {
          "skill_id": "nlm-strategy-hub",
          "tokens": 2207
        },
        {
          "skill_id": "docs-tutor-setup",
          "tokens": 2288
        },
        {
          "skill_id": "nlm-launch-kit",
          "tokens": 2368
        },
        {
          "skill_id": "kwp-marketing-brand-voice",
          "tokens": 2392
        },
        {
          "skill_id": "kwp-cowork-create-plugin",
          "tokens": 3214
        },
        {
          "skill_id": "nlm-deep-learn",
          "tokens": 3214
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        },
        {
          "skill_id": "agency-data-analytics-reporter",
          "tokens": 1003
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Guideline Generation. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-brand-voice-guideline-generation\", \"kwp-brand-voice-discover-brand\", \"kwp-brand-voice-brand-voice-enforcement\", \"executive-briefing\", \"cognee\", \"issue-to-code\", \"nlm-slides\", \"nlm-video\", \"agency-feedback-synthesizer\", \"pipeline-builder\", \"kwp-cowork-create-cowork-plugin\", \"kwp-human-resources-people-analytics\", \"daily-strategy-post\", \"role-sales\", \"skill-guide-generator\", \"nlm-strategy-hub\", \"docs-tutor-setup\", \"nlm-launch-kit\", \"kwp-marketing-brand-voice\", \"kwp-cowork-create-plugin\", \"nlm-deep-learn\", \"kwp-design-design-system-management\", \"knowledge-to-paper\", \"agency-data-consolidation-agent\", \"agency-data-analytics-reporter\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0105",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Coding Standards & Best Practices. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "ecc-coding-standards",
        "backend-expert",
        "ui-suite",
        "code-review-all",
        "kwp-engineering-tech-debt",
        "frontend-expert",
        "deep-review",
        "agency-api-tester",
        "agency-legal-compliance-checker",
        "agency-test-results-analyzer",
        "kwp-engineering-code-review",
        "kwp-design-design-system-management",
        "anthropic-frontend-design",
        "sre-devops-expert",
        "db-expert",
        "evals-skills",
        "performance-profiler",
        "design-architect",
        "agency-social-media-strategist",
        "agency-frontend-developer",
        "simplify",
        "agency-xiaohongshu-specialist",
        "agency-studio-operations",
        "agency-wechat-official-account-manager",
        "agency-tool-evaluator"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "ecc-coding-standards",
          "tokens": 1050
        },
        {
          "skill_id": "backend-expert",
          "tokens": 1132
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "code-review-all",
          "tokens": 2524
        },
        {
          "skill_id": "kwp-engineering-tech-debt",
          "tokens": 648
        },
        {
          "skill_id": "frontend-expert",
          "tokens": 1234
        },
        {
          "skill_id": "deep-review",
          "tokens": 1923
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "kwp-engineering-code-review",
          "tokens": 610
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "anthropic-frontend-design",
          "tokens": 1312
        },
        {
          "skill_id": "sre-devops-expert",
          "tokens": 1450
        },
        {
          "skill_id": "db-expert",
          "tokens": 1523
        },
        {
          "skill_id": "evals-skills",
          "tokens": 1524
        },
        {
          "skill_id": "performance-profiler",
          "tokens": 1738
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "agency-social-media-strategist",
          "tokens": 2084
        },
        {
          "skill_id": "agency-frontend-developer",
          "tokens": 2533
        },
        {
          "skill_id": "simplify",
          "tokens": 2596
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-tool-evaluator",
          "tokens": 4549
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Coding Standards & Best Practices. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"ecc-coding-standards\", \"backend-expert\", \"ui-suite\", \"code-review-all\", \"kwp-engineering-tech-debt\", \"frontend-expert\", \"deep-review\", \"agency-api-tester\", \"agency-legal-compliance-checker\", \"agency-test-results-analyzer\", \"kwp-engineering-code-review\", \"kwp-design-design-system-management\", \"anthropic-frontend-design\", \"sre-devops-expert\", \"db-expert\", \"evals-skills\", \"performance-profiler\", \"design-architect\", \"agency-social-media-strategist\", \"agency-frontend-developer\", \"simplify\", \"agency-xiaohongshu-specialist\", \"agency-studio-operations\", \"agency-wechat-official-account-manager\", \"agency-tool-evaluator\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0106",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: MiroFish Orchestrator. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "mirofish",
        "mirofish-financial-sim",
        "mirofish-opinion-sim",
        "mirofish-graph-explorer",
        "nlm-strategy-hub",
        "ai-workflow-integrator",
        "knowledge-to-paper",
        "ui-suite",
        "hf-trending-intelligence",
        "nlm-discovery-lab",
        "test-suite",
        "deep-review",
        "cognee",
        "deep-research-pipeline",
        "nlm-deep-learn",
        "daily-strategy-post",
        "portfolio-report-generator",
        "paperclip-agents",
        "full-stack-planner",
        "context-engineer",
        "nlm-launch-kit",
        "simplify",
        "incident-to-improvement",
        "cursor-automations",
        "morning-ship"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "mirofish",
          "tokens": 3657
        },
        {
          "skill_id": "mirofish-financial-sim",
          "tokens": 1330
        },
        {
          "skill_id": "mirofish-opinion-sim",
          "tokens": 1383
        },
        {
          "skill_id": "mirofish-graph-explorer",
          "tokens": 1175
        },
        {
          "skill_id": "nlm-strategy-hub",
          "tokens": 2207
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "hf-trending-intelligence",
          "tokens": 1832
        },
        {
          "skill_id": "nlm-discovery-lab",
          "tokens": 2808
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "deep-review",
          "tokens": 1923
        },
        {
          "skill_id": "cognee",
          "tokens": 2070
        },
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        },
        {
          "skill_id": "nlm-deep-learn",
          "tokens": 3214
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "portfolio-report-generator",
          "tokens": 1406
        },
        {
          "skill_id": "paperclip-agents",
          "tokens": 1887
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        },
        {
          "skill_id": "context-engineer",
          "tokens": 2324
        },
        {
          "skill_id": "nlm-launch-kit",
          "tokens": 2368
        },
        {
          "skill_id": "simplify",
          "tokens": 2596
        },
        {
          "skill_id": "incident-to-improvement",
          "tokens": 2624
        },
        {
          "skill_id": "cursor-automations",
          "tokens": 2911
        },
        {
          "skill_id": "morning-ship",
          "tokens": 2943
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: MiroFish Orchestrator. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"mirofish\", \"mirofish-financial-sim\", \"mirofish-opinion-sim\", \"mirofish-graph-explorer\", \"nlm-strategy-hub\", \"ai-workflow-integrator\", \"knowledge-to-paper\", \"ui-suite\", \"hf-trending-intelligence\", \"nlm-discovery-lab\", \"test-suite\", \"deep-review\", \"cognee\", \"deep-research-pipeline\", \"nlm-deep-learn\", \"daily-strategy-post\", \"portfolio-report-generator\", \"paperclip-agents\", \"full-stack-planner\", \"context-engineer\", \"nlm-launch-kit\", \"simplify\", \"incident-to-improvement\", \"cursor-automations\", \"morning-ship\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0107",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: MCP Server Development Guide. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "anthropic-mcp-builder",
        "hf-trackio",
        "anthropic-claude-api",
        "skill-seekers",
        "public-apis",
        "kwp-data-interactive-dashboard-builder",
        "setup-doctor",
        "ecc-coding-standards",
        "prompt-transformer",
        "mission-control copy",
        "mission-control",
        "kwp-data-data-visualization",
        "ecc-continuous-learning",
        "agency-api-tester",
        "ecc-configure",
        "hf-evaluation",
        "planning-meeting-sentinel",
        "gws-chat",
        "kwp-engineering-system-design",
        "ci-quality-gate",
        "ppt-template-engine",
        "hf-endpoints",
        "gws-workspace",
        "docx-template-engine",
        "anthropic-frontend-design"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "anthropic-mcp-builder",
          "tokens": 2825
        },
        {
          "skill_id": "hf-trackio",
          "tokens": 1424
        },
        {
          "skill_id": "anthropic-claude-api",
          "tokens": 4709
        },
        {
          "skill_id": "skill-seekers",
          "tokens": 1568
        },
        {
          "skill_id": "public-apis",
          "tokens": 2246
        },
        {
          "skill_id": "kwp-data-interactive-dashboard-builder",
          "tokens": 2350
        },
        {
          "skill_id": "setup-doctor",
          "tokens": 2578
        },
        {
          "skill_id": "ecc-coding-standards",
          "tokens": 1050
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "kwp-data-data-visualization",
          "tokens": 3020
        },
        {
          "skill_id": "ecc-continuous-learning",
          "tokens": 3352
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "ecc-configure",
          "tokens": 3534
        },
        {
          "skill_id": "hf-evaluation",
          "tokens": 4549
        },
        {
          "skill_id": "planning-meeting-sentinel",
          "tokens": 604
        },
        {
          "skill_id": "gws-chat",
          "tokens": 751
        },
        {
          "skill_id": "kwp-engineering-system-design",
          "tokens": 756
        },
        {
          "skill_id": "ci-quality-gate",
          "tokens": 1054
        },
        {
          "skill_id": "ppt-template-engine",
          "tokens": 1101
        },
        {
          "skill_id": "hf-endpoints",
          "tokens": 1177
        },
        {
          "skill_id": "gws-workspace",
          "tokens": 1186
        },
        {
          "skill_id": "docx-template-engine",
          "tokens": 1193
        },
        {
          "skill_id": "anthropic-frontend-design",
          "tokens": 1312
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: MCP Server Development Guide. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-mcp-builder\", \"hf-trackio\", \"anthropic-claude-api\", \"skill-seekers\", \"public-apis\", \"kwp-data-interactive-dashboard-builder\", \"setup-doctor\", \"ecc-coding-standards\", \"prompt-transformer\", \"mission-control copy\", \"mission-control\", \"kwp-data-data-visualization\", \"ecc-continuous-learning\", \"agency-api-tester\", \"ecc-configure\", \"hf-evaluation\", \"planning-meeting-sentinel\", \"gws-chat\", \"kwp-engineering-system-design\", \"ci-quality-gate\", \"ppt-template-engine\", \"hf-endpoints\", \"gws-workspace\", \"docx-template-engine\", \"anthropic-frontend-design\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0108",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Visual Storyteller Agent. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-visual-storyteller",
        "agency-instagram-curator",
        "agency-xiaohongshu-specialist",
        "agency-developer-advocate",
        "agency-content-creator",
        "agency-twitter-engager",
        "agency-whimsy-injector",
        "agency-zhihu-strategist",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "agency-support-responder",
        "agency-ai-engineer",
        "agency-experiment-tracker",
        "agency-wechat-official-account-manager",
        "agency-trend-researcher",
        "agency-image-prompt-engineer",
        "agency-infrastructure-maintainer",
        "agency-legal-compliance-checker",
        "agency-data-consolidation-agent",
        "agency-studio-operations",
        "agency-workflow-optimizer",
        "agency-studio-producer",
        "agency-api-tester",
        "agency-app-store-optimizer",
        "agency-test-results-analyzer"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-zhihu-strategist",
          "tokens": 3239
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-support-responder",
          "tokens": 3296
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "agency-image-prompt-engineer",
          "tokens": 3027
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-studio-producer",
          "tokens": 2953
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-app-store-optimizer",
          "tokens": 3512
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Visual Storyteller Agent. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-visual-storyteller\", \"agency-instagram-curator\", \"agency-xiaohongshu-specialist\", \"agency-developer-advocate\", \"agency-content-creator\", \"agency-twitter-engager\", \"agency-whimsy-injector\", \"agency-zhihu-strategist\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"agency-support-responder\", \"agency-ai-engineer\", \"agency-experiment-tracker\", \"agency-wechat-official-account-manager\", \"agency-trend-researcher\", \"agency-image-prompt-engineer\", \"agency-infrastructure-maintainer\", \"agency-legal-compliance-checker\", \"agency-data-consolidation-agent\", \"agency-studio-operations\", \"agency-workflow-optimizer\", \"agency-studio-producer\", \"agency-api-tester\", \"agency-app-store-optimizer\", \"agency-test-results-analyzer\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0109",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-turtle-refresh. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "tab-turtle-refresh",
        "tab-bollinger-refresh",
        "tab-dualma-refresh",
        "tab-strategy-comparison",
        "tab-market-breadth",
        "daily-stock-check",
        "tab-llm-agents",
        "tab-technical-analysis",
        "tab-market-environment",
        "tab-report-generate",
        "tab-fundamental-sync",
        "tab-stock-sync",
        "tab-hot-stock-discovery",
        "tab-screening",
        "tab-news-fetch",
        "tab-patterns",
        "alphaear-predictor",
        "weekly-stock-update",
        "trading-backtest-expert",
        "tab-genai-features",
        "tab-kiwoom",
        "tab-glossary",
        "trading-trade-hypothesis-ideator",
        "stock-csv-downloader",
        "morning-ship"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "tab-turtle-refresh",
          "tokens": 371
        },
        {
          "skill_id": "tab-bollinger-refresh",
          "tokens": 383
        },
        {
          "skill_id": "tab-dualma-refresh",
          "tokens": 391
        },
        {
          "skill_id": "tab-strategy-comparison",
          "tokens": 571
        },
        {
          "skill_id": "tab-market-breadth",
          "tokens": 504
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "tab-llm-agents",
          "tokens": 382
        },
        {
          "skill_id": "tab-technical-analysis",
          "tokens": 414
        },
        {
          "skill_id": "tab-market-environment",
          "tokens": 566
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "tab-fundamental-sync",
          "tokens": 331
        },
        {
          "skill_id": "tab-stock-sync",
          "tokens": 345
        },
        {
          "skill_id": "tab-hot-stock-discovery",
          "tokens": 359
        },
        {
          "skill_id": "tab-screening",
          "tokens": 410
        },
        {
          "skill_id": "tab-news-fetch",
          "tokens": 474
        },
        {
          "skill_id": "tab-patterns",
          "tokens": 509
        },
        {
          "skill_id": "alphaear-predictor",
          "tokens": 803
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "trading-backtest-expert",
          "tokens": 2754
        },
        {
          "skill_id": "tab-genai-features",
          "tokens": 354
        },
        {
          "skill_id": "tab-kiwoom",
          "tokens": 452
        },
        {
          "skill_id": "tab-glossary",
          "tokens": 538
        },
        {
          "skill_id": "trading-trade-hypothesis-ideator",
          "tokens": 909
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "morning-ship",
          "tokens": 2943
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-turtle-refresh. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"tab-turtle-refresh\", \"tab-bollinger-refresh\", \"tab-dualma-refresh\", \"tab-strategy-comparison\", \"tab-market-breadth\", \"daily-stock-check\", \"tab-llm-agents\", \"tab-technical-analysis\", \"tab-market-environment\", \"tab-report-generate\", \"tab-fundamental-sync\", \"tab-stock-sync\", \"tab-hot-stock-discovery\", \"tab-screening\", \"tab-news-fetch\", \"tab-patterns\", \"alphaear-predictor\", \"weekly-stock-update\", \"trading-backtest-expert\", \"tab-genai-features\", \"tab-kiwoom\", \"tab-glossary\", \"trading-trade-hypothesis-ideator\", \"stock-csv-downloader\", \"morning-ship\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0110",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: XR Interface Architect Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-xr-interface-architect",
        "agency-xr-cockpit-interaction-specialist",
        "agency-agentic-identity-trust-architect",
        "agency-xr-immersive-developer",
        "agency-growth-hacker",
        "agency-content-creator",
        "agency-visionos-spatial-engineer",
        "agency-behavioral-nudge-engine",
        "agency-cultural-intelligence-strategist",
        "agency-social-media-strategist",
        "agency-tiktok-strategist",
        "agency-autonomous-optimization-architect",
        "agency-wechat-official-account-manager",
        "agency-ux-architect",
        "agency-brand-guardian",
        "agency-macos-spatial-metal-engineer",
        "agency-zhihu-strategist",
        "agency-data-consolidation-agent",
        "agency-sales-data-extraction-agent",
        "agency-terminal-integration-specialist",
        "agency-senior-project-manager",
        "agency-instagram-curator",
        "agency-inclusive-visuals-specialist",
        "agency-feedback-synthesizer",
        "agency-visual-storyteller"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-xr-interface-architect",
          "tokens": 657
        },
        {
          "skill_id": "agency-xr-cockpit-interaction-specialist",
          "tokens": 693
        },
        {
          "skill_id": "agency-agentic-identity-trust-architect",
          "tokens": 4533
        },
        {
          "skill_id": "agency-xr-immersive-developer",
          "tokens": 697
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "agency-visionos-spatial-engineer",
          "tokens": 1091
        },
        {
          "skill_id": "agency-behavioral-nudge-engine",
          "tokens": 1553
        },
        {
          "skill_id": "agency-cultural-intelligence-strategist",
          "tokens": 1892
        },
        {
          "skill_id": "agency-social-media-strategist",
          "tokens": 2084
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-autonomous-optimization-architect",
          "tokens": 2220
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-ux-architect",
          "tokens": 3119
        },
        {
          "skill_id": "agency-brand-guardian",
          "tokens": 3135
        },
        {
          "skill_id": "agency-macos-spatial-metal-engineer",
          "tokens": 3229
        },
        {
          "skill_id": "agency-zhihu-strategist",
          "tokens": 3239
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        },
        {
          "skill_id": "agency-sales-data-extraction-agent",
          "tokens": 922
        },
        {
          "skill_id": "agency-terminal-integration-specialist",
          "tokens": 1273
        },
        {
          "skill_id": "agency-senior-project-manager",
          "tokens": 1555
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-inclusive-visuals-specialist",
          "tokens": 1931
        },
        {
          "skill_id": "agency-feedback-synthesizer",
          "tokens": 2041
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: XR Interface Architect Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-xr-interface-architect\", \"agency-xr-cockpit-interaction-specialist\", \"agency-agentic-identity-trust-architect\", \"agency-xr-immersive-developer\", \"agency-growth-hacker\", \"agency-content-creator\", \"agency-visionos-spatial-engineer\", \"agency-behavioral-nudge-engine\", \"agency-cultural-intelligence-strategist\", \"agency-social-media-strategist\", \"agency-tiktok-strategist\", \"agency-autonomous-optimization-architect\", \"agency-wechat-official-account-manager\", \"agency-ux-architect\", \"agency-brand-guardian\", \"agency-macos-spatial-metal-engineer\", \"agency-zhihu-strategist\", \"agency-data-consolidation-agent\", \"agency-sales-data-extraction-agent\", \"agency-terminal-integration-specialist\", \"agency-senior-project-manager\", \"agency-instagram-curator\", \"agency-inclusive-visuals-specialist\", \"agency-feedback-synthesizer\", \"agency-visual-storyteller\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0111",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: UX Expert. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "ux-expert",
        "design-architect",
        "frontend-expert",
        "ui-suite",
        "kwp-design-accessibility-review",
        "kwp-design-design-system-management",
        "compliance-governance",
        "kwp-design-user-research",
        "ci-quality-gate",
        "role-ux-designer",
        "performance-profiler",
        "agency-accessibility-auditor",
        "kwp-design-design-critique",
        "backend-expert",
        "i18n-sync",
        "anthropic-frontend-design",
        "evals-skills",
        "security-expert",
        "agency-instagram-curator",
        "agency-visual-storyteller",
        "agency-ai-engineer",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "agency-twitter-engager",
        "code-review-all"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "ux-expert",
          "tokens": 1210
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "frontend-expert",
          "tokens": 1234
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "kwp-design-accessibility-review",
          "tokens": 679
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "compliance-governance",
          "tokens": 1501
        },
        {
          "skill_id": "kwp-design-user-research",
          "tokens": 718
        },
        {
          "skill_id": "ci-quality-gate",
          "tokens": 1054
        },
        {
          "skill_id": "role-ux-designer",
          "tokens": 1268
        },
        {
          "skill_id": "performance-profiler",
          "tokens": 1738
        },
        {
          "skill_id": "agency-accessibility-auditor",
          "tokens": 4031
        },
        {
          "skill_id": "kwp-design-design-critique",
          "tokens": 671
        },
        {
          "skill_id": "backend-expert",
          "tokens": 1132
        },
        {
          "skill_id": "i18n-sync",
          "tokens": 1282
        },
        {
          "skill_id": "anthropic-frontend-design",
          "tokens": 1312
        },
        {
          "skill_id": "evals-skills",
          "tokens": 1524
        },
        {
          "skill_id": "security-expert",
          "tokens": 1562
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "code-review-all",
          "tokens": 2524
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: UX Expert. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"ux-expert\", \"design-architect\", \"frontend-expert\", \"ui-suite\", \"kwp-design-accessibility-review\", \"kwp-design-design-system-management\", \"compliance-governance\", \"kwp-design-user-research\", \"ci-quality-gate\", \"role-ux-designer\", \"performance-profiler\", \"agency-accessibility-auditor\", \"kwp-design-design-critique\", \"backend-expert\", \"i18n-sync\", \"anthropic-frontend-design\", \"evals-skills\", \"security-expert\", \"agency-instagram-curator\", \"agency-visual-storyteller\", \"agency-ai-engineer\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"agency-twitter-engager\", \"code-review-all\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0112",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Automation Strategist. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "automation-strategist",
        "pipeline-builder",
        "system-thinker",
        "qa-test-expert",
        "prompt-architect",
        "kwp-human-resources-org-planning",
        "kwp-engineering-testing-strategy",
        "kwp-engineering-tech-debt",
        "trading-trade-hypothesis-ideator",
        "agency-data-analytics-reporter",
        "trading-options-strategy-advisor",
        "codebase-archaeologist",
        "trading-position-sizer",
        "ai-workflow-integrator",
        "plans",
        "kwp-bio-research-scientific-problem-selection",
        "agency-analytics-reporter",
        "kwp-design-design-system-management",
        "daily-strategy-post",
        "pm-execution",
        "frontend-expert",
        "knowledge-daily-aggregator",
        "issue-to-code",
        "design-architect",
        "hf-trending-intelligence"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "automation-strategist",
          "tokens": 2482
        },
        {
          "skill_id": "pipeline-builder",
          "tokens": 2234
        },
        {
          "skill_id": "system-thinker",
          "tokens": 2684
        },
        {
          "skill_id": "qa-test-expert",
          "tokens": 1517
        },
        {
          "skill_id": "prompt-architect",
          "tokens": 1939
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "kwp-engineering-tech-debt",
          "tokens": 648
        },
        {
          "skill_id": "trading-trade-hypothesis-ideator",
          "tokens": 909
        },
        {
          "skill_id": "agency-data-analytics-reporter",
          "tokens": 1003
        },
        {
          "skill_id": "trading-options-strategy-advisor",
          "tokens": 1657
        },
        {
          "skill_id": "codebase-archaeologist",
          "tokens": 1977
        },
        {
          "skill_id": "trading-position-sizer",
          "tokens": 2013
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "plans",
          "tokens": 2297
        },
        {
          "skill_id": "kwp-bio-research-scientific-problem-selection",
          "tokens": 2990
        },
        {
          "skill_id": "agency-analytics-reporter",
          "tokens": 4157
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "pm-execution",
          "tokens": 1178
        },
        {
          "skill_id": "frontend-expert",
          "tokens": 1234
        },
        {
          "skill_id": "knowledge-daily-aggregator",
          "tokens": 1425
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "hf-trending-intelligence",
          "tokens": 1832
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Automation Strategist. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"automation-strategist\", \"pipeline-builder\", \"system-thinker\", \"qa-test-expert\", \"prompt-architect\", \"kwp-human-resources-org-planning\", \"kwp-engineering-testing-strategy\", \"kwp-engineering-tech-debt\", \"trading-trade-hypothesis-ideator\", \"agency-data-analytics-reporter\", \"trading-options-strategy-advisor\", \"codebase-archaeologist\", \"trading-position-sizer\", \"ai-workflow-integrator\", \"plans\", \"kwp-bio-research-scientific-problem-selection\", \"agency-analytics-reporter\", \"kwp-design-design-system-management\", \"daily-strategy-post\", \"pm-execution\", \"frontend-expert\", \"knowledge-daily-aggregator\", \"issue-to-code\", \"design-architect\", \"hf-trending-intelligence\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0113",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Agentic Identity & Trust Architect. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-agentic-identity-trust-architect",
        "agency-ux-architect",
        "agency-xr-interface-architect",
        "agency-xr-cockpit-interaction-specialist",
        "agency-autonomous-optimization-architect",
        "agency-agents-orchestrator",
        "agency-content-creator",
        "agency-instagram-curator",
        "agency-cultural-intelligence-strategist",
        "agency-inclusive-visuals-specialist",
        "agency-ai-engineer",
        "agency-wechat-official-account-manager",
        "agency-studio-producer",
        "agency-lsp-index-engineer",
        "agency-brand-guardian",
        "agency-macos-spatial-metal-engineer",
        "agency-support-responder",
        "agency-api-tester",
        "agency-data-engineer",
        "agency-xr-immersive-developer",
        "agency-data-consolidation-agent",
        "agency-sales-data-extraction-agent",
        "agency-growth-hacker",
        "agency-visionos-spatial-engineer",
        "agency-terminal-integration-specialist"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-agentic-identity-trust-architect",
          "tokens": 4533
        },
        {
          "skill_id": "agency-ux-architect",
          "tokens": 3119
        },
        {
          "skill_id": "agency-xr-interface-architect",
          "tokens": 657
        },
        {
          "skill_id": "agency-xr-cockpit-interaction-specialist",
          "tokens": 693
        },
        {
          "skill_id": "agency-autonomous-optimization-architect",
          "tokens": 2220
        },
        {
          "skill_id": "agency-agents-orchestrator",
          "tokens": 4128
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-cultural-intelligence-strategist",
          "tokens": 1892
        },
        {
          "skill_id": "agency-inclusive-visuals-specialist",
          "tokens": 1931
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-studio-producer",
          "tokens": 2953
        },
        {
          "skill_id": "agency-lsp-index-engineer",
          "tokens": 3006
        },
        {
          "skill_id": "agency-brand-guardian",
          "tokens": 3135
        },
        {
          "skill_id": "agency-macos-spatial-metal-engineer",
          "tokens": 3229
        },
        {
          "skill_id": "agency-support-responder",
          "tokens": 3296
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-data-engineer",
          "tokens": 3843
        },
        {
          "skill_id": "agency-xr-immersive-developer",
          "tokens": 697
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        },
        {
          "skill_id": "agency-sales-data-extraction-agent",
          "tokens": 922
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-visionos-spatial-engineer",
          "tokens": 1091
        },
        {
          "skill_id": "agency-terminal-integration-specialist",
          "tokens": 1273
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Agentic Identity & Trust Architect. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-agentic-identity-trust-architect\", \"agency-ux-architect\", \"agency-xr-interface-architect\", \"agency-xr-cockpit-interaction-specialist\", \"agency-autonomous-optimization-architect\", \"agency-agents-orchestrator\", \"agency-content-creator\", \"agency-instagram-curator\", \"agency-cultural-intelligence-strategist\", \"agency-inclusive-visuals-specialist\", \"agency-ai-engineer\", \"agency-wechat-official-account-manager\", \"agency-studio-producer\", \"agency-lsp-index-engineer\", \"agency-brand-guardian\", \"agency-macos-spatial-metal-engineer\", \"agency-support-responder\", \"agency-api-tester\", \"agency-data-engineer\", \"agency-xr-immersive-developer\", \"agency-data-consolidation-agent\", \"agency-sales-data-extraction-agent\", \"agency-growth-hacker\", \"agency-visionos-spatial-engineer\", \"agency-terminal-integration-specialist\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0114",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: md-to-slack-canvas. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "md-to-slack-canvas",
        "md-to-notion",
        "gws-docs",
        "hf-buckets",
        "nlm-video",
        "anthropic-pdf",
        "slack-agent",
        "cursor-automations",
        "gws-chat",
        "domain-commit",
        "nlm-slides",
        "md-enhance-publish",
        "release-ship",
        "twitter-timeline-to-slack",
        "gws-sheets",
        "hf-repos",
        "office-template-enforcer",
        "stock-csv-downloader",
        "issue-to-code",
        "x-to-slack",
        "kwp-brand-voice-guideline-generation",
        "docs-tutor-setup",
        "skill-optimizer",
        "kwp-cowork-create-cowork-plugin",
        "cursor-sync"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "md-to-slack-canvas",
          "tokens": 1960
        },
        {
          "skill_id": "md-to-notion",
          "tokens": 2096
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "hf-buckets",
          "tokens": 1334
        },
        {
          "skill_id": "nlm-video",
          "tokens": 1763
        },
        {
          "skill_id": "anthropic-pdf",
          "tokens": 2219
        },
        {
          "skill_id": "slack-agent",
          "tokens": 2564
        },
        {
          "skill_id": "cursor-automations",
          "tokens": 2911
        },
        {
          "skill_id": "gws-chat",
          "tokens": 751
        },
        {
          "skill_id": "domain-commit",
          "tokens": 915
        },
        {
          "skill_id": "nlm-slides",
          "tokens": 1685
        },
        {
          "skill_id": "md-enhance-publish",
          "tokens": 1740
        },
        {
          "skill_id": "release-ship",
          "tokens": 3056
        },
        {
          "skill_id": "twitter-timeline-to-slack",
          "tokens": 6229
        },
        {
          "skill_id": "gws-sheets",
          "tokens": 918
        },
        {
          "skill_id": "hf-repos",
          "tokens": 1106
        },
        {
          "skill_id": "office-template-enforcer",
          "tokens": 1273
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "x-to-slack",
          "tokens": 2001
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "docs-tutor-setup",
          "tokens": 2288
        },
        {
          "skill_id": "skill-optimizer",
          "tokens": 2940
        },
        {
          "skill_id": "kwp-cowork-create-cowork-plugin",
          "tokens": 3017
        },
        {
          "skill_id": "cursor-sync",
          "tokens": 3267
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: md-to-slack-canvas. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"md-to-slack-canvas\", \"md-to-notion\", \"gws-docs\", \"hf-buckets\", \"nlm-video\", \"anthropic-pdf\", \"slack-agent\", \"cursor-automations\", \"gws-chat\", \"domain-commit\", \"nlm-slides\", \"md-enhance-publish\", \"release-ship\", \"twitter-timeline-to-slack\", \"gws-sheets\", \"hf-repos\", \"office-template-enforcer\", \"stock-csv-downloader\", \"issue-to-code\", \"x-to-slack\", \"kwp-brand-voice-guideline-generation\", \"docs-tutor-setup\", \"skill-optimizer\", \"kwp-cowork-create-cowork-plugin\", \"cursor-sync\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0115",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: HR Perspective Analyzer. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "role-hr",
        "role-finance",
        "role-developer",
        "role-security-engineer",
        "role-sales",
        "role-ux-designer",
        "role-pm",
        "role-cto",
        "role-cso",
        "role-data-scientist",
        "role-ceo",
        "role-trading-expert",
        "deep-research-pipeline",
        "executive-briefing",
        "kwp-operations-change-management",
        "paper-review",
        "role-dispatcher",
        "problem-definition",
        "agency-project-shepherd",
        "meeting-digest",
        "kwp-operations-compliance-tracking",
        "pm-data-analytics",
        "alphaxiv-paper-lookup",
        "refactor-simulator",
        "full-stack-planner"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "role-hr",
          "tokens": 1350
        },
        {
          "skill_id": "role-finance",
          "tokens": 1353
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "role-security-engineer",
          "tokens": 1406
        },
        {
          "skill_id": "role-sales",
          "tokens": 1221
        },
        {
          "skill_id": "role-ux-designer",
          "tokens": 1268
        },
        {
          "skill_id": "role-pm",
          "tokens": 1304
        },
        {
          "skill_id": "role-cto",
          "tokens": 1413
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        },
        {
          "skill_id": "role-data-scientist",
          "tokens": 1695
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "role-trading-expert",
          "tokens": 1791
        },
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        },
        {
          "skill_id": "executive-briefing",
          "tokens": 1419
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "paper-review",
          "tokens": 5815
        },
        {
          "skill_id": "role-dispatcher",
          "tokens": 2075
        },
        {
          "skill_id": "problem-definition",
          "tokens": 2601
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "meeting-digest",
          "tokens": 4491
        },
        {
          "skill_id": "kwp-operations-compliance-tracking",
          "tokens": 686
        },
        {
          "skill_id": "pm-data-analytics",
          "tokens": 956
        },
        {
          "skill_id": "alphaxiv-paper-lookup",
          "tokens": 1299
        },
        {
          "skill_id": "refactor-simulator",
          "tokens": 2044
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: HR Perspective Analyzer. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"role-hr\", \"role-finance\", \"role-developer\", \"role-security-engineer\", \"role-sales\", \"role-ux-designer\", \"role-pm\", \"role-cto\", \"role-cso\", \"role-data-scientist\", \"role-ceo\", \"role-trading-expert\", \"deep-research-pipeline\", \"executive-briefing\", \"kwp-operations-change-management\", \"paper-review\", \"role-dispatcher\", \"problem-definition\", \"agency-project-shepherd\", \"meeting-digest\", \"kwp-operations-compliance-tracking\", \"pm-data-analytics\", \"alphaxiv-paper-lookup\", \"refactor-simulator\", \"full-stack-planner\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0116",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Deep Research Pipeline. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "deep-research-pipeline",
        "role-ceo",
        "role-cso",
        "role-sales",
        "role-developer",
        "role-ux-designer",
        "role-cto",
        "role-pm",
        "role-hr",
        "role-finance",
        "role-security-engineer",
        "paper-review",
        "email-research-dispatcher",
        "critical-review",
        "daily-strategy-post",
        "executive-briefing",
        "ui-suite",
        "hf-trending-intelligence",
        "role-dispatcher",
        "auto-research",
        "meeting-digest",
        "weekly-status-report",
        "portfolio-report-generator",
        "role-data-scientist",
        "full-stack-planner"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        },
        {
          "skill_id": "role-sales",
          "tokens": 1221
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "role-ux-designer",
          "tokens": 1268
        },
        {
          "skill_id": "role-cto",
          "tokens": 1413
        },
        {
          "skill_id": "role-pm",
          "tokens": 1304
        },
        {
          "skill_id": "role-hr",
          "tokens": 1350
        },
        {
          "skill_id": "role-finance",
          "tokens": 1353
        },
        {
          "skill_id": "role-security-engineer",
          "tokens": 1406
        },
        {
          "skill_id": "paper-review",
          "tokens": 5815
        },
        {
          "skill_id": "email-research-dispatcher",
          "tokens": 703
        },
        {
          "skill_id": "critical-review",
          "tokens": 3039
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "executive-briefing",
          "tokens": 1419
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "hf-trending-intelligence",
          "tokens": 1832
        },
        {
          "skill_id": "role-dispatcher",
          "tokens": 2075
        },
        {
          "skill_id": "auto-research",
          "tokens": 2139
        },
        {
          "skill_id": "meeting-digest",
          "tokens": 4491
        },
        {
          "skill_id": "weekly-status-report",
          "tokens": 1282
        },
        {
          "skill_id": "portfolio-report-generator",
          "tokens": 1406
        },
        {
          "skill_id": "role-data-scientist",
          "tokens": 1695
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Deep Research Pipeline. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"deep-research-pipeline\", \"role-ceo\", \"role-cso\", \"role-sales\", \"role-developer\", \"role-ux-designer\", \"role-cto\", \"role-pm\", \"role-hr\", \"role-finance\", \"role-security-engineer\", \"paper-review\", \"email-research-dispatcher\", \"critical-review\", \"daily-strategy-post\", \"executive-briefing\", \"ui-suite\", \"hf-trending-intelligence\", \"role-dispatcher\", \"auto-research\", \"meeting-digest\", \"weekly-status-report\", \"portfolio-report-generator\", \"role-data-scientist\", \"full-stack-planner\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0117",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: App Store Optimizer Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-app-store-optimizer",
        "agency-instagram-curator",
        "agency-tiktok-strategist",
        "agency-reddit-community-builder",
        "agency-twitter-engager",
        "agency-xiaohongshu-specialist",
        "agency-wechat-official-account-manager",
        "agency-workflow-optimizer",
        "agency-zhihu-strategist",
        "agency-infrastructure-maintainer",
        "agency-tool-evaluator",
        "agency-visual-storyteller",
        "agency-whimsy-injector",
        "agency-studio-operations",
        "agency-support-responder",
        "agency-api-tester",
        "agency-test-results-analyzer",
        "agency-growth-hacker",
        "agency-terminal-integration-specialist",
        "agency-inclusive-visuals-specialist",
        "agency-ai-engineer",
        "agency-experiment-tracker",
        "agency-project-shepherd",
        "agency-trend-researcher",
        "agency-legal-compliance-checker"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-app-store-optimizer",
          "tokens": 3512
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-zhihu-strategist",
          "tokens": 3239
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-tool-evaluator",
          "tokens": 4549
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-support-responder",
          "tokens": 3296
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-terminal-integration-specialist",
          "tokens": 1273
        },
        {
          "skill_id": "agency-inclusive-visuals-specialist",
          "tokens": 1931
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: App Store Optimizer Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-app-store-optimizer\", \"agency-instagram-curator\", \"agency-tiktok-strategist\", \"agency-reddit-community-builder\", \"agency-twitter-engager\", \"agency-xiaohongshu-specialist\", \"agency-wechat-official-account-manager\", \"agency-workflow-optimizer\", \"agency-zhihu-strategist\", \"agency-infrastructure-maintainer\", \"agency-tool-evaluator\", \"agency-visual-storyteller\", \"agency-whimsy-injector\", \"agency-studio-operations\", \"agency-support-responder\", \"agency-api-tester\", \"agency-test-results-analyzer\", \"agency-growth-hacker\", \"agency-terminal-integration-specialist\", \"agency-inclusive-visuals-specialist\", \"agency-ai-engineer\", \"agency-experiment-tracker\", \"agency-project-shepherd\", \"agency-trend-researcher\", \"agency-legal-compliance-checker\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0118",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Weekly Prep Brief. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-common-room-weekly-prep-brief",
        "kwp-common-room-call-prep",
        "kwp-common-room-compose-outreach",
        "kwp-common-room-account-research",
        "kwp-common-room-prospect",
        "kwp-sales-daily-briefing",
        "kwp-common-room-contact-research",
        "kwp-brand-voice-discover-brand",
        "planning-weekly-pulse",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-operations-change-management",
        "kwp-operations-resource-planning",
        "kwp-finance-journal-entry-prep",
        "kwp-sales-call-prep",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-legal-meeting-briefing",
        "kwp-marketing-content-creation",
        "kwp-legal-canned-responses",
        "kwp-design-system",
        "kwp-operations-vendor-management",
        "kwp-operations-process-optimization",
        "kwp-design-design-handoff",
        "kwp-apollo-enrich-lead",
        "mission-control copy"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-common-room-weekly-prep-brief",
          "tokens": 1607
        },
        {
          "skill_id": "kwp-common-room-call-prep",
          "tokens": 1729
        },
        {
          "skill_id": "kwp-common-room-compose-outreach",
          "tokens": 1627
        },
        {
          "skill_id": "kwp-common-room-account-research",
          "tokens": 1784
        },
        {
          "skill_id": "kwp-common-room-prospect",
          "tokens": 1803
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "kwp-common-room-contact-research",
          "tokens": 1611
        },
        {
          "skill_id": "kwp-brand-voice-discover-brand",
          "tokens": 1897
        },
        {
          "skill_id": "planning-weekly-pulse",
          "tokens": 528
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-finance-journal-entry-prep",
          "tokens": 2215
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "kwp-legal-canned-responses",
          "tokens": 3164
        },
        {
          "skill_id": "kwp-design-system",
          "tokens": 4960
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "kwp-design-design-handoff",
          "tokens": 649
        },
        {
          "skill_id": "kwp-apollo-enrich-lead",
          "tokens": 837
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Weekly Prep Brief. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-common-room-weekly-prep-brief\", \"kwp-common-room-call-prep\", \"kwp-common-room-compose-outreach\", \"kwp-common-room-account-research\", \"kwp-common-room-prospect\", \"kwp-sales-daily-briefing\", \"kwp-common-room-contact-research\", \"kwp-brand-voice-discover-brand\", \"planning-weekly-pulse\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-operations-change-management\", \"kwp-operations-resource-planning\", \"kwp-finance-journal-entry-prep\", \"kwp-sales-call-prep\", \"kwp-bio-research-single-cell-rna-qc\", \"kwp-legal-meeting-briefing\", \"kwp-marketing-content-creation\", \"kwp-legal-canned-responses\", \"kwp-design-system\", \"kwp-operations-vendor-management\", \"kwp-operations-process-optimization\", \"kwp-design-design-handoff\", \"kwp-apollo-enrich-lead\", \"mission-control copy\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0119",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Design System Management. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-design-design-system-management",
        "kwp-design-design-critique",
        "kwp-engineering-tech-debt",
        "kwp-design-accessibility-review",
        "kwp-engineering-system-design",
        "kwp-design-user-research",
        "kwp-engineering-code-review",
        "kwp-engineering-testing-strategy",
        "kwp-engineering-incident-response",
        "kwp-engineering-documentation",
        "ux-expert",
        "frontend-expert",
        "i18n-sync",
        "design-architect",
        "agency-ui-designer",
        "kwp-design-ux-writing",
        "kwp-operations-compliance-tracking",
        "kwp-operations-risk-assessment",
        "role-ux-designer",
        "agency-visual-storyteller",
        "agency-backend-architect",
        "agency-studio-operations",
        "agency-infrastructure-maintainer",
        "agency-legal-compliance-checker",
        "agency-xr-immersive-developer"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "kwp-design-design-critique",
          "tokens": 671
        },
        {
          "skill_id": "kwp-engineering-tech-debt",
          "tokens": 648
        },
        {
          "skill_id": "kwp-design-accessibility-review",
          "tokens": 679
        },
        {
          "skill_id": "kwp-engineering-system-design",
          "tokens": 756
        },
        {
          "skill_id": "kwp-design-user-research",
          "tokens": 718
        },
        {
          "skill_id": "kwp-engineering-code-review",
          "tokens": 610
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "kwp-engineering-incident-response",
          "tokens": 645
        },
        {
          "skill_id": "kwp-engineering-documentation",
          "tokens": 654
        },
        {
          "skill_id": "ux-expert",
          "tokens": 1210
        },
        {
          "skill_id": "frontend-expert",
          "tokens": 1234
        },
        {
          "skill_id": "i18n-sync",
          "tokens": 1282
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "agency-ui-designer",
          "tokens": 3538
        },
        {
          "skill_id": "kwp-design-ux-writing",
          "tokens": 675
        },
        {
          "skill_id": "kwp-operations-compliance-tracking",
          "tokens": 686
        },
        {
          "skill_id": "kwp-operations-risk-assessment",
          "tokens": 808
        },
        {
          "skill_id": "role-ux-designer",
          "tokens": 1268
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-backend-architect",
          "tokens": 2600
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-xr-immersive-developer",
          "tokens": 697
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Design System Management. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-design-design-system-management\", \"kwp-design-design-critique\", \"kwp-engineering-tech-debt\", \"kwp-design-accessibility-review\", \"kwp-engineering-system-design\", \"kwp-design-user-research\", \"kwp-engineering-code-review\", \"kwp-engineering-testing-strategy\", \"kwp-engineering-incident-response\", \"kwp-engineering-documentation\", \"ux-expert\", \"frontend-expert\", \"i18n-sync\", \"design-architect\", \"agency-ui-designer\", \"kwp-design-ux-writing\", \"kwp-operations-compliance-tracking\", \"kwp-operations-risk-assessment\", \"role-ux-designer\", \"agency-visual-storyteller\", \"agency-backend-architect\", \"agency-studio-operations\", \"agency-infrastructure-maintainer\", \"agency-legal-compliance-checker\", \"agency-xr-immersive-developer\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0120",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Stock CSV Downloader. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "stock-csv-downloader",
        "weekly-stock-update",
        "daily-stock-check",
        "alphaear-stock",
        "alphaear-deepear-lite",
        "trading-sector-analyst",
        "alphaear-search",
        "alphaear-news",
        "trading-technical-analyst",
        "tab-stock-sync",
        "trading-market-environment-analysis",
        "alphaear-reporter",
        "trading-trade-hypothesis-ideator",
        "weekly-status-report",
        "ai-workflow-integrator",
        "trading-backtest-expert",
        "trading-us-stock-analysis",
        "trading-theme-detector",
        "tab-kiwoom",
        "trading-market-breadth-analyzer",
        "trading-edge-candidate-agent",
        "trading-us-market-bubble-detector",
        "tab-report-generate",
        "github-sprint-digest",
        "gws-recipe-events-from-sheet"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "alphaear-stock",
          "tokens": 820
        },
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        },
        {
          "skill_id": "trading-sector-analyst",
          "tokens": 2657
        },
        {
          "skill_id": "alphaear-search",
          "tokens": 672
        },
        {
          "skill_id": "alphaear-news",
          "tokens": 873
        },
        {
          "skill_id": "trading-technical-analyst",
          "tokens": 2868
        },
        {
          "skill_id": "tab-stock-sync",
          "tokens": 345
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "alphaear-reporter",
          "tokens": 810
        },
        {
          "skill_id": "trading-trade-hypothesis-ideator",
          "tokens": 909
        },
        {
          "skill_id": "weekly-status-report",
          "tokens": 1282
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "trading-backtest-expert",
          "tokens": 2754
        },
        {
          "skill_id": "trading-us-stock-analysis",
          "tokens": 2926
        },
        {
          "skill_id": "trading-theme-detector",
          "tokens": 3487
        },
        {
          "skill_id": "tab-kiwoom",
          "tokens": 452
        },
        {
          "skill_id": "trading-market-breadth-analyzer",
          "tokens": 1830
        },
        {
          "skill_id": "trading-edge-candidate-agent",
          "tokens": 1971
        },
        {
          "skill_id": "trading-us-market-bubble-detector",
          "tokens": 3062
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "gws-recipe-events-from-sheet",
          "tokens": 674
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Stock CSV Downloader. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"stock-csv-downloader\", \"weekly-stock-update\", \"daily-stock-check\", \"alphaear-stock\", \"alphaear-deepear-lite\", \"trading-sector-analyst\", \"alphaear-search\", \"alphaear-news\", \"trading-technical-analyst\", \"tab-stock-sync\", \"trading-market-environment-analysis\", \"alphaear-reporter\", \"trading-trade-hypothesis-ideator\", \"weekly-status-report\", \"ai-workflow-integrator\", \"trading-backtest-expert\", \"trading-us-stock-analysis\", \"trading-theme-detector\", \"tab-kiwoom\", \"trading-market-breadth-analyzer\", \"trading-edge-candidate-agent\", \"trading-us-market-bubble-detector\", \"tab-report-generate\", \"github-sprint-digest\", \"gws-recipe-events-from-sheet\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0121",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Design System Guide. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-design-system",
        "kwp-design-design-system-management",
        "kwp-data-data-visualization",
        "kwp-cowork-create-plugin",
        "kwp-operations-vendor-management",
        "kwp-human-resources-org-planning",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "ui-suite",
        "kwp-finance-journal-entry-prep",
        "kwp-data-interactive-dashboard-builder",
        "kwp-data-data-exploration",
        "kwp-product-management-roadmap-management",
        "kwp-legal-meeting-briefing",
        "kwp-marketing-content-creation",
        "kwp-customer-support-knowledge-management",
        "kwp-finance-financial-statements",
        "kwp-customer-support-response-drafting",
        "kwp-cowork-create-cowork-plugin",
        "kwp-legal-contract-review",
        "kwp-product-management-metrics-tracking",
        "kwp-human-resources-people-analytics",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-engineering-tech-debt"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-design-system",
          "tokens": 4960
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "kwp-data-data-visualization",
          "tokens": 3020
        },
        {
          "skill_id": "kwp-cowork-create-plugin",
          "tokens": 3214
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "kwp-finance-journal-entry-prep",
          "tokens": 2215
        },
        {
          "skill_id": "kwp-data-interactive-dashboard-builder",
          "tokens": 2350
        },
        {
          "skill_id": "kwp-data-data-exploration",
          "tokens": 2358
        },
        {
          "skill_id": "kwp-product-management-roadmap-management",
          "tokens": 2467
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "kwp-customer-support-knowledge-management",
          "tokens": 2624
        },
        {
          "skill_id": "kwp-finance-financial-statements",
          "tokens": 2768
        },
        {
          "skill_id": "kwp-customer-support-response-drafting",
          "tokens": 2944
        },
        {
          "skill_id": "kwp-cowork-create-cowork-plugin",
          "tokens": 3017
        },
        {
          "skill_id": "kwp-legal-contract-review",
          "tokens": 3137
        },
        {
          "skill_id": "kwp-product-management-metrics-tracking",
          "tokens": 3606
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-human-resources-recruiting-pipeline",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-engineering-tech-debt",
          "tokens": 648
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Design System Guide. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-design-system\", \"kwp-design-design-system-management\", \"kwp-data-data-visualization\", \"kwp-cowork-create-plugin\", \"kwp-operations-vendor-management\", \"kwp-human-resources-org-planning\", \"kwp-operations-process-optimization\", \"kwp-operations-change-management\", \"ui-suite\", \"kwp-finance-journal-entry-prep\", \"kwp-data-interactive-dashboard-builder\", \"kwp-data-data-exploration\", \"kwp-product-management-roadmap-management\", \"kwp-legal-meeting-briefing\", \"kwp-marketing-content-creation\", \"kwp-customer-support-knowledge-management\", \"kwp-finance-financial-statements\", \"kwp-customer-support-response-drafting\", \"kwp-cowork-create-cowork-plugin\", \"kwp-legal-contract-review\", \"kwp-product-management-metrics-tracking\", \"kwp-human-resources-people-analytics\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-engineering-tech-debt\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0122",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Recruiting Pipeline. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-interview-prep",
        "kwp-operations-process-optimization",
        "kwp-human-resources-people-analytics",
        "kwp-operations-vendor-management",
        "kwp-human-resources-org-planning",
        "kwp-operations-change-management",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-operations-resource-planning",
        "kwp-sales-call-prep",
        "kwp-finance-close-management",
        "commit-to-issue",
        "paperclip-tasks",
        "kwp-sales-daily-briefing",
        "kwp-sales-account-research",
        "mission-control",
        "kwp-legal-meeting-briefing",
        "kwp-product-management-feature-spec",
        "kwp-sales-competitive-intelligence",
        "kwp-product-management-metrics-tracking",
        "kwp-engineering-testing-strategy",
        "kwp-human-resources-employee-handbook",
        "kwp-engineering-incident-response",
        "kwp-design-design-handoff",
        "kwp-design-design-system-management"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-human-resources-recruiting-pipeline",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-finance-close-management",
          "tokens": 2782
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "paperclip-tasks",
          "tokens": 1534
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-product-management-feature-spec",
          "tokens": 2614
        },
        {
          "skill_id": "kwp-sales-competitive-intelligence",
          "tokens": 3361
        },
        {
          "skill_id": "kwp-product-management-metrics-tracking",
          "tokens": 3606
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "kwp-human-resources-employee-handbook",
          "tokens": 629
        },
        {
          "skill_id": "kwp-engineering-incident-response",
          "tokens": 645
        },
        {
          "skill_id": "kwp-design-design-handoff",
          "tokens": 649
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Recruiting Pipeline. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-human-resources-recruiting-pipeline\", \"kwp-human-resources-interview-prep\", \"kwp-operations-process-optimization\", \"kwp-human-resources-people-analytics\", \"kwp-operations-vendor-management\", \"kwp-human-resources-org-planning\", \"kwp-operations-change-management\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-operations-resource-planning\", \"kwp-sales-call-prep\", \"kwp-finance-close-management\", \"commit-to-issue\", \"paperclip-tasks\", \"kwp-sales-daily-briefing\", \"kwp-sales-account-research\", \"mission-control\", \"kwp-legal-meeting-briefing\", \"kwp-product-management-feature-spec\", \"kwp-sales-competitive-intelligence\", \"kwp-product-management-metrics-tracking\", \"kwp-engineering-testing-strategy\", \"kwp-human-resources-employee-handbook\", \"kwp-engineering-incident-response\", \"kwp-design-design-handoff\", \"kwp-design-design-system-management\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0123",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Skill Composer. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "skill-composer",
        "autoskill-extractor",
        "workflow-miner",
        "autoskill-judge",
        "skill-guide-generator",
        "ai-workflow-integrator",
        "agency-roster",
        "kwp-sync",
        "release-commander",
        "anthropic-skill-creator",
        "autoskill-merger",
        "autoskill-evolve",
        "hf-auto-finetune-planner",
        "pipeline-builder",
        "setup-doctor",
        "skill-optimizer",
        "kwp-cowork-create-cowork-plugin",
        "knowledge-to-paper",
        "docx-template-engine",
        "skill-seekers",
        "prompt-transformer",
        "kwp-brand-voice-guideline-generation",
        "mission-control",
        "sefo-orchestrator",
        "pm-data-analytics"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "skill-composer",
          "tokens": 599
        },
        {
          "skill_id": "autoskill-extractor",
          "tokens": 1235
        },
        {
          "skill_id": "workflow-miner",
          "tokens": 893
        },
        {
          "skill_id": "autoskill-judge",
          "tokens": 1176
        },
        {
          "skill_id": "skill-guide-generator",
          "tokens": 1858
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "agency-roster",
          "tokens": 3411
        },
        {
          "skill_id": "kwp-sync",
          "tokens": 1472
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "anthropic-skill-creator",
          "tokens": 6862
        },
        {
          "skill_id": "autoskill-merger",
          "tokens": 1142
        },
        {
          "skill_id": "autoskill-evolve",
          "tokens": 1566
        },
        {
          "skill_id": "hf-auto-finetune-planner",
          "tokens": 2091
        },
        {
          "skill_id": "pipeline-builder",
          "tokens": 2234
        },
        {
          "skill_id": "setup-doctor",
          "tokens": 2578
        },
        {
          "skill_id": "skill-optimizer",
          "tokens": 2940
        },
        {
          "skill_id": "kwp-cowork-create-cowork-plugin",
          "tokens": 3017
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "docx-template-engine",
          "tokens": 1193
        },
        {
          "skill_id": "skill-seekers",
          "tokens": 1568
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "sefo-orchestrator",
          "tokens": 821
        },
        {
          "skill_id": "pm-data-analytics",
          "tokens": 956
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Skill Composer. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"skill-composer\", \"autoskill-extractor\", \"workflow-miner\", \"autoskill-judge\", \"skill-guide-generator\", \"ai-workflow-integrator\", \"agency-roster\", \"kwp-sync\", \"release-commander\", \"anthropic-skill-creator\", \"autoskill-merger\", \"autoskill-evolve\", \"hf-auto-finetune-planner\", \"pipeline-builder\", \"setup-doctor\", \"skill-optimizer\", \"kwp-cowork-create-cowork-plugin\", \"knowledge-to-paper\", \"docx-template-engine\", \"skill-seekers\", \"prompt-transformer\", \"kwp-brand-voice-guideline-generation\", \"mission-control\", \"sefo-orchestrator\", \"pm-data-analytics\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0124",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AgentsOrchestrator Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-agents-orchestrator",
        "agency-agentic-identity-trust-architect",
        "agency-data-consolidation-agent",
        "agency-autonomous-optimization-architect",
        "agency-studio-operations",
        "agency-workflow-optimizer",
        "agency-report-distribution-agent",
        "agency-behavioral-nudge-engine",
        "agency-senior-project-manager",
        "agency-instagram-curator",
        "agency-cultural-intelligence-strategist",
        "agency-visual-storyteller",
        "agency-ai-engineer",
        "agency-whimsy-injector",
        "agency-rapid-prototyper",
        "agency-experiment-tracker",
        "agency-mobile-app-builder",
        "agency-project-shepherd",
        "agency-wechat-official-account-manager",
        "agency-trend-researcher",
        "agency-studio-producer",
        "agency-image-prompt-engineer",
        "agency-brand-guardian",
        "agency-xr-interface-architect",
        "agency-xr-cockpit-interaction-specialist"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-agents-orchestrator",
          "tokens": 4128
        },
        {
          "skill_id": "agency-agentic-identity-trust-architect",
          "tokens": 4533
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        },
        {
          "skill_id": "agency-autonomous-optimization-architect",
          "tokens": 2220
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-report-distribution-agent",
          "tokens": 907
        },
        {
          "skill_id": "agency-behavioral-nudge-engine",
          "tokens": 1553
        },
        {
          "skill_id": "agency-senior-project-manager",
          "tokens": 1555
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-cultural-intelligence-strategist",
          "tokens": 1892
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-rapid-prototyper",
          "tokens": 2496
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-mobile-app-builder",
          "tokens": 2595
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "agency-studio-producer",
          "tokens": 2953
        },
        {
          "skill_id": "agency-image-prompt-engineer",
          "tokens": 3027
        },
        {
          "skill_id": "agency-brand-guardian",
          "tokens": 3135
        },
        {
          "skill_id": "agency-xr-interface-architect",
          "tokens": 657
        },
        {
          "skill_id": "agency-xr-cockpit-interaction-specialist",
          "tokens": 693
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AgentsOrchestrator Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-agents-orchestrator\", \"agency-agentic-identity-trust-architect\", \"agency-data-consolidation-agent\", \"agency-autonomous-optimization-architect\", \"agency-studio-operations\", \"agency-workflow-optimizer\", \"agency-report-distribution-agent\", \"agency-behavioral-nudge-engine\", \"agency-senior-project-manager\", \"agency-instagram-curator\", \"agency-cultural-intelligence-strategist\", \"agency-visual-storyteller\", \"agency-ai-engineer\", \"agency-whimsy-injector\", \"agency-rapid-prototyper\", \"agency-experiment-tracker\", \"agency-mobile-app-builder\", \"agency-project-shepherd\", \"agency-wechat-official-account-manager\", \"agency-trend-researcher\", \"agency-studio-producer\", \"agency-image-prompt-engineer\", \"agency-brand-guardian\", \"agency-xr-interface-architect\", \"agency-xr-cockpit-interaction-specialist\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0125",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Slack Search. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-slack-slack-search",
        "kwp-operations-change-management",
        "kwp-operations-resource-planning",
        "kwp-brand-voice-discover-brand",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-data-interactive-dashboard-builder",
        "kwp-enterprise-search-knowledge-synthesis",
        "kwp-customer-support-customer-research",
        "kwp-marketing-content-creation",
        "kwp-marketing-campaign-planning",
        "kwp-human-resources-people-analytics",
        "kwp-operations-vendor-management",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-operations-process-optimization",
        "gws-recipe-audit-sharing",
        "kwp-design-design-handoff",
        "kwp-human-resources-interview-prep",
        "kwp-apollo-enrich-lead",
        "kwp-slack-slack-messaging",
        "kwp-apollo-prospect",
        "gws-calendar",
        "kwp-enterprise-search-source-management",
        "kwp-common-room-compose-outreach",
        "kwp-common-room-prospect"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-slack-slack-search",
          "tokens": 1186
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-brand-voice-discover-brand",
          "tokens": 1897
        },
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "kwp-data-interactive-dashboard-builder",
          "tokens": 2350
        },
        {
          "skill_id": "kwp-enterprise-search-knowledge-synthesis",
          "tokens": 2363
        },
        {
          "skill_id": "kwp-customer-support-customer-research",
          "tokens": 2529
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "kwp-marketing-campaign-planning",
          "tokens": 2803
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-recruiting-pipeline",
          "tokens": 578
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "gws-recipe-audit-sharing",
          "tokens": 645
        },
        {
          "skill_id": "kwp-design-design-handoff",
          "tokens": 649
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "kwp-apollo-enrich-lead",
          "tokens": 837
        },
        {
          "skill_id": "kwp-slack-slack-messaging",
          "tokens": 987
        },
        {
          "skill_id": "kwp-apollo-prospect",
          "tokens": 1006
        },
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "kwp-enterprise-search-source-management",
          "tokens": 1624
        },
        {
          "skill_id": "kwp-common-room-compose-outreach",
          "tokens": 1627
        },
        {
          "skill_id": "kwp-common-room-prospect",
          "tokens": 1803
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Slack Search. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-slack-slack-search\", \"kwp-operations-change-management\", \"kwp-operations-resource-planning\", \"kwp-brand-voice-discover-brand\", \"kwp-bio-research-single-cell-rna-qc\", \"kwp-data-interactive-dashboard-builder\", \"kwp-enterprise-search-knowledge-synthesis\", \"kwp-customer-support-customer-research\", \"kwp-marketing-content-creation\", \"kwp-marketing-campaign-planning\", \"kwp-human-resources-people-analytics\", \"kwp-operations-vendor-management\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-operations-process-optimization\", \"gws-recipe-audit-sharing\", \"kwp-design-design-handoff\", \"kwp-human-resources-interview-prep\", \"kwp-apollo-enrich-lead\", \"kwp-slack-slack-messaging\", \"kwp-apollo-prospect\", \"gws-calendar\", \"kwp-enterprise-search-source-management\", \"kwp-common-room-compose-outreach\", \"kwp-common-room-prospect\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0126",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-market-breadth. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "tab-market-breadth",
        "tab-market-environment",
        "tab-turtle-refresh",
        "tab-bollinger-refresh",
        "tab-dualma-refresh",
        "tab-news-fetch",
        "tab-llm-agents",
        "tab-dashboard-summary",
        "tab-patterns",
        "tab-glossary",
        "tab-strategy-comparison",
        "tab-report-generate",
        "tab-technical-analysis",
        "daily-stock-check",
        "tab-fundamental-sync",
        "tab-stock-sync",
        "weekly-stock-update",
        "tab-hot-stock-discovery",
        "tab-kiwoom",
        "alphaear-news",
        "stock-csv-downloader",
        "trading-market-environment-analysis",
        "trading-market-breadth-analyzer",
        "trading-sector-analyst",
        "trading-technical-analyst"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "tab-market-breadth",
          "tokens": 504
        },
        {
          "skill_id": "tab-market-environment",
          "tokens": 566
        },
        {
          "skill_id": "tab-turtle-refresh",
          "tokens": 371
        },
        {
          "skill_id": "tab-bollinger-refresh",
          "tokens": 383
        },
        {
          "skill_id": "tab-dualma-refresh",
          "tokens": 391
        },
        {
          "skill_id": "tab-news-fetch",
          "tokens": 474
        },
        {
          "skill_id": "tab-llm-agents",
          "tokens": 382
        },
        {
          "skill_id": "tab-dashboard-summary",
          "tokens": 419
        },
        {
          "skill_id": "tab-patterns",
          "tokens": 509
        },
        {
          "skill_id": "tab-glossary",
          "tokens": 538
        },
        {
          "skill_id": "tab-strategy-comparison",
          "tokens": 571
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "tab-technical-analysis",
          "tokens": 414
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "tab-fundamental-sync",
          "tokens": 331
        },
        {
          "skill_id": "tab-stock-sync",
          "tokens": 345
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "tab-hot-stock-discovery",
          "tokens": 359
        },
        {
          "skill_id": "tab-kiwoom",
          "tokens": 452
        },
        {
          "skill_id": "alphaear-news",
          "tokens": 873
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "trading-market-breadth-analyzer",
          "tokens": 1830
        },
        {
          "skill_id": "trading-sector-analyst",
          "tokens": 2657
        },
        {
          "skill_id": "trading-technical-analyst",
          "tokens": 2868
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-market-breadth. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"tab-market-breadth\", \"tab-market-environment\", \"tab-turtle-refresh\", \"tab-bollinger-refresh\", \"tab-dualma-refresh\", \"tab-news-fetch\", \"tab-llm-agents\", \"tab-dashboard-summary\", \"tab-patterns\", \"tab-glossary\", \"tab-strategy-comparison\", \"tab-report-generate\", \"tab-technical-analysis\", \"daily-stock-check\", \"tab-fundamental-sync\", \"tab-stock-sync\", \"weekly-stock-update\", \"tab-hot-stock-discovery\", \"tab-kiwoom\", \"alphaear-news\", \"stock-csv-downloader\", \"trading-market-environment-analysis\", \"trading-market-breadth-analyzer\", \"trading-sector-analyst\", \"trading-technical-analyst\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0127",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Design Critique. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-design-design-critique",
        "kwp-design-design-system-management",
        "kwp-engineering-code-review",
        "kwp-engineering-system-design",
        "kwp-engineering-tech-debt",
        "kwp-design-accessibility-review",
        "kwp-design-user-research",
        "kwp-engineering-testing-strategy",
        "kwp-design-ux-writing",
        "kwp-operations-compliance-tracking",
        "kwp-operations-risk-assessment",
        "design-architect",
        "agency-visual-storyteller",
        "agency-agentic-identity-trust-architect",
        "kwp-engineering-incident-response",
        "kwp-engineering-documentation",
        "ux-expert",
        "role-ux-designer",
        "problem-definition",
        "agency-ux-architect",
        "agency-ui-designer",
        "kwp-operations-vendor-management",
        "agency-xr-interface-architect",
        "agency-xr-immersive-developer",
        "agency-data-consolidation-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-design-design-critique",
          "tokens": 671
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "kwp-engineering-code-review",
          "tokens": 610
        },
        {
          "skill_id": "kwp-engineering-system-design",
          "tokens": 756
        },
        {
          "skill_id": "kwp-engineering-tech-debt",
          "tokens": 648
        },
        {
          "skill_id": "kwp-design-accessibility-review",
          "tokens": 679
        },
        {
          "skill_id": "kwp-design-user-research",
          "tokens": 718
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "kwp-design-ux-writing",
          "tokens": 675
        },
        {
          "skill_id": "kwp-operations-compliance-tracking",
          "tokens": 686
        },
        {
          "skill_id": "kwp-operations-risk-assessment",
          "tokens": 808
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-agentic-identity-trust-architect",
          "tokens": 4533
        },
        {
          "skill_id": "kwp-engineering-incident-response",
          "tokens": 645
        },
        {
          "skill_id": "kwp-engineering-documentation",
          "tokens": 654
        },
        {
          "skill_id": "ux-expert",
          "tokens": 1210
        },
        {
          "skill_id": "role-ux-designer",
          "tokens": 1268
        },
        {
          "skill_id": "problem-definition",
          "tokens": 2601
        },
        {
          "skill_id": "agency-ux-architect",
          "tokens": 3119
        },
        {
          "skill_id": "agency-ui-designer",
          "tokens": 3538
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "agency-xr-interface-architect",
          "tokens": 657
        },
        {
          "skill_id": "agency-xr-immersive-developer",
          "tokens": 697
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Design Critique. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-design-design-critique\", \"kwp-design-design-system-management\", \"kwp-engineering-code-review\", \"kwp-engineering-system-design\", \"kwp-engineering-tech-debt\", \"kwp-design-accessibility-review\", \"kwp-design-user-research\", \"kwp-engineering-testing-strategy\", \"kwp-design-ux-writing\", \"kwp-operations-compliance-tracking\", \"kwp-operations-risk-assessment\", \"design-architect\", \"agency-visual-storyteller\", \"agency-agentic-identity-trust-architect\", \"kwp-engineering-incident-response\", \"kwp-engineering-documentation\", \"ux-expert\", \"role-ux-designer\", \"problem-definition\", \"agency-ux-architect\", \"agency-ui-designer\", \"kwp-operations-vendor-management\", \"agency-xr-interface-architect\", \"agency-xr-immersive-developer\", \"agency-data-consolidation-agent\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0128",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Cursor Automations. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "cursor-automations",
        "slack-agent",
        "smart-meeting-scheduler",
        "github-sprint-digest",
        "paperclip-agents",
        "pipeline-builder",
        "mirofish-financial-sim",
        "paperclip-tasks",
        "paperclip-setup",
        "gws-calendar",
        "commit-to-issue",
        "ralph-loop",
        "sefo-federation",
        "hf-hub",
        "mirofish-opinion-sim",
        "local-dev-runner",
        "proactive-meeting-scheduler",
        "ui-suite",
        "ship",
        "scrapling",
        "prompt-transformer",
        "md-to-slack-canvas",
        "hf-auto-finetune-planner",
        "context-engineer",
        "kwp-cowork-cowork-plugin-customizer"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "cursor-automations",
          "tokens": 2911
        },
        {
          "skill_id": "slack-agent",
          "tokens": 2564
        },
        {
          "skill_id": "smart-meeting-scheduler",
          "tokens": 639
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "paperclip-agents",
          "tokens": 1887
        },
        {
          "skill_id": "pipeline-builder",
          "tokens": 2234
        },
        {
          "skill_id": "mirofish-financial-sim",
          "tokens": 1330
        },
        {
          "skill_id": "paperclip-tasks",
          "tokens": 1534
        },
        {
          "skill_id": "paperclip-setup",
          "tokens": 1708
        },
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "ralph-loop",
          "tokens": 1925
        },
        {
          "skill_id": "sefo-federation",
          "tokens": 756
        },
        {
          "skill_id": "hf-hub",
          "tokens": 1151
        },
        {
          "skill_id": "mirofish-opinion-sim",
          "tokens": 1383
        },
        {
          "skill_id": "local-dev-runner",
          "tokens": 1419
        },
        {
          "skill_id": "proactive-meeting-scheduler",
          "tokens": 1444
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "scrapling",
          "tokens": 1675
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "md-to-slack-canvas",
          "tokens": 1960
        },
        {
          "skill_id": "hf-auto-finetune-planner",
          "tokens": 2091
        },
        {
          "skill_id": "context-engineer",
          "tokens": 2324
        },
        {
          "skill_id": "kwp-cowork-cowork-plugin-customizer",
          "tokens": 2437
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Cursor Automations. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"cursor-automations\", \"slack-agent\", \"smart-meeting-scheduler\", \"github-sprint-digest\", \"paperclip-agents\", \"pipeline-builder\", \"mirofish-financial-sim\", \"paperclip-tasks\", \"paperclip-setup\", \"gws-calendar\", \"commit-to-issue\", \"ralph-loop\", \"sefo-federation\", \"hf-hub\", \"mirofish-opinion-sim\", \"local-dev-runner\", \"proactive-meeting-scheduler\", \"ui-suite\", \"ship\", \"scrapling\", \"prompt-transformer\", \"md-to-slack-canvas\", \"hf-auto-finetune-planner\", \"context-engineer\", \"kwp-cowork-cowork-plugin-customizer\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0129",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Calendar Daily Briefing. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "calendar-daily-briefing",
        "smart-meeting-scheduler",
        "gws-workflows",
        "gws-calendar",
        "ai-chief-of-staff",
        "proactive-meeting-scheduler",
        "gws-gmail",
        "email-auto-reply",
        "kwp-sales-daily-briefing",
        "morning-ship",
        "bespin-news-digest",
        "gws-recipe-events-from-sheet",
        "gws-workspace",
        "feedback-meeting-scheduler",
        "standup-digest",
        "morning",
        "gws-email-reply",
        "gmail-daily-triage",
        "gws-recipe-personalized-emails",
        "github-sprint-digest",
        "gws-chat",
        "gws-drive",
        "weekly-status-report",
        "setup-doctor",
        "daily-am-orchestrator"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "calendar-daily-briefing",
          "tokens": 1042
        },
        {
          "skill_id": "smart-meeting-scheduler",
          "tokens": 639
        },
        {
          "skill_id": "gws-workflows",
          "tokens": 952
        },
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "ai-chief-of-staff",
          "tokens": 1152
        },
        {
          "skill_id": "proactive-meeting-scheduler",
          "tokens": 1444
        },
        {
          "skill_id": "gws-gmail",
          "tokens": 1796
        },
        {
          "skill_id": "email-auto-reply",
          "tokens": 638
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "morning-ship",
          "tokens": 2943
        },
        {
          "skill_id": "bespin-news-digest",
          "tokens": 3949
        },
        {
          "skill_id": "gws-recipe-events-from-sheet",
          "tokens": 674
        },
        {
          "skill_id": "gws-workspace",
          "tokens": 1186
        },
        {
          "skill_id": "feedback-meeting-scheduler",
          "tokens": 1238
        },
        {
          "skill_id": "standup-digest",
          "tokens": 1330
        },
        {
          "skill_id": "morning",
          "tokens": 2088
        },
        {
          "skill_id": "gws-email-reply",
          "tokens": 2222
        },
        {
          "skill_id": "gmail-daily-triage",
          "tokens": 2229
        },
        {
          "skill_id": "gws-recipe-personalized-emails",
          "tokens": 641
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "gws-chat",
          "tokens": 751
        },
        {
          "skill_id": "gws-drive",
          "tokens": 997
        },
        {
          "skill_id": "weekly-status-report",
          "tokens": 1282
        },
        {
          "skill_id": "setup-doctor",
          "tokens": 2578
        },
        {
          "skill_id": "daily-am-orchestrator",
          "tokens": 3613
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Calendar Daily Briefing. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"calendar-daily-briefing\", \"smart-meeting-scheduler\", \"gws-workflows\", \"gws-calendar\", \"ai-chief-of-staff\", \"proactive-meeting-scheduler\", \"gws-gmail\", \"email-auto-reply\", \"kwp-sales-daily-briefing\", \"morning-ship\", \"bespin-news-digest\", \"gws-recipe-events-from-sheet\", \"gws-workspace\", \"feedback-meeting-scheduler\", \"standup-digest\", \"morning\", \"gws-email-reply\", \"gmail-daily-triage\", \"gws-recipe-personalized-emails\", \"github-sprint-digest\", \"gws-chat\", \"gws-drive\", \"weekly-status-report\", \"setup-doctor\", \"daily-am-orchestrator\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0130",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: User Research Synthesis Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-product-management-user-research-synthesis",
        "agency-feedback-synthesizer",
        "kwp-human-resources-interview-prep",
        "kwp-product-management-feature-spec",
        "kwp-legal-canned-responses",
        "kwp-human-resources-people-analytics",
        "kwp-design-user-research",
        "agency-data-analytics-reporter",
        "pm-product-discovery",
        "kwp-brand-voice-guideline-generation",
        "kwp-sales-call-prep",
        "kwp-customer-support-escalation",
        "kwp-data-interactive-dashboard-builder",
        "kwp-product-management-roadmap-management",
        "kwp-legal-meeting-briefing",
        "kwp-customer-support-customer-research",
        "kwp-marketing-competitive-analysis",
        "nlm-discovery-lab",
        "kwp-data-statistical-analysis",
        "kwp-customer-support-response-drafting",
        "kwp-finance-variance-analysis",
        "agency-ux-researcher",
        "agency-analytics-reporter",
        "feature-intake-router",
        "kwp-operations-vendor-management"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-product-management-user-research-synthesis",
          "tokens": 3071
        },
        {
          "skill_id": "agency-feedback-synthesizer",
          "tokens": 2041
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "kwp-product-management-feature-spec",
          "tokens": 2614
        },
        {
          "skill_id": "kwp-legal-canned-responses",
          "tokens": 3164
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-design-user-research",
          "tokens": 718
        },
        {
          "skill_id": "agency-data-analytics-reporter",
          "tokens": 1003
        },
        {
          "skill_id": "pm-product-discovery",
          "tokens": 1198
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-customer-support-escalation",
          "tokens": 2329
        },
        {
          "skill_id": "kwp-data-interactive-dashboard-builder",
          "tokens": 2350
        },
        {
          "skill_id": "kwp-product-management-roadmap-management",
          "tokens": 2467
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-customer-support-customer-research",
          "tokens": 2529
        },
        {
          "skill_id": "kwp-marketing-competitive-analysis",
          "tokens": 2723
        },
        {
          "skill_id": "nlm-discovery-lab",
          "tokens": 2808
        },
        {
          "skill_id": "kwp-data-statistical-analysis",
          "tokens": 2865
        },
        {
          "skill_id": "kwp-customer-support-response-drafting",
          "tokens": 2944
        },
        {
          "skill_id": "kwp-finance-variance-analysis",
          "tokens": 3208
        },
        {
          "skill_id": "agency-ux-researcher",
          "tokens": 3473
        },
        {
          "skill_id": "agency-analytics-reporter",
          "tokens": 4157
        },
        {
          "skill_id": "feature-intake-router",
          "tokens": 568
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: User Research Synthesis Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-product-management-user-research-synthesis\", \"agency-feedback-synthesizer\", \"kwp-human-resources-interview-prep\", \"kwp-product-management-feature-spec\", \"kwp-legal-canned-responses\", \"kwp-human-resources-people-analytics\", \"kwp-design-user-research\", \"agency-data-analytics-reporter\", \"pm-product-discovery\", \"kwp-brand-voice-guideline-generation\", \"kwp-sales-call-prep\", \"kwp-customer-support-escalation\", \"kwp-data-interactive-dashboard-builder\", \"kwp-product-management-roadmap-management\", \"kwp-legal-meeting-briefing\", \"kwp-customer-support-customer-research\", \"kwp-marketing-competitive-analysis\", \"nlm-discovery-lab\", \"kwp-data-statistical-analysis\", \"kwp-customer-support-response-drafting\", \"kwp-finance-variance-analysis\", \"agency-ux-researcher\", \"agency-analytics-reporter\", \"feature-intake-router\", \"kwp-operations-vendor-management\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0131",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: NotebookLM Studio: Content Generation. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "notebooklm-studio",
        "nlm-deep-learn",
        "notebooklm",
        "nlm-strategy-hub",
        "nlm-slides",
        "nlm-video",
        "nlm-launch-kit",
        "notebooklm-research",
        "nlm-discovery-lab",
        "gws-recipe-doc-from-template",
        "knowledge-to-paper",
        "gws-drive",
        "nlm-arxiv-slides",
        "notion-meeting-sync",
        "gws-docs",
        "gws-sheets",
        "office-template-enforcer",
        "bespin-news-digest",
        "gws-recipe-events-from-sheet",
        "gws-recipe-save-email-to-doc",
        "gws-workspace",
        "hf-buckets",
        "kwp-brand-voice-brand-voice-enforcement",
        "kwp-brand-voice-guideline-generation",
        "cognee"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "notebooklm-studio",
          "tokens": 1812
        },
        {
          "skill_id": "nlm-deep-learn",
          "tokens": 3214
        },
        {
          "skill_id": "notebooklm",
          "tokens": 1844
        },
        {
          "skill_id": "nlm-strategy-hub",
          "tokens": 2207
        },
        {
          "skill_id": "nlm-slides",
          "tokens": 1685
        },
        {
          "skill_id": "nlm-video",
          "tokens": 1763
        },
        {
          "skill_id": "nlm-launch-kit",
          "tokens": 2368
        },
        {
          "skill_id": "notebooklm-research",
          "tokens": 1362
        },
        {
          "skill_id": "nlm-discovery-lab",
          "tokens": 2808
        },
        {
          "skill_id": "gws-recipe-doc-from-template",
          "tokens": 743
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "gws-drive",
          "tokens": 997
        },
        {
          "skill_id": "nlm-arxiv-slides",
          "tokens": 3484
        },
        {
          "skill_id": "notion-meeting-sync",
          "tokens": 4347
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "gws-sheets",
          "tokens": 918
        },
        {
          "skill_id": "office-template-enforcer",
          "tokens": 1273
        },
        {
          "skill_id": "bespin-news-digest",
          "tokens": 3949
        },
        {
          "skill_id": "gws-recipe-events-from-sheet",
          "tokens": 674
        },
        {
          "skill_id": "gws-recipe-save-email-to-doc",
          "tokens": 676
        },
        {
          "skill_id": "gws-workspace",
          "tokens": 1186
        },
        {
          "skill_id": "hf-buckets",
          "tokens": 1334
        },
        {
          "skill_id": "kwp-brand-voice-brand-voice-enforcement",
          "tokens": 1634
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "cognee",
          "tokens": 2070
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: NotebookLM Studio: Content Generation. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"notebooklm-studio\", \"nlm-deep-learn\", \"notebooklm\", \"nlm-strategy-hub\", \"nlm-slides\", \"nlm-video\", \"nlm-launch-kit\", \"notebooklm-research\", \"nlm-discovery-lab\", \"gws-recipe-doc-from-template\", \"knowledge-to-paper\", \"gws-drive\", \"nlm-arxiv-slides\", \"notion-meeting-sync\", \"gws-docs\", \"gws-sheets\", \"office-template-enforcer\", \"bespin-news-digest\", \"gws-recipe-events-from-sheet\", \"gws-recipe-save-email-to-doc\", \"gws-workspace\", \"hf-buckets\", \"kwp-brand-voice-brand-voice-enforcement\", \"kwp-brand-voice-guideline-generation\", \"cognee\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0132",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Ticket Triage Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-customer-support-ticket-triage",
        "kwp-legal-nda-triage",
        "kwp-customer-support-knowledge-management",
        "kwp-customer-support-escalation",
        "kwp-data-data-exploration",
        "tech-support-brain",
        "kwp-legal-legal-risk-assessment",
        "kwp-human-resources-org-planning",
        "sprint-orchestrator",
        "kwp-customer-support-response-drafting",
        "kwp-product-management-user-research-synthesis",
        "kwp-legal-compliance",
        "feature-intake-router",
        "kwp-operations-vendor-management",
        "kwp-operations-resource-planning",
        "kwp-human-resources-interview-prep",
        "jira-linear-adapter",
        "kwp-enterprise-search-source-management",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-finance-reconciliation",
        "kwp-customer-support-customer-research",
        "simplify",
        "kwp-product-management-feature-spec",
        "kwp-product-management-competitive-analysis",
        "kwp-cowork-create-cowork-plugin"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-customer-support-ticket-triage",
          "tokens": 2413
        },
        {
          "skill_id": "kwp-legal-nda-triage",
          "tokens": 3139
        },
        {
          "skill_id": "kwp-customer-support-knowledge-management",
          "tokens": 2624
        },
        {
          "skill_id": "kwp-customer-support-escalation",
          "tokens": 2329
        },
        {
          "skill_id": "kwp-data-data-exploration",
          "tokens": 2358
        },
        {
          "skill_id": "tech-support-brain",
          "tokens": 620
        },
        {
          "skill_id": "kwp-legal-legal-risk-assessment",
          "tokens": 3406
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "sprint-orchestrator",
          "tokens": 1320
        },
        {
          "skill_id": "kwp-customer-support-response-drafting",
          "tokens": 2944
        },
        {
          "skill_id": "kwp-product-management-user-research-synthesis",
          "tokens": 3071
        },
        {
          "skill_id": "kwp-legal-compliance",
          "tokens": 3473
        },
        {
          "skill_id": "feature-intake-router",
          "tokens": 568
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "jira-linear-adapter",
          "tokens": 1365
        },
        {
          "skill_id": "kwp-enterprise-search-source-management",
          "tokens": 1624
        },
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "kwp-finance-reconciliation",
          "tokens": 2287
        },
        {
          "skill_id": "kwp-customer-support-customer-research",
          "tokens": 2529
        },
        {
          "skill_id": "simplify",
          "tokens": 2596
        },
        {
          "skill_id": "kwp-product-management-feature-spec",
          "tokens": 2614
        },
        {
          "skill_id": "kwp-product-management-competitive-analysis",
          "tokens": 2899
        },
        {
          "skill_id": "kwp-cowork-create-cowork-plugin",
          "tokens": 3017
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Ticket Triage Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-customer-support-ticket-triage\", \"kwp-legal-nda-triage\", \"kwp-customer-support-knowledge-management\", \"kwp-customer-support-escalation\", \"kwp-data-data-exploration\", \"tech-support-brain\", \"kwp-legal-legal-risk-assessment\", \"kwp-human-resources-org-planning\", \"sprint-orchestrator\", \"kwp-customer-support-response-drafting\", \"kwp-product-management-user-research-synthesis\", \"kwp-legal-compliance\", \"feature-intake-router\", \"kwp-operations-vendor-management\", \"kwp-operations-resource-planning\", \"kwp-human-resources-interview-prep\", \"jira-linear-adapter\", \"kwp-enterprise-search-source-management\", \"kwp-bio-research-single-cell-rna-qc\", \"kwp-finance-reconciliation\", \"kwp-customer-support-customer-research\", \"simplify\", \"kwp-product-management-feature-spec\", \"kwp-product-management-competitive-analysis\", \"kwp-cowork-create-cowork-plugin\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0133",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Contact Research. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-common-room-contact-research",
        "kwp-common-room-compose-outreach",
        "kwp-common-room-account-research",
        "kwp-common-room-prospect",
        "kwp-sales-account-research",
        "kwp-apollo-enrich-lead",
        "kwp-sales-draft-outreach",
        "kwp-common-room-weekly-prep-brief",
        "kwp-common-room-call-prep",
        "kwp-marketing-content-creation",
        "mission-control copy",
        "kwp-data-data-context-extractor",
        "kwp-sales-call-prep",
        "kwp-bio-research-single-cell-rna-qc",
        "mission-control",
        "kwp-legal-meeting-briefing",
        "kwp-customer-support-customer-research",
        "kwp-product-management-user-research-synthesis",
        "kwp-legal-canned-responses",
        "kwp-human-resources-people-analytics",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-operations-resource-planning",
        "kwp-sales-daily-briefing",
        "kwp-enterprise-search-search-strategy"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-common-room-contact-research",
          "tokens": 1611
        },
        {
          "skill_id": "kwp-common-room-compose-outreach",
          "tokens": 1627
        },
        {
          "skill_id": "kwp-common-room-account-research",
          "tokens": 1784
        },
        {
          "skill_id": "kwp-common-room-prospect",
          "tokens": 1803
        },
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "kwp-apollo-enrich-lead",
          "tokens": 837
        },
        {
          "skill_id": "kwp-sales-draft-outreach",
          "tokens": 2782
        },
        {
          "skill_id": "kwp-common-room-weekly-prep-brief",
          "tokens": 1607
        },
        {
          "skill_id": "kwp-common-room-call-prep",
          "tokens": 1729
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "kwp-data-data-context-extractor",
          "tokens": 1883
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-customer-support-customer-research",
          "tokens": 2529
        },
        {
          "skill_id": "kwp-product-management-user-research-synthesis",
          "tokens": 3071
        },
        {
          "skill_id": "kwp-legal-canned-responses",
          "tokens": 3164
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "kwp-enterprise-search-search-strategy",
          "tokens": 2009
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Contact Research. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-common-room-contact-research\", \"kwp-common-room-compose-outreach\", \"kwp-common-room-account-research\", \"kwp-common-room-prospect\", \"kwp-sales-account-research\", \"kwp-apollo-enrich-lead\", \"kwp-sales-draft-outreach\", \"kwp-common-room-weekly-prep-brief\", \"kwp-common-room-call-prep\", \"kwp-marketing-content-creation\", \"mission-control copy\", \"kwp-data-data-context-extractor\", \"kwp-sales-call-prep\", \"kwp-bio-research-single-cell-rna-qc\", \"mission-control\", \"kwp-legal-meeting-briefing\", \"kwp-customer-support-customer-research\", \"kwp-product-management-user-research-synthesis\", \"kwp-legal-canned-responses\", \"kwp-human-resources-people-analytics\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-operations-resource-planning\", \"kwp-sales-daily-briefing\", \"kwp-enterprise-search-search-strategy\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0134",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Incident to Improvement. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "incident-to-improvement",
        "ui-suite",
        "test-suite",
        "pr-to-issue-linker",
        "role-developer",
        "diagnose",
        "release-commander",
        "e2e-overhaul",
        "problem-definition",
        "bespin-news-digest",
        "issue-to-code",
        "infra-runbook-generator",
        "hf-trending-intelligence",
        "deep-research-pipeline",
        "onboarding-accelerator",
        "kwp-engineering-incident-response",
        "github-sprint-digest",
        "deep-review",
        "codebase-archaeologist",
        "ai-quality-evaluator",
        "skill-optimizer",
        "critical-review",
        "intent-alignment-tracker",
        "incident-lifecycle-manager",
        "email-research-dispatcher"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "incident-to-improvement",
          "tokens": 2624
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "pr-to-issue-linker",
          "tokens": 1223
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "diagnose",
          "tokens": 1433
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "e2e-overhaul",
          "tokens": 2425
        },
        {
          "skill_id": "problem-definition",
          "tokens": 2601
        },
        {
          "skill_id": "bespin-news-digest",
          "tokens": 3949
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "infra-runbook-generator",
          "tokens": 1623
        },
        {
          "skill_id": "hf-trending-intelligence",
          "tokens": 1832
        },
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        },
        {
          "skill_id": "onboarding-accelerator",
          "tokens": 2419
        },
        {
          "skill_id": "kwp-engineering-incident-response",
          "tokens": 645
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "deep-review",
          "tokens": 1923
        },
        {
          "skill_id": "codebase-archaeologist",
          "tokens": 1977
        },
        {
          "skill_id": "ai-quality-evaluator",
          "tokens": 2228
        },
        {
          "skill_id": "skill-optimizer",
          "tokens": 2940
        },
        {
          "skill_id": "critical-review",
          "tokens": 3039
        },
        {
          "skill_id": "intent-alignment-tracker",
          "tokens": 558
        },
        {
          "skill_id": "incident-lifecycle-manager",
          "tokens": 610
        },
        {
          "skill_id": "email-research-dispatcher",
          "tokens": 703
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Incident to Improvement. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"incident-to-improvement\", \"ui-suite\", \"test-suite\", \"pr-to-issue-linker\", \"role-developer\", \"diagnose\", \"release-commander\", \"e2e-overhaul\", \"problem-definition\", \"bespin-news-digest\", \"issue-to-code\", \"infra-runbook-generator\", \"hf-trending-intelligence\", \"deep-research-pipeline\", \"onboarding-accelerator\", \"kwp-engineering-incident-response\", \"github-sprint-digest\", \"deep-review\", \"codebase-archaeologist\", \"ai-quality-evaluator\", \"skill-optimizer\", \"critical-review\", \"intent-alignment-tracker\", \"incident-lifecycle-manager\", \"email-research-dispatcher\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0135",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Brand Voice Enforcement. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-brand-voice-brand-voice-enforcement",
        "kwp-brand-voice-discover-brand",
        "kwp-brand-voice-guideline-generation",
        "kwp-design-ux-writing",
        "notebooklm-studio",
        "bespin-news-digest",
        "kwp-engineering-documentation",
        "gws-recipe-save-email-to-doc",
        "pm-go-to-market",
        "agency-tiktok-strategist",
        "gws-email-reply",
        "kwp-marketing-brand-voice",
        "slack-agent",
        "notion-meeting-sync",
        "kwp-engineering-testing-strategy",
        "kwp-engineering-tech-debt",
        "gws-docs",
        "daily-strategy-post",
        "gws-recipe-doc-from-template",
        "gws-chat",
        "anthropic-brand-guidelines",
        "pm-toolkit",
        "agency-content-creator",
        "gws-calendar",
        "office-template-enforcer"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-brand-voice-brand-voice-enforcement",
          "tokens": 1634
        },
        {
          "skill_id": "kwp-brand-voice-discover-brand",
          "tokens": 1897
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "kwp-design-ux-writing",
          "tokens": 675
        },
        {
          "skill_id": "notebooklm-studio",
          "tokens": 1812
        },
        {
          "skill_id": "bespin-news-digest",
          "tokens": 3949
        },
        {
          "skill_id": "kwp-engineering-documentation",
          "tokens": 654
        },
        {
          "skill_id": "gws-recipe-save-email-to-doc",
          "tokens": 676
        },
        {
          "skill_id": "pm-go-to-market",
          "tokens": 992
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "gws-email-reply",
          "tokens": 2222
        },
        {
          "skill_id": "kwp-marketing-brand-voice",
          "tokens": 2392
        },
        {
          "skill_id": "slack-agent",
          "tokens": 2564
        },
        {
          "skill_id": "notion-meeting-sync",
          "tokens": 4347
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "kwp-engineering-tech-debt",
          "tokens": 648
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "gws-recipe-doc-from-template",
          "tokens": 743
        },
        {
          "skill_id": "gws-chat",
          "tokens": 751
        },
        {
          "skill_id": "anthropic-brand-guidelines",
          "tokens": 781
        },
        {
          "skill_id": "pm-toolkit",
          "tokens": 935
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "office-template-enforcer",
          "tokens": 1273
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Brand Voice Enforcement. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-brand-voice-brand-voice-enforcement\", \"kwp-brand-voice-discover-brand\", \"kwp-brand-voice-guideline-generation\", \"kwp-design-ux-writing\", \"notebooklm-studio\", \"bespin-news-digest\", \"kwp-engineering-documentation\", \"gws-recipe-save-email-to-doc\", \"pm-go-to-market\", \"agency-tiktok-strategist\", \"gws-email-reply\", \"kwp-marketing-brand-voice\", \"slack-agent\", \"notion-meeting-sync\", \"kwp-engineering-testing-strategy\", \"kwp-engineering-tech-debt\", \"gws-docs\", \"daily-strategy-post\", \"gws-recipe-doc-from-template\", \"gws-chat\", \"anthropic-brand-guidelines\", \"pm-toolkit\", \"agency-content-creator\", \"gws-calendar\", \"office-template-enforcer\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0136",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Sales Data Extraction Agent. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-sales-data-extraction-agent",
        "agency-data-consolidation-agent",
        "agency-growth-hacker",
        "agency-ai-engineer",
        "agency-rapid-prototyper",
        "agency-experiment-tracker",
        "agency-mobile-app-builder",
        "agency-legal-compliance-checker",
        "agency-test-results-analyzer",
        "agency-data-engineer",
        "agency-xr-interface-architect",
        "agency-xr-cockpit-interaction-specialist",
        "agency-xr-immersive-developer",
        "agency-report-distribution-agent",
        "agency-content-creator",
        "agency-visionos-spatial-engineer",
        "agency-terminal-integration-specialist",
        "agency-behavioral-nudge-engine",
        "agency-senior-project-manager",
        "agency-instagram-curator",
        "agency-cultural-intelligence-strategist",
        "agency-inclusive-visuals-specialist",
        "agency-feedback-synthesizer",
        "agency-visual-storyteller",
        "agency-social-media-strategist"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-sales-data-extraction-agent",
          "tokens": 922
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-rapid-prototyper",
          "tokens": 2496
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-mobile-app-builder",
          "tokens": 2595
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "agency-data-engineer",
          "tokens": 3843
        },
        {
          "skill_id": "agency-xr-interface-architect",
          "tokens": 657
        },
        {
          "skill_id": "agency-xr-cockpit-interaction-specialist",
          "tokens": 693
        },
        {
          "skill_id": "agency-xr-immersive-developer",
          "tokens": 697
        },
        {
          "skill_id": "agency-report-distribution-agent",
          "tokens": 907
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "agency-visionos-spatial-engineer",
          "tokens": 1091
        },
        {
          "skill_id": "agency-terminal-integration-specialist",
          "tokens": 1273
        },
        {
          "skill_id": "agency-behavioral-nudge-engine",
          "tokens": 1553
        },
        {
          "skill_id": "agency-senior-project-manager",
          "tokens": 1555
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-cultural-intelligence-strategist",
          "tokens": 1892
        },
        {
          "skill_id": "agency-inclusive-visuals-specialist",
          "tokens": 1931
        },
        {
          "skill_id": "agency-feedback-synthesizer",
          "tokens": 2041
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-social-media-strategist",
          "tokens": 2084
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Sales Data Extraction Agent. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-sales-data-extraction-agent\", \"agency-data-consolidation-agent\", \"agency-growth-hacker\", \"agency-ai-engineer\", \"agency-rapid-prototyper\", \"agency-experiment-tracker\", \"agency-mobile-app-builder\", \"agency-legal-compliance-checker\", \"agency-test-results-analyzer\", \"agency-data-engineer\", \"agency-xr-interface-architect\", \"agency-xr-cockpit-interaction-specialist\", \"agency-xr-immersive-developer\", \"agency-report-distribution-agent\", \"agency-content-creator\", \"agency-visionos-spatial-engineer\", \"agency-terminal-integration-specialist\", \"agency-behavioral-nudge-engine\", \"agency-senior-project-manager\", \"agency-instagram-curator\", \"agency-cultural-intelligence-strategist\", \"agency-inclusive-visuals-specialist\", \"agency-feedback-synthesizer\", \"agency-visual-storyteller\", \"agency-social-media-strategist\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0137",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Create an Asset. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-sales-create-an-asset",
        "kwp-sales-call-prep",
        "kwp-sales-account-research",
        "kwp-apollo-prospect",
        "kwp-sales-daily-briefing",
        "kwp-sales-draft-outreach",
        "kwp-sales-competitive-intelligence",
        "kwp-human-resources-people-analytics",
        "kwp-operations-process-optimization",
        "kwp-apollo-enrich-lead",
        "kwp-finance-journal-entry-prep",
        "kwp-customer-support-escalation",
        "kwp-enterprise-search-knowledge-synthesis",
        "kwp-legal-meeting-briefing",
        "kwp-marketing-content-creation",
        "kwp-finance-financial-statements",
        "kwp-marketing-campaign-planning",
        "kwp-product-management-stakeholder-comms",
        "kwp-legal-nda-triage",
        "kwp-legal-legal-risk-assessment",
        "kwp-product-management-metrics-tracking",
        "kwp-operations-vendor-management",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-operations-change-management"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-sales-create-an-asset",
          "tokens": 4079
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "kwp-apollo-prospect",
          "tokens": 1006
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "kwp-sales-draft-outreach",
          "tokens": 2782
        },
        {
          "skill_id": "kwp-sales-competitive-intelligence",
          "tokens": 3361
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "kwp-apollo-enrich-lead",
          "tokens": 837
        },
        {
          "skill_id": "kwp-finance-journal-entry-prep",
          "tokens": 2215
        },
        {
          "skill_id": "kwp-customer-support-escalation",
          "tokens": 2329
        },
        {
          "skill_id": "kwp-enterprise-search-knowledge-synthesis",
          "tokens": 2363
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "kwp-finance-financial-statements",
          "tokens": 2768
        },
        {
          "skill_id": "kwp-marketing-campaign-planning",
          "tokens": 2803
        },
        {
          "skill_id": "kwp-product-management-stakeholder-comms",
          "tokens": 2952
        },
        {
          "skill_id": "kwp-legal-nda-triage",
          "tokens": 3139
        },
        {
          "skill_id": "kwp-legal-legal-risk-assessment",
          "tokens": 3406
        },
        {
          "skill_id": "kwp-product-management-metrics-tracking",
          "tokens": 3606
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-recruiting-pipeline",
          "tokens": 578
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Create an Asset. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-sales-create-an-asset\", \"kwp-sales-call-prep\", \"kwp-sales-account-research\", \"kwp-apollo-prospect\", \"kwp-sales-daily-briefing\", \"kwp-sales-draft-outreach\", \"kwp-sales-competitive-intelligence\", \"kwp-human-resources-people-analytics\", \"kwp-operations-process-optimization\", \"kwp-apollo-enrich-lead\", \"kwp-finance-journal-entry-prep\", \"kwp-customer-support-escalation\", \"kwp-enterprise-search-knowledge-synthesis\", \"kwp-legal-meeting-briefing\", \"kwp-marketing-content-creation\", \"kwp-finance-financial-statements\", \"kwp-marketing-campaign-planning\", \"kwp-product-management-stakeholder-comms\", \"kwp-legal-nda-triage\", \"kwp-legal-legal-risk-assessment\", \"kwp-product-management-metrics-tracking\", \"kwp-operations-vendor-management\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-operations-change-management\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0138",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Create an Asset. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-sales-create-an-asset",
        "kwp-sales-call-prep",
        "kwp-sales-account-research",
        "kwp-apollo-prospect",
        "kwp-sales-daily-briefing",
        "kwp-sales-draft-outreach",
        "kwp-sales-competitive-intelligence",
        "kwp-human-resources-people-analytics",
        "kwp-operations-process-optimization",
        "kwp-apollo-enrich-lead",
        "kwp-finance-journal-entry-prep",
        "kwp-customer-support-escalation",
        "kwp-enterprise-search-knowledge-synthesis",
        "kwp-legal-meeting-briefing",
        "kwp-marketing-content-creation",
        "kwp-finance-financial-statements",
        "kwp-marketing-campaign-planning",
        "kwp-product-management-stakeholder-comms",
        "kwp-legal-nda-triage",
        "kwp-legal-legal-risk-assessment",
        "kwp-product-management-metrics-tracking",
        "kwp-operations-vendor-management",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-operations-change-management"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-sales-create-an-asset",
          "tokens": 4079
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "kwp-apollo-prospect",
          "tokens": 1006
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "kwp-sales-draft-outreach",
          "tokens": 2782
        },
        {
          "skill_id": "kwp-sales-competitive-intelligence",
          "tokens": 3361
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "kwp-apollo-enrich-lead",
          "tokens": 837
        },
        {
          "skill_id": "kwp-finance-journal-entry-prep",
          "tokens": 2215
        },
        {
          "skill_id": "kwp-customer-support-escalation",
          "tokens": 2329
        },
        {
          "skill_id": "kwp-enterprise-search-knowledge-synthesis",
          "tokens": 2363
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "kwp-finance-financial-statements",
          "tokens": 2768
        },
        {
          "skill_id": "kwp-marketing-campaign-planning",
          "tokens": 2803
        },
        {
          "skill_id": "kwp-product-management-stakeholder-comms",
          "tokens": 2952
        },
        {
          "skill_id": "kwp-legal-nda-triage",
          "tokens": 3139
        },
        {
          "skill_id": "kwp-legal-legal-risk-assessment",
          "tokens": 3406
        },
        {
          "skill_id": "kwp-product-management-metrics-tracking",
          "tokens": 3606
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-recruiting-pipeline",
          "tokens": 578
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Create an Asset. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-sales-create-an-asset\", \"kwp-sales-call-prep\", \"kwp-sales-account-research\", \"kwp-apollo-prospect\", \"kwp-sales-daily-briefing\", \"kwp-sales-draft-outreach\", \"kwp-sales-competitive-intelligence\", \"kwp-human-resources-people-analytics\", \"kwp-operations-process-optimization\", \"kwp-apollo-enrich-lead\", \"kwp-finance-journal-entry-prep\", \"kwp-customer-support-escalation\", \"kwp-enterprise-search-knowledge-synthesis\", \"kwp-legal-meeting-briefing\", \"kwp-marketing-content-creation\", \"kwp-finance-financial-statements\", \"kwp-marketing-campaign-planning\", \"kwp-product-management-stakeholder-comms\", \"kwp-legal-nda-triage\", \"kwp-legal-legal-risk-assessment\", \"kwp-product-management-metrics-tracking\", \"kwp-operations-vendor-management\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-operations-change-management\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0139",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AI Engineer Agent. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-ai-engineer",
        "agency-data-engineer",
        "agency-growth-hacker",
        "agency-instagram-curator",
        "agency-visual-storyteller",
        "agency-tiktok-strategist",
        "agency-experiment-tracker",
        "agency-studio-operations",
        "agency-infrastructure-maintainer",
        "agency-developer-advocate",
        "agency-xr-immersive-developer",
        "agency-data-consolidation-agent",
        "agency-reddit-community-builder",
        "agency-twitter-engager",
        "agency-project-shepherd",
        "agency-trend-researcher",
        "agency-studio-producer",
        "agency-lsp-index-engineer",
        "agency-image-prompt-engineer",
        "agency-brand-guardian",
        "agency-macos-spatial-metal-engineer",
        "agency-api-tester",
        "agency-legal-compliance-checker",
        "agency-terminal-integration-specialist",
        "agency-feedback-synthesizer"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-data-engineer",
          "tokens": 3843
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "agency-xr-immersive-developer",
          "tokens": 697
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "agency-studio-producer",
          "tokens": 2953
        },
        {
          "skill_id": "agency-lsp-index-engineer",
          "tokens": 3006
        },
        {
          "skill_id": "agency-image-prompt-engineer",
          "tokens": 3027
        },
        {
          "skill_id": "agency-brand-guardian",
          "tokens": 3135
        },
        {
          "skill_id": "agency-macos-spatial-metal-engineer",
          "tokens": 3229
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-terminal-integration-specialist",
          "tokens": 1273
        },
        {
          "skill_id": "agency-feedback-synthesizer",
          "tokens": 2041
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AI Engineer Agent. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-ai-engineer\", \"agency-data-engineer\", \"agency-growth-hacker\", \"agency-instagram-curator\", \"agency-visual-storyteller\", \"agency-tiktok-strategist\", \"agency-experiment-tracker\", \"agency-studio-operations\", \"agency-infrastructure-maintainer\", \"agency-developer-advocate\", \"agency-xr-immersive-developer\", \"agency-data-consolidation-agent\", \"agency-reddit-community-builder\", \"agency-twitter-engager\", \"agency-project-shepherd\", \"agency-trend-researcher\", \"agency-studio-producer\", \"agency-lsp-index-engineer\", \"agency-image-prompt-engineer\", \"agency-brand-guardian\", \"agency-macos-spatial-metal-engineer\", \"agency-api-tester\", \"agency-legal-compliance-checker\", \"agency-terminal-integration-specialist\", \"agency-feedback-synthesizer\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0140",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: daily-strategy-post. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "daily-strategy-post",
        "role-ceo",
        "morning-ship",
        "email-research-dispatcher",
        "knowledge-to-paper",
        "knowledge-daily-aggregator",
        "daily-am-orchestrator",
        "ai-chief-of-staff",
        "nlm-strategy-hub",
        "deep-research-pipeline",
        "nlm-discovery-lab",
        "bespin-news-digest",
        "standup-digest",
        "weekly-stock-update",
        "autoskill-pipeline",
        "unified-intel-intake",
        "hf-trending-intelligence",
        "agency-feedback-synthesizer",
        "morning",
        "ai-workflow-integrator",
        "eod-ship",
        "agency-trend-researcher",
        "agency-data-analytics-reporter",
        "agency-content-creator",
        "weekly-status-report"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "morning-ship",
          "tokens": 2943
        },
        {
          "skill_id": "email-research-dispatcher",
          "tokens": 703
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "knowledge-daily-aggregator",
          "tokens": 1425
        },
        {
          "skill_id": "daily-am-orchestrator",
          "tokens": 3613
        },
        {
          "skill_id": "ai-chief-of-staff",
          "tokens": 1152
        },
        {
          "skill_id": "nlm-strategy-hub",
          "tokens": 2207
        },
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        },
        {
          "skill_id": "nlm-discovery-lab",
          "tokens": 2808
        },
        {
          "skill_id": "bespin-news-digest",
          "tokens": 3949
        },
        {
          "skill_id": "standup-digest",
          "tokens": 1330
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "autoskill-pipeline",
          "tokens": 1508
        },
        {
          "skill_id": "unified-intel-intake",
          "tokens": 1593
        },
        {
          "skill_id": "hf-trending-intelligence",
          "tokens": 1832
        },
        {
          "skill_id": "agency-feedback-synthesizer",
          "tokens": 2041
        },
        {
          "skill_id": "morning",
          "tokens": 2088
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "eod-ship",
          "tokens": 2420
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "agency-data-analytics-reporter",
          "tokens": 1003
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "weekly-status-report",
          "tokens": 1282
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: daily-strategy-post. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"daily-strategy-post\", \"role-ceo\", \"morning-ship\", \"email-research-dispatcher\", \"knowledge-to-paper\", \"knowledge-daily-aggregator\", \"daily-am-orchestrator\", \"ai-chief-of-staff\", \"nlm-strategy-hub\", \"deep-research-pipeline\", \"nlm-discovery-lab\", \"bespin-news-digest\", \"standup-digest\", \"weekly-stock-update\", \"autoskill-pipeline\", \"unified-intel-intake\", \"hf-trending-intelligence\", \"agency-feedback-synthesizer\", \"morning\", \"ai-workflow-integrator\", \"eod-ship\", \"agency-trend-researcher\", \"agency-data-analytics-reporter\", \"agency-content-creator\", \"weekly-status-report\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0141",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-market-breadth. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "tab-market-breadth",
        "tab-market-environment",
        "tab-turtle-refresh",
        "tab-bollinger-refresh",
        "tab-dualma-refresh",
        "tab-news-fetch",
        "tab-llm-agents",
        "tab-dashboard-summary",
        "tab-patterns",
        "tab-glossary",
        "tab-strategy-comparison",
        "tab-report-generate",
        "tab-technical-analysis",
        "daily-stock-check",
        "tab-fundamental-sync",
        "tab-stock-sync",
        "weekly-stock-update",
        "tab-hot-stock-discovery",
        "tab-kiwoom",
        "alphaear-news",
        "stock-csv-downloader",
        "trading-market-environment-analysis",
        "trading-market-breadth-analyzer",
        "trading-sector-analyst",
        "trading-technical-analyst"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "tab-market-breadth",
          "tokens": 504
        },
        {
          "skill_id": "tab-market-environment",
          "tokens": 566
        },
        {
          "skill_id": "tab-turtle-refresh",
          "tokens": 371
        },
        {
          "skill_id": "tab-bollinger-refresh",
          "tokens": 383
        },
        {
          "skill_id": "tab-dualma-refresh",
          "tokens": 391
        },
        {
          "skill_id": "tab-news-fetch",
          "tokens": 474
        },
        {
          "skill_id": "tab-llm-agents",
          "tokens": 382
        },
        {
          "skill_id": "tab-dashboard-summary",
          "tokens": 419
        },
        {
          "skill_id": "tab-patterns",
          "tokens": 509
        },
        {
          "skill_id": "tab-glossary",
          "tokens": 538
        },
        {
          "skill_id": "tab-strategy-comparison",
          "tokens": 571
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "tab-technical-analysis",
          "tokens": 414
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "tab-fundamental-sync",
          "tokens": 331
        },
        {
          "skill_id": "tab-stock-sync",
          "tokens": 345
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "tab-hot-stock-discovery",
          "tokens": 359
        },
        {
          "skill_id": "tab-kiwoom",
          "tokens": 452
        },
        {
          "skill_id": "alphaear-news",
          "tokens": 873
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "trading-market-breadth-analyzer",
          "tokens": 1830
        },
        {
          "skill_id": "trading-sector-analyst",
          "tokens": 2657
        },
        {
          "skill_id": "trading-technical-analyst",
          "tokens": 2868
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-market-breadth. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"tab-market-breadth\", \"tab-market-environment\", \"tab-turtle-refresh\", \"tab-bollinger-refresh\", \"tab-dualma-refresh\", \"tab-news-fetch\", \"tab-llm-agents\", \"tab-dashboard-summary\", \"tab-patterns\", \"tab-glossary\", \"tab-strategy-comparison\", \"tab-report-generate\", \"tab-technical-analysis\", \"daily-stock-check\", \"tab-fundamental-sync\", \"tab-stock-sync\", \"weekly-stock-update\", \"tab-hot-stock-discovery\", \"tab-kiwoom\", \"alphaear-news\", \"stock-csv-downloader\", \"trading-market-environment-analysis\", \"trading-market-breadth-analyzer\", \"trading-sector-analyst\", \"trading-technical-analyst\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0142",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Compliance Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-legal-compliance",
        "kwp-legal-canned-responses",
        "kwp-legal-contract-review",
        "kwp-legal-legal-risk-assessment",
        "kwp-legal-nda-triage",
        "gws-recipe-audit-sharing",
        "compliance-governance",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-data-data-exploration",
        "kwp-marketing-brand-voice",
        "kwp-legal-meeting-briefing",
        "kwp-data-data-validation",
        "kwp-customer-support-response-drafting",
        "agency-legal-compliance-checker",
        "kwp-human-resources-people-analytics",
        "kwp-operations-vendor-management",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-operations-compliance-tracking",
        "kwp-finance-reconciliation",
        "kwp-customer-support-escalation",
        "kwp-data-interactive-dashboard-builder",
        "mission-control",
        "kwp-product-management-feature-spec",
        "kwp-data-statistical-analysis",
        "kwp-product-management-competitive-analysis"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-legal-compliance",
          "tokens": 3473
        },
        {
          "skill_id": "kwp-legal-canned-responses",
          "tokens": 3164
        },
        {
          "skill_id": "kwp-legal-contract-review",
          "tokens": 3137
        },
        {
          "skill_id": "kwp-legal-legal-risk-assessment",
          "tokens": 3406
        },
        {
          "skill_id": "kwp-legal-nda-triage",
          "tokens": 3139
        },
        {
          "skill_id": "gws-recipe-audit-sharing",
          "tokens": 645
        },
        {
          "skill_id": "compliance-governance",
          "tokens": 1501
        },
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "kwp-data-data-exploration",
          "tokens": 2358
        },
        {
          "skill_id": "kwp-marketing-brand-voice",
          "tokens": 2392
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-data-data-validation",
          "tokens": 2650
        },
        {
          "skill_id": "kwp-customer-support-response-drafting",
          "tokens": 2944
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-operations-compliance-tracking",
          "tokens": 686
        },
        {
          "skill_id": "kwp-finance-reconciliation",
          "tokens": 2287
        },
        {
          "skill_id": "kwp-customer-support-escalation",
          "tokens": 2329
        },
        {
          "skill_id": "kwp-data-interactive-dashboard-builder",
          "tokens": 2350
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "kwp-product-management-feature-spec",
          "tokens": 2614
        },
        {
          "skill_id": "kwp-data-statistical-analysis",
          "tokens": 2865
        },
        {
          "skill_id": "kwp-product-management-competitive-analysis",
          "tokens": 2899
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Compliance Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-legal-compliance\", \"kwp-legal-canned-responses\", \"kwp-legal-contract-review\", \"kwp-legal-legal-risk-assessment\", \"kwp-legal-nda-triage\", \"gws-recipe-audit-sharing\", \"compliance-governance\", \"kwp-bio-research-single-cell-rna-qc\", \"kwp-data-data-exploration\", \"kwp-marketing-brand-voice\", \"kwp-legal-meeting-briefing\", \"kwp-data-data-validation\", \"kwp-customer-support-response-drafting\", \"agency-legal-compliance-checker\", \"kwp-human-resources-people-analytics\", \"kwp-operations-vendor-management\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-operations-compliance-tracking\", \"kwp-finance-reconciliation\", \"kwp-customer-support-escalation\", \"kwp-data-interactive-dashboard-builder\", \"mission-control\", \"kwp-product-management-feature-spec\", \"kwp-data-statistical-analysis\", \"kwp-product-management-competitive-analysis\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0143",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Domain-Split Commit. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "domain-commit",
        "release-ship",
        "commit-to-issue",
        "ship",
        "sod-ship",
        "kwp-productivity-task-management",
        "simplify",
        "hf-discussions",
        "pr-to-issue-linker",
        "pr-review-captain",
        "ui-suite",
        "deep-review",
        "cursor-sync",
        "gws-docs",
        "hf-buckets",
        "diagnose",
        "test-suite",
        "release-commander",
        "kwp-operations-change-management",
        "gws-recipe-personalized-emails",
        "gws-recipe-events-from-sheet",
        "gws-recipe-doc-from-template",
        "gws-sheets",
        "gws-workflows",
        "hf-repos"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "domain-commit",
          "tokens": 915
        },
        {
          "skill_id": "release-ship",
          "tokens": 3056
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "sod-ship",
          "tokens": 4151
        },
        {
          "skill_id": "kwp-productivity-task-management",
          "tokens": 776
        },
        {
          "skill_id": "simplify",
          "tokens": 2596
        },
        {
          "skill_id": "hf-discussions",
          "tokens": 1118
        },
        {
          "skill_id": "pr-to-issue-linker",
          "tokens": 1223
        },
        {
          "skill_id": "pr-review-captain",
          "tokens": 1496
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "deep-review",
          "tokens": 1923
        },
        {
          "skill_id": "cursor-sync",
          "tokens": 3267
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "hf-buckets",
          "tokens": 1334
        },
        {
          "skill_id": "diagnose",
          "tokens": 1433
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "gws-recipe-personalized-emails",
          "tokens": 641
        },
        {
          "skill_id": "gws-recipe-events-from-sheet",
          "tokens": 674
        },
        {
          "skill_id": "gws-recipe-doc-from-template",
          "tokens": 743
        },
        {
          "skill_id": "gws-sheets",
          "tokens": 918
        },
        {
          "skill_id": "gws-workflows",
          "tokens": 952
        },
        {
          "skill_id": "hf-repos",
          "tokens": 1106
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Domain-Split Commit. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"domain-commit\", \"release-ship\", \"commit-to-issue\", \"ship\", \"sod-ship\", \"kwp-productivity-task-management\", \"simplify\", \"hf-discussions\", \"pr-to-issue-linker\", \"pr-review-captain\", \"ui-suite\", \"deep-review\", \"cursor-sync\", \"gws-docs\", \"hf-buckets\", \"diagnose\", \"test-suite\", \"release-commander\", \"kwp-operations-change-management\", \"gws-recipe-personalized-emails\", \"gws-recipe-events-from-sheet\", \"gws-recipe-doc-from-template\", \"gws-sheets\", \"gws-workflows\", \"hf-repos\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0144",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: md-to-notion. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "md-to-notion",
        "md-to-slack-canvas",
        "md-enhance-publish",
        "meeting-digest",
        "gws-docs",
        "demo-forge",
        "anthropic-pdf",
        "visual-explainer",
        "hf-buckets",
        "issue-to-code",
        "nlm-video",
        "release-ship",
        "github-sprint-digest",
        "knowledge-to-paper",
        "transcribee",
        "knowledge-daily-aggregator",
        "nlm-slides",
        "mission-control copy",
        "ralph-loop",
        "kwp-brand-voice-guideline-generation",
        "docs-tutor-setup",
        "mission-control",
        "nlm-discovery-lab",
        "cursor-sync",
        "paper-review"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "md-to-notion",
          "tokens": 2096
        },
        {
          "skill_id": "md-to-slack-canvas",
          "tokens": 1960
        },
        {
          "skill_id": "md-enhance-publish",
          "tokens": 1740
        },
        {
          "skill_id": "meeting-digest",
          "tokens": 4491
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "demo-forge",
          "tokens": 2152
        },
        {
          "skill_id": "anthropic-pdf",
          "tokens": 2219
        },
        {
          "skill_id": "visual-explainer",
          "tokens": 5408
        },
        {
          "skill_id": "hf-buckets",
          "tokens": 1334
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "nlm-video",
          "tokens": 1763
        },
        {
          "skill_id": "release-ship",
          "tokens": 3056
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "transcribee",
          "tokens": 1255
        },
        {
          "skill_id": "knowledge-daily-aggregator",
          "tokens": 1425
        },
        {
          "skill_id": "nlm-slides",
          "tokens": 1685
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "ralph-loop",
          "tokens": 1925
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "docs-tutor-setup",
          "tokens": 2288
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "nlm-discovery-lab",
          "tokens": 2808
        },
        {
          "skill_id": "cursor-sync",
          "tokens": 3267
        },
        {
          "skill_id": "paper-review",
          "tokens": 5815
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: md-to-notion. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"md-to-notion\", \"md-to-slack-canvas\", \"md-enhance-publish\", \"meeting-digest\", \"gws-docs\", \"demo-forge\", \"anthropic-pdf\", \"visual-explainer\", \"hf-buckets\", \"issue-to-code\", \"nlm-video\", \"release-ship\", \"github-sprint-digest\", \"knowledge-to-paper\", \"transcribee\", \"knowledge-daily-aggregator\", \"nlm-slides\", \"mission-control copy\", \"ralph-loop\", \"kwp-brand-voice-guideline-generation\", \"docs-tutor-setup\", \"mission-control\", \"nlm-discovery-lab\", \"cursor-sync\", \"paper-review\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0145",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: MiroFish Financial Scenario Simulation. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "mirofish-financial-sim",
        "mirofish-opinion-sim",
        "mirofish",
        "cursor-automations",
        "alphaear-deepear-lite",
        "trading-market-news-analyst",
        "trading-sector-analyst",
        "trading-theme-detector",
        "alphaear-predictor",
        "alphaear-sentiment",
        "mirofish-graph-explorer",
        "terraform-reviewer",
        "weekly-stock-update",
        "daily-strategy-post",
        "alphaear-news",
        "helm-validator",
        "ui-suite",
        "trading-market-environment-analysis",
        "daily-stock-check",
        "trading-market-breadth-analyzer",
        "hf-trending-intelligence",
        "codebase-archaeologist",
        "trading-position-sizer",
        "refactor-simulator",
        "agency-twitter-engager"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "mirofish-financial-sim",
          "tokens": 1330
        },
        {
          "skill_id": "mirofish-opinion-sim",
          "tokens": 1383
        },
        {
          "skill_id": "mirofish",
          "tokens": 3657
        },
        {
          "skill_id": "cursor-automations",
          "tokens": 2911
        },
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        },
        {
          "skill_id": "trading-market-news-analyst",
          "tokens": 1467
        },
        {
          "skill_id": "trading-sector-analyst",
          "tokens": 2657
        },
        {
          "skill_id": "trading-theme-detector",
          "tokens": 3487
        },
        {
          "skill_id": "alphaear-predictor",
          "tokens": 803
        },
        {
          "skill_id": "alphaear-sentiment",
          "tokens": 928
        },
        {
          "skill_id": "mirofish-graph-explorer",
          "tokens": 1175
        },
        {
          "skill_id": "terraform-reviewer",
          "tokens": 1457
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "alphaear-news",
          "tokens": 873
        },
        {
          "skill_id": "helm-validator",
          "tokens": 1398
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "trading-market-breadth-analyzer",
          "tokens": 1830
        },
        {
          "skill_id": "hf-trending-intelligence",
          "tokens": 1832
        },
        {
          "skill_id": "codebase-archaeologist",
          "tokens": 1977
        },
        {
          "skill_id": "trading-position-sizer",
          "tokens": 2013
        },
        {
          "skill_id": "refactor-simulator",
          "tokens": 2044
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: MiroFish Financial Scenario Simulation. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"mirofish-financial-sim\", \"mirofish-opinion-sim\", \"mirofish\", \"cursor-automations\", \"alphaear-deepear-lite\", \"trading-market-news-analyst\", \"trading-sector-analyst\", \"trading-theme-detector\", \"alphaear-predictor\", \"alphaear-sentiment\", \"mirofish-graph-explorer\", \"terraform-reviewer\", \"weekly-stock-update\", \"daily-strategy-post\", \"alphaear-news\", \"helm-validator\", \"ui-suite\", \"trading-market-environment-analysis\", \"daily-stock-check\", \"trading-market-breadth-analyzer\", \"hf-trending-intelligence\", \"codebase-archaeologist\", \"trading-position-sizer\", \"refactor-simulator\", \"agency-twitter-engager\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0146",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Brand Voice Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-marketing-brand-voice",
        "kwp-brand-voice-discover-brand",
        "kwp-brand-voice-guideline-generation",
        "kwp-marketing-content-creation",
        "kwp-marketing-competitive-analysis",
        "kwp-marketing-campaign-planning",
        "kwp-data-data-validation",
        "kwp-legal-compliance",
        "kwp-marketing-performance-analytics",
        "kwp-operations-resource-planning",
        "kwp-design-design-system-management",
        "kwp-brand-voice-brand-voice-enforcement",
        "kwp-legal-meeting-briefing",
        "kwp-customer-support-customer-research",
        "kwp-product-management-feature-spec",
        "kwp-data-statistical-analysis",
        "kwp-legal-contract-review",
        "kwp-finance-audit-support",
        "kwp-human-resources-people-analytics",
        "kwp-operations-vendor-management",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-human-resources-interview-prep"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-marketing-brand-voice",
          "tokens": 2392
        },
        {
          "skill_id": "kwp-brand-voice-discover-brand",
          "tokens": 1897
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "kwp-marketing-competitive-analysis",
          "tokens": 2723
        },
        {
          "skill_id": "kwp-marketing-campaign-planning",
          "tokens": 2803
        },
        {
          "skill_id": "kwp-data-data-validation",
          "tokens": 2650
        },
        {
          "skill_id": "kwp-legal-compliance",
          "tokens": 3473
        },
        {
          "skill_id": "kwp-marketing-performance-analytics",
          "tokens": 3966
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "kwp-brand-voice-brand-voice-enforcement",
          "tokens": 1634
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-customer-support-customer-research",
          "tokens": 2529
        },
        {
          "skill_id": "kwp-product-management-feature-spec",
          "tokens": 2614
        },
        {
          "skill_id": "kwp-data-statistical-analysis",
          "tokens": 2865
        },
        {
          "skill_id": "kwp-legal-contract-review",
          "tokens": 3137
        },
        {
          "skill_id": "kwp-finance-audit-support",
          "tokens": 4426
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-recruiting-pipeline",
          "tokens": 578
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Brand Voice Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-marketing-brand-voice\", \"kwp-brand-voice-discover-brand\", \"kwp-brand-voice-guideline-generation\", \"kwp-marketing-content-creation\", \"kwp-marketing-competitive-analysis\", \"kwp-marketing-campaign-planning\", \"kwp-data-data-validation\", \"kwp-legal-compliance\", \"kwp-marketing-performance-analytics\", \"kwp-operations-resource-planning\", \"kwp-design-design-system-management\", \"kwp-brand-voice-brand-voice-enforcement\", \"kwp-legal-meeting-briefing\", \"kwp-customer-support-customer-research\", \"kwp-product-management-feature-spec\", \"kwp-data-statistical-analysis\", \"kwp-legal-contract-review\", \"kwp-finance-audit-support\", \"kwp-human-resources-people-analytics\", \"kwp-operations-vendor-management\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-operations-process-optimization\", \"kwp-operations-change-management\", \"kwp-human-resources-interview-prep\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0147",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: SOD Ship. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "sod-ship",
        "morning-ship",
        "eod-ship",
        "morning",
        "cursor-sync",
        "domain-commit",
        "release-ship",
        "daily-am-orchestrator",
        "commit-to-issue",
        "ship",
        "jira-linear-adapter",
        "portfolio-report-generator",
        "release-commander",
        "daily-strategy-post",
        "knowledge-daily-aggregator",
        "trading-data-quality-checker",
        "bespin-news-digest",
        "github-sprint-digest",
        "knowledge-to-paper",
        "ai-chief-of-staff",
        "pr-to-issue-linker",
        "sprint-orchestrator",
        "hf-buckets",
        "compliance-gate",
        "proactive-meeting-scheduler"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "sod-ship",
          "tokens": 4151
        },
        {
          "skill_id": "morning-ship",
          "tokens": 2943
        },
        {
          "skill_id": "eod-ship",
          "tokens": 2420
        },
        {
          "skill_id": "morning",
          "tokens": 2088
        },
        {
          "skill_id": "cursor-sync",
          "tokens": 3267
        },
        {
          "skill_id": "domain-commit",
          "tokens": 915
        },
        {
          "skill_id": "release-ship",
          "tokens": 3056
        },
        {
          "skill_id": "daily-am-orchestrator",
          "tokens": 3613
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "jira-linear-adapter",
          "tokens": 1365
        },
        {
          "skill_id": "portfolio-report-generator",
          "tokens": 1406
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "knowledge-daily-aggregator",
          "tokens": 1425
        },
        {
          "skill_id": "trading-data-quality-checker",
          "tokens": 1962
        },
        {
          "skill_id": "bespin-news-digest",
          "tokens": 3949
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "ai-chief-of-staff",
          "tokens": 1152
        },
        {
          "skill_id": "pr-to-issue-linker",
          "tokens": 1223
        },
        {
          "skill_id": "sprint-orchestrator",
          "tokens": 1320
        },
        {
          "skill_id": "hf-buckets",
          "tokens": 1334
        },
        {
          "skill_id": "compliance-gate",
          "tokens": 1376
        },
        {
          "skill_id": "proactive-meeting-scheduler",
          "tokens": 1444
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: SOD Ship. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"sod-ship\", \"morning-ship\", \"eod-ship\", \"morning\", \"cursor-sync\", \"domain-commit\", \"release-ship\", \"daily-am-orchestrator\", \"commit-to-issue\", \"ship\", \"jira-linear-adapter\", \"portfolio-report-generator\", \"release-commander\", \"daily-strategy-post\", \"knowledge-daily-aggregator\", \"trading-data-quality-checker\", \"bespin-news-digest\", \"github-sprint-digest\", \"knowledge-to-paper\", \"ai-chief-of-staff\", \"pr-to-issue-linker\", \"sprint-orchestrator\", \"hf-buckets\", \"compliance-gate\", \"proactive-meeting-scheduler\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0148",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Systematic Debugging. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "sp-debugging",
        "ecc-verification-loop",
        "diagnose",
        "ui-suite",
        "ship",
        "deep-review",
        "code-review-all",
        "problem-definition",
        "sp-verification",
        "test-suite",
        "refactor-simulator",
        "docs-freshness-guardian",
        "role-developer",
        "sp-receive-review",
        "trading-data-quality-checker",
        "codebase-archaeologist",
        "ai-quality-evaluator",
        "release-commander",
        "agency-whimsy-injector",
        "sp-tdd",
        "simplify",
        "skill-optimizer",
        "agency-test-results-analyzer",
        "visual-explainer",
        "dependency-radar"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "sp-debugging",
          "tokens": 2574
        },
        {
          "skill_id": "ecc-verification-loop",
          "tokens": 886
        },
        {
          "skill_id": "diagnose",
          "tokens": 1433
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "deep-review",
          "tokens": 1923
        },
        {
          "skill_id": "code-review-all",
          "tokens": 2524
        },
        {
          "skill_id": "problem-definition",
          "tokens": 2601
        },
        {
          "skill_id": "sp-verification",
          "tokens": 1146
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "refactor-simulator",
          "tokens": 2044
        },
        {
          "skill_id": "docs-freshness-guardian",
          "tokens": 557
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "sp-receive-review",
          "tokens": 1761
        },
        {
          "skill_id": "trading-data-quality-checker",
          "tokens": 1962
        },
        {
          "skill_id": "codebase-archaeologist",
          "tokens": 1977
        },
        {
          "skill_id": "ai-quality-evaluator",
          "tokens": 2228
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "sp-tdd",
          "tokens": 2511
        },
        {
          "skill_id": "simplify",
          "tokens": 2596
        },
        {
          "skill_id": "skill-optimizer",
          "tokens": 2940
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "visual-explainer",
          "tokens": 5408
        },
        {
          "skill_id": "dependency-radar",
          "tokens": 531
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Systematic Debugging. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"sp-debugging\", \"ecc-verification-loop\", \"diagnose\", \"ui-suite\", \"ship\", \"deep-review\", \"code-review-all\", \"problem-definition\", \"sp-verification\", \"test-suite\", \"refactor-simulator\", \"docs-freshness-guardian\", \"role-developer\", \"sp-receive-review\", \"trading-data-quality-checker\", \"codebase-archaeologist\", \"ai-quality-evaluator\", \"release-commander\", \"agency-whimsy-injector\", \"sp-tdd\", \"simplify\", \"skill-optimizer\", \"agency-test-results-analyzer\", \"visual-explainer\", \"dependency-radar\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0149",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AutoSkill Judge. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "autoskill-judge",
        "autoskill-merger",
        "autoskill-extractor",
        "autoskill-evolve",
        "skill-optimizer",
        "workflow-miner",
        "skill-guide-generator",
        "skill-composer",
        "evals-skills",
        "prompt-transformer",
        "intent-alignment-tracker",
        "kwp-sync",
        "ui-suite",
        "release-commander",
        "kwp-cowork-create-cowork-plugin",
        "agency-roster",
        "anthropic-skill-creator",
        "agency-data-consolidation-agent",
        "commit-to-issue",
        "skill-seekers",
        "nlm-slides",
        "mission-control copy",
        "deep-review",
        "kwp-brand-voice-guideline-generation",
        "agency-feedback-synthesizer"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "autoskill-judge",
          "tokens": 1176
        },
        {
          "skill_id": "autoskill-merger",
          "tokens": 1142
        },
        {
          "skill_id": "autoskill-extractor",
          "tokens": 1235
        },
        {
          "skill_id": "autoskill-evolve",
          "tokens": 1566
        },
        {
          "skill_id": "skill-optimizer",
          "tokens": 2940
        },
        {
          "skill_id": "workflow-miner",
          "tokens": 893
        },
        {
          "skill_id": "skill-guide-generator",
          "tokens": 1858
        },
        {
          "skill_id": "skill-composer",
          "tokens": 599
        },
        {
          "skill_id": "evals-skills",
          "tokens": 1524
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "intent-alignment-tracker",
          "tokens": 558
        },
        {
          "skill_id": "kwp-sync",
          "tokens": 1472
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "kwp-cowork-create-cowork-plugin",
          "tokens": 3017
        },
        {
          "skill_id": "agency-roster",
          "tokens": 3411
        },
        {
          "skill_id": "anthropic-skill-creator",
          "tokens": 6862
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "skill-seekers",
          "tokens": 1568
        },
        {
          "skill_id": "nlm-slides",
          "tokens": 1685
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "deep-review",
          "tokens": 1923
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "agency-feedback-synthesizer",
          "tokens": 2041
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AutoSkill Judge. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"autoskill-judge\", \"autoskill-merger\", \"autoskill-extractor\", \"autoskill-evolve\", \"skill-optimizer\", \"workflow-miner\", \"skill-guide-generator\", \"skill-composer\", \"evals-skills\", \"prompt-transformer\", \"intent-alignment-tracker\", \"kwp-sync\", \"ui-suite\", \"release-commander\", \"kwp-cowork-create-cowork-plugin\", \"agency-roster\", \"anthropic-skill-creator\", \"agency-data-consolidation-agent\", \"commit-to-issue\", \"skill-seekers\", \"nlm-slides\", \"mission-control copy\", \"deep-review\", \"kwp-brand-voice-guideline-generation\", \"agency-feedback-synthesizer\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0150",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Image Prompt Engineer Agent. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-image-prompt-engineer",
        "agency-visual-storyteller",
        "agency-data-engineer",
        "agency-ai-engineer",
        "agency-wechat-official-account-manager",
        "agency-growth-hacker",
        "agency-instagram-curator",
        "agency-whimsy-injector",
        "agency-experiment-tracker",
        "agency-xiaohongshu-specialist",
        "agency-studio-operations",
        "agency-trend-researcher",
        "agency-lsp-index-engineer",
        "agency-zhihu-strategist",
        "agency-test-results-analyzer",
        "agency-developer-advocate",
        "agency-data-consolidation-agent",
        "agency-feedback-synthesizer",
        "agency-social-media-strategist",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "agency-twitter-engager",
        "agency-project-shepherd",
        "agency-brand-guardian",
        "agency-finance-tracker"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-image-prompt-engineer",
          "tokens": 3027
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-data-engineer",
          "tokens": 3843
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "agency-lsp-index-engineer",
          "tokens": 3006
        },
        {
          "skill_id": "agency-zhihu-strategist",
          "tokens": 3239
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        },
        {
          "skill_id": "agency-feedback-synthesizer",
          "tokens": 2041
        },
        {
          "skill_id": "agency-social-media-strategist",
          "tokens": 2084
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-brand-guardian",
          "tokens": 3135
        },
        {
          "skill_id": "agency-finance-tracker",
          "tokens": 3177
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Image Prompt Engineer Agent. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-image-prompt-engineer\", \"agency-visual-storyteller\", \"agency-data-engineer\", \"agency-ai-engineer\", \"agency-wechat-official-account-manager\", \"agency-growth-hacker\", \"agency-instagram-curator\", \"agency-whimsy-injector\", \"agency-experiment-tracker\", \"agency-xiaohongshu-specialist\", \"agency-studio-operations\", \"agency-trend-researcher\", \"agency-lsp-index-engineer\", \"agency-zhihu-strategist\", \"agency-test-results-analyzer\", \"agency-developer-advocate\", \"agency-data-consolidation-agent\", \"agency-feedback-synthesizer\", \"agency-social-media-strategist\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"agency-twitter-engager\", \"agency-project-shepherd\", \"agency-brand-guardian\", \"agency-finance-tracker\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0151",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PPTX Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "anthropic-pptx",
        "anthropic-pdf",
        "anthropic-docx",
        "ppt-template-engine",
        "anthropic-xlsx",
        "docx-template-engine",
        "office-template-enforcer",
        "skill-seekers",
        "nlm-slides",
        "notion-meeting-sync",
        "gws-docs",
        "alphaxiv-paper-lookup",
        "cognee",
        "docs-tutor-setup",
        "skill-optimizer",
        "anthropic-canvas-design",
        "nlm-arxiv-slides",
        "anthropic-doc-coauthoring",
        "anthropic-skill-creator",
        "gws-recipe-doc-from-template",
        "gws-sheets",
        "autoskill-merger",
        "video-compress",
        "transcribee",
        "hf-buckets"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "anthropic-pptx",
          "tokens": 1588
        },
        {
          "skill_id": "anthropic-pdf",
          "tokens": 2219
        },
        {
          "skill_id": "anthropic-docx",
          "tokens": 4265
        },
        {
          "skill_id": "ppt-template-engine",
          "tokens": 1101
        },
        {
          "skill_id": "anthropic-xlsx",
          "tokens": 2876
        },
        {
          "skill_id": "docx-template-engine",
          "tokens": 1193
        },
        {
          "skill_id": "office-template-enforcer",
          "tokens": 1273
        },
        {
          "skill_id": "skill-seekers",
          "tokens": 1568
        },
        {
          "skill_id": "nlm-slides",
          "tokens": 1685
        },
        {
          "skill_id": "notion-meeting-sync",
          "tokens": 4347
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "alphaxiv-paper-lookup",
          "tokens": 1299
        },
        {
          "skill_id": "cognee",
          "tokens": 2070
        },
        {
          "skill_id": "docs-tutor-setup",
          "tokens": 2288
        },
        {
          "skill_id": "skill-optimizer",
          "tokens": 2940
        },
        {
          "skill_id": "anthropic-canvas-design",
          "tokens": 3199
        },
        {
          "skill_id": "nlm-arxiv-slides",
          "tokens": 3484
        },
        {
          "skill_id": "anthropic-doc-coauthoring",
          "tokens": 4138
        },
        {
          "skill_id": "anthropic-skill-creator",
          "tokens": 6862
        },
        {
          "skill_id": "gws-recipe-doc-from-template",
          "tokens": 743
        },
        {
          "skill_id": "gws-sheets",
          "tokens": 918
        },
        {
          "skill_id": "autoskill-merger",
          "tokens": 1142
        },
        {
          "skill_id": "video-compress",
          "tokens": 1230
        },
        {
          "skill_id": "transcribee",
          "tokens": 1255
        },
        {
          "skill_id": "hf-buckets",
          "tokens": 1334
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PPTX Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-pptx\", \"anthropic-pdf\", \"anthropic-docx\", \"ppt-template-engine\", \"anthropic-xlsx\", \"docx-template-engine\", \"office-template-enforcer\", \"skill-seekers\", \"nlm-slides\", \"notion-meeting-sync\", \"gws-docs\", \"alphaxiv-paper-lookup\", \"cognee\", \"docs-tutor-setup\", \"skill-optimizer\", \"anthropic-canvas-design\", \"nlm-arxiv-slides\", \"anthropic-doc-coauthoring\", \"anthropic-skill-creator\", \"gws-recipe-doc-from-template\", \"gws-sheets\", \"autoskill-merger\", \"video-compress\", \"transcribee\", \"hf-buckets\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0152",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AlphaEar Logic Visualizer. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "alphaear-logic-visualizer",
        "kwp-engineering-system-design",
        "alphaear-reporter",
        "alphaear-signal-tracker",
        "visual-explainer",
        "alphaear-search",
        "alphaear-news",
        "weekly-stock-update",
        "design-architect",
        "agency-visual-storyteller",
        "agency-backend-architect",
        "agency-ui-designer",
        "hf-evaluation",
        "dependency-radar",
        "kwp-engineering-testing-strategy",
        "kwp-design-design-system-management",
        "kwp-design-design-critique",
        "anthropic-brand-guidelines",
        "alphaear-sentiment",
        "agency-data-analytics-reporter",
        "backend-expert",
        "ux-expert",
        "frontend-expert",
        "technical-writer",
        "stock-csv-downloader"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "alphaear-logic-visualizer",
          "tokens": 740
        },
        {
          "skill_id": "kwp-engineering-system-design",
          "tokens": 756
        },
        {
          "skill_id": "alphaear-reporter",
          "tokens": 810
        },
        {
          "skill_id": "alphaear-signal-tracker",
          "tokens": 854
        },
        {
          "skill_id": "visual-explainer",
          "tokens": 5408
        },
        {
          "skill_id": "alphaear-search",
          "tokens": 672
        },
        {
          "skill_id": "alphaear-news",
          "tokens": 873
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-backend-architect",
          "tokens": 2600
        },
        {
          "skill_id": "agency-ui-designer",
          "tokens": 3538
        },
        {
          "skill_id": "hf-evaluation",
          "tokens": 4549
        },
        {
          "skill_id": "dependency-radar",
          "tokens": 531
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "kwp-design-design-critique",
          "tokens": 671
        },
        {
          "skill_id": "anthropic-brand-guidelines",
          "tokens": 781
        },
        {
          "skill_id": "alphaear-sentiment",
          "tokens": 928
        },
        {
          "skill_id": "agency-data-analytics-reporter",
          "tokens": 1003
        },
        {
          "skill_id": "backend-expert",
          "tokens": 1132
        },
        {
          "skill_id": "ux-expert",
          "tokens": 1210
        },
        {
          "skill_id": "frontend-expert",
          "tokens": 1234
        },
        {
          "skill_id": "technical-writer",
          "tokens": 1250
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AlphaEar Logic Visualizer. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-logic-visualizer\", \"kwp-engineering-system-design\", \"alphaear-reporter\", \"alphaear-signal-tracker\", \"visual-explainer\", \"alphaear-search\", \"alphaear-news\", \"weekly-stock-update\", \"design-architect\", \"agency-visual-storyteller\", \"agency-backend-architect\", \"agency-ui-designer\", \"hf-evaluation\", \"dependency-radar\", \"kwp-engineering-testing-strategy\", \"kwp-design-design-system-management\", \"kwp-design-design-critique\", \"anthropic-brand-guidelines\", \"alphaear-sentiment\", \"agency-data-analytics-reporter\", \"backend-expert\", \"ux-expert\", \"frontend-expert\", \"technical-writer\", \"stock-csv-downloader\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0153",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Marketing Instagram Curator. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-instagram-curator",
        "agency-xiaohongshu-specialist",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "agency-wechat-official-account-manager",
        "agency-visual-storyteller",
        "agency-twitter-engager",
        "agency-zhihu-strategist",
        "agency-content-creator",
        "agency-app-store-optimizer",
        "agency-ai-engineer",
        "agency-infrastructure-maintainer",
        "agency-support-responder",
        "agency-developer-advocate",
        "agency-tool-evaluator",
        "agency-whimsy-injector",
        "agency-studio-operations",
        "agency-workflow-optimizer",
        "agency-api-tester",
        "agency-legal-compliance-checker",
        "agency-test-results-analyzer",
        "agency-data-engineer",
        "agency-terminal-integration-specialist",
        "agency-inclusive-visuals-specialist",
        "agency-social-media-strategist"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-zhihu-strategist",
          "tokens": 3239
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "agency-app-store-optimizer",
          "tokens": 3512
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-support-responder",
          "tokens": 3296
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "agency-tool-evaluator",
          "tokens": 4549
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "agency-data-engineer",
          "tokens": 3843
        },
        {
          "skill_id": "agency-terminal-integration-specialist",
          "tokens": 1273
        },
        {
          "skill_id": "agency-inclusive-visuals-specialist",
          "tokens": 1931
        },
        {
          "skill_id": "agency-social-media-strategist",
          "tokens": 2084
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Marketing Instagram Curator. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-instagram-curator\", \"agency-xiaohongshu-specialist\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"agency-wechat-official-account-manager\", \"agency-visual-storyteller\", \"agency-twitter-engager\", \"agency-zhihu-strategist\", \"agency-content-creator\", \"agency-app-store-optimizer\", \"agency-ai-engineer\", \"agency-infrastructure-maintainer\", \"agency-support-responder\", \"agency-developer-advocate\", \"agency-tool-evaluator\", \"agency-whimsy-injector\", \"agency-studio-operations\", \"agency-workflow-optimizer\", \"agency-api-tester\", \"agency-legal-compliance-checker\", \"agency-test-results-analyzer\", \"agency-data-engineer\", \"agency-terminal-integration-specialist\", \"agency-inclusive-visuals-specialist\", \"agency-social-media-strategist\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0154",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: NLM Video: Local Document to Expert Video Pipeline. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "nlm-video",
        "nlm-slides",
        "nlm-arxiv-slides",
        "nlm-launch-kit",
        "notebooklm-studio",
        "nlm-strategy-hub",
        "nlm-discovery-lab",
        "nlm-deep-learn",
        "gws-docs",
        "kwp-brand-voice-guideline-generation",
        "docs-tutor-setup",
        "gws-recipe-doc-from-template",
        "prompt-transformer",
        "md-to-slack-canvas",
        "hf-auto-finetune-planner",
        "agency-technical-writer",
        "paper-review",
        "gws-recipe-save-email-to-doc",
        "knowledge-to-paper",
        "ppt-template-engine",
        "docx-template-engine",
        "commit-to-issue",
        "skill-seekers",
        "md-enhance-publish",
        "skill-guide-generator"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "nlm-video",
          "tokens": 1763
        },
        {
          "skill_id": "nlm-slides",
          "tokens": 1685
        },
        {
          "skill_id": "nlm-arxiv-slides",
          "tokens": 3484
        },
        {
          "skill_id": "nlm-launch-kit",
          "tokens": 2368
        },
        {
          "skill_id": "notebooklm-studio",
          "tokens": 1812
        },
        {
          "skill_id": "nlm-strategy-hub",
          "tokens": 2207
        },
        {
          "skill_id": "nlm-discovery-lab",
          "tokens": 2808
        },
        {
          "skill_id": "nlm-deep-learn",
          "tokens": 3214
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "docs-tutor-setup",
          "tokens": 2288
        },
        {
          "skill_id": "gws-recipe-doc-from-template",
          "tokens": 743
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "md-to-slack-canvas",
          "tokens": 1960
        },
        {
          "skill_id": "hf-auto-finetune-planner",
          "tokens": 2091
        },
        {
          "skill_id": "agency-technical-writer",
          "tokens": 3748
        },
        {
          "skill_id": "paper-review",
          "tokens": 5815
        },
        {
          "skill_id": "gws-recipe-save-email-to-doc",
          "tokens": 676
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "ppt-template-engine",
          "tokens": 1101
        },
        {
          "skill_id": "docx-template-engine",
          "tokens": 1193
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "skill-seekers",
          "tokens": 1568
        },
        {
          "skill_id": "md-enhance-publish",
          "tokens": 1740
        },
        {
          "skill_id": "skill-guide-generator",
          "tokens": 1858
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: NLM Video: Local Document to Expert Video Pipeline. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"nlm-video\", \"nlm-slides\", \"nlm-arxiv-slides\", \"nlm-launch-kit\", \"notebooklm-studio\", \"nlm-strategy-hub\", \"nlm-discovery-lab\", \"nlm-deep-learn\", \"gws-docs\", \"kwp-brand-voice-guideline-generation\", \"docs-tutor-setup\", \"gws-recipe-doc-from-template\", \"prompt-transformer\", \"md-to-slack-canvas\", \"hf-auto-finetune-planner\", \"agency-technical-writer\", \"paper-review\", \"gws-recipe-save-email-to-doc\", \"knowledge-to-paper\", \"ppt-template-engine\", \"docx-template-engine\", \"commit-to-issue\", \"skill-seekers\", \"md-enhance-publish\", \"skill-guide-generator\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0155",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Recipe: Create a Google Doc from a Template. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "gws-recipe-doc-from-template",
        "gws-docs",
        "gws-recipe-events-from-sheet",
        "gws-recipe-save-email-to-doc",
        "gws-sheets",
        "gws-drive",
        "gws-workspace",
        "docx-template-engine",
        "gws-recipe-personalized-emails",
        "ppt-template-engine",
        "notebooklm-studio",
        "docs-tutor-setup",
        "anthropic-docx",
        "smart-meeting-scheduler",
        "gws-calendar",
        "office-template-enforcer",
        "commit-to-issue",
        "nlm-slides",
        "nlm-video",
        "skill-guide-generator",
        "proposal-factory",
        "gws-recipe-audit-sharing",
        "kwp-engineering-documentation",
        "domain-commit",
        "hf-repos"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "gws-recipe-doc-from-template",
          "tokens": 743
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "gws-recipe-events-from-sheet",
          "tokens": 674
        },
        {
          "skill_id": "gws-recipe-save-email-to-doc",
          "tokens": 676
        },
        {
          "skill_id": "gws-sheets",
          "tokens": 918
        },
        {
          "skill_id": "gws-drive",
          "tokens": 997
        },
        {
          "skill_id": "gws-workspace",
          "tokens": 1186
        },
        {
          "skill_id": "docx-template-engine",
          "tokens": 1193
        },
        {
          "skill_id": "gws-recipe-personalized-emails",
          "tokens": 641
        },
        {
          "skill_id": "ppt-template-engine",
          "tokens": 1101
        },
        {
          "skill_id": "notebooklm-studio",
          "tokens": 1812
        },
        {
          "skill_id": "docs-tutor-setup",
          "tokens": 2288
        },
        {
          "skill_id": "anthropic-docx",
          "tokens": 4265
        },
        {
          "skill_id": "smart-meeting-scheduler",
          "tokens": 639
        },
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "office-template-enforcer",
          "tokens": 1273
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "nlm-slides",
          "tokens": 1685
        },
        {
          "skill_id": "nlm-video",
          "tokens": 1763
        },
        {
          "skill_id": "skill-guide-generator",
          "tokens": 1858
        },
        {
          "skill_id": "proposal-factory",
          "tokens": 582
        },
        {
          "skill_id": "gws-recipe-audit-sharing",
          "tokens": 645
        },
        {
          "skill_id": "kwp-engineering-documentation",
          "tokens": 654
        },
        {
          "skill_id": "domain-commit",
          "tokens": 915
        },
        {
          "skill_id": "hf-repos",
          "tokens": 1106
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Recipe: Create a Google Doc from a Template. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"gws-recipe-doc-from-template\", \"gws-docs\", \"gws-recipe-events-from-sheet\", \"gws-recipe-save-email-to-doc\", \"gws-sheets\", \"gws-drive\", \"gws-workspace\", \"docx-template-engine\", \"gws-recipe-personalized-emails\", \"ppt-template-engine\", \"notebooklm-studio\", \"docs-tutor-setup\", \"anthropic-docx\", \"smart-meeting-scheduler\", \"gws-calendar\", \"office-template-enforcer\", \"commit-to-issue\", \"nlm-slides\", \"nlm-video\", \"skill-guide-generator\", \"proposal-factory\", \"gws-recipe-audit-sharing\", \"kwp-engineering-documentation\", \"domain-commit\", \"hf-repos\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0156",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: QA Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-evidence-collector",
        "agency-reality-checker",
        "agency-instagram-curator",
        "agency-visual-storyteller",
        "agency-xr-cockpit-interaction-specialist",
        "agency-terminal-integration-specialist",
        "agency-behavioral-nudge-engine",
        "agency-senior-developer",
        "agency-cultural-intelligence-strategist",
        "agency-inclusive-visuals-specialist",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "agency-twitter-engager",
        "agency-whimsy-injector",
        "agency-rapid-prototyper",
        "agency-xiaohongshu-specialist",
        "agency-workflow-optimizer",
        "agency-lsp-index-engineer",
        "agency-image-prompt-engineer",
        "agency-ux-architect",
        "agency-macos-spatial-metal-engineer",
        "agency-zhihu-strategist",
        "agency-infrastructure-maintainer",
        "agency-support-responder",
        "agency-api-tester"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-evidence-collector",
          "tokens": 2300
        },
        {
          "skill_id": "agency-reality-checker",
          "tokens": 2764
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-xr-cockpit-interaction-specialist",
          "tokens": 693
        },
        {
          "skill_id": "agency-terminal-integration-specialist",
          "tokens": 1273
        },
        {
          "skill_id": "agency-behavioral-nudge-engine",
          "tokens": 1553
        },
        {
          "skill_id": "agency-senior-developer",
          "tokens": 1781
        },
        {
          "skill_id": "agency-cultural-intelligence-strategist",
          "tokens": 1892
        },
        {
          "skill_id": "agency-inclusive-visuals-specialist",
          "tokens": 1931
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-rapid-prototyper",
          "tokens": 2496
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-lsp-index-engineer",
          "tokens": 3006
        },
        {
          "skill_id": "agency-image-prompt-engineer",
          "tokens": 3027
        },
        {
          "skill_id": "agency-ux-architect",
          "tokens": 3119
        },
        {
          "skill_id": "agency-macos-spatial-metal-engineer",
          "tokens": 3229
        },
        {
          "skill_id": "agency-zhihu-strategist",
          "tokens": 3239
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-support-responder",
          "tokens": 3296
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: QA Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-evidence-collector\", \"agency-reality-checker\", \"agency-instagram-curator\", \"agency-visual-storyteller\", \"agency-xr-cockpit-interaction-specialist\", \"agency-terminal-integration-specialist\", \"agency-behavioral-nudge-engine\", \"agency-senior-developer\", \"agency-cultural-intelligence-strategist\", \"agency-inclusive-visuals-specialist\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"agency-twitter-engager\", \"agency-whimsy-injector\", \"agency-rapid-prototyper\", \"agency-xiaohongshu-specialist\", \"agency-workflow-optimizer\", \"agency-lsp-index-engineer\", \"agency-image-prompt-engineer\", \"agency-ux-architect\", \"agency-macos-spatial-metal-engineer\", \"agency-zhihu-strategist\", \"agency-infrastructure-maintainer\", \"agency-support-responder\", \"agency-api-tester\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0157",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AlphaEar DeepEar Lite. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "alphaear-deepear-lite",
        "alphaear-news",
        "alphaear-sentiment",
        "trading-market-environment-analysis",
        "trading-us-stock-analysis",
        "alphaear-predictor",
        "alphaear-stock",
        "stock-csv-downloader",
        "daily-stock-check",
        "alphaear-search",
        "weekly-stock-update",
        "trading-sector-analyst",
        "trading-technical-analyst",
        "alphaear-signal-tracker",
        "trading-scenario-analyzer",
        "mirofish-financial-sim",
        "trading-market-news-analyst",
        "trading-market-breadth-analyzer",
        "trading-us-market-bubble-detector",
        "trading-breadth-chart-analyst",
        "alphaear-reporter",
        "trading-uptrend-analyzer",
        "ai-workflow-integrator",
        "release-commander",
        "mission-control"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        },
        {
          "skill_id": "alphaear-news",
          "tokens": 873
        },
        {
          "skill_id": "alphaear-sentiment",
          "tokens": 928
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "trading-us-stock-analysis",
          "tokens": 2926
        },
        {
          "skill_id": "alphaear-predictor",
          "tokens": 803
        },
        {
          "skill_id": "alphaear-stock",
          "tokens": 820
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "alphaear-search",
          "tokens": 672
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "trading-sector-analyst",
          "tokens": 2657
        },
        {
          "skill_id": "trading-technical-analyst",
          "tokens": 2868
        },
        {
          "skill_id": "alphaear-signal-tracker",
          "tokens": 854
        },
        {
          "skill_id": "trading-scenario-analyzer",
          "tokens": 2068
        },
        {
          "skill_id": "mirofish-financial-sim",
          "tokens": 1330
        },
        {
          "skill_id": "trading-market-news-analyst",
          "tokens": 1467
        },
        {
          "skill_id": "trading-market-breadth-analyzer",
          "tokens": 1830
        },
        {
          "skill_id": "trading-us-market-bubble-detector",
          "tokens": 3062
        },
        {
          "skill_id": "trading-breadth-chart-analyst",
          "tokens": 3912
        },
        {
          "skill_id": "alphaear-reporter",
          "tokens": 810
        },
        {
          "skill_id": "trading-uptrend-analyzer",
          "tokens": 1780
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AlphaEar DeepEar Lite. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-deepear-lite\", \"alphaear-news\", \"alphaear-sentiment\", \"trading-market-environment-analysis\", \"trading-us-stock-analysis\", \"alphaear-predictor\", \"alphaear-stock\", \"stock-csv-downloader\", \"daily-stock-check\", \"alphaear-search\", \"weekly-stock-update\", \"trading-sector-analyst\", \"trading-technical-analyst\", \"alphaear-signal-tracker\", \"trading-scenario-analyzer\", \"mirofish-financial-sim\", \"trading-market-news-analyst\", \"trading-market-breadth-analyzer\", \"trading-us-market-bubble-detector\", \"trading-breadth-chart-analyst\", \"alphaear-reporter\", \"trading-uptrend-analyzer\", \"ai-workflow-integrator\", \"release-commander\", \"mission-control\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0158",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Stock CSV Downloader. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "stock-csv-downloader",
        "weekly-stock-update",
        "daily-stock-check",
        "alphaear-stock",
        "alphaear-deepear-lite",
        "trading-sector-analyst",
        "alphaear-search",
        "alphaear-news",
        "trading-technical-analyst",
        "tab-stock-sync",
        "trading-market-environment-analysis",
        "alphaear-reporter",
        "trading-trade-hypothesis-ideator",
        "weekly-status-report",
        "ai-workflow-integrator",
        "trading-backtest-expert",
        "trading-us-stock-analysis",
        "trading-theme-detector",
        "tab-kiwoom",
        "trading-market-breadth-analyzer",
        "trading-edge-candidate-agent",
        "trading-us-market-bubble-detector",
        "tab-report-generate",
        "github-sprint-digest",
        "gws-recipe-events-from-sheet"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "alphaear-stock",
          "tokens": 820
        },
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        },
        {
          "skill_id": "trading-sector-analyst",
          "tokens": 2657
        },
        {
          "skill_id": "alphaear-search",
          "tokens": 672
        },
        {
          "skill_id": "alphaear-news",
          "tokens": 873
        },
        {
          "skill_id": "trading-technical-analyst",
          "tokens": 2868
        },
        {
          "skill_id": "tab-stock-sync",
          "tokens": 345
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "alphaear-reporter",
          "tokens": 810
        },
        {
          "skill_id": "trading-trade-hypothesis-ideator",
          "tokens": 909
        },
        {
          "skill_id": "weekly-status-report",
          "tokens": 1282
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "trading-backtest-expert",
          "tokens": 2754
        },
        {
          "skill_id": "trading-us-stock-analysis",
          "tokens": 2926
        },
        {
          "skill_id": "trading-theme-detector",
          "tokens": 3487
        },
        {
          "skill_id": "tab-kiwoom",
          "tokens": 452
        },
        {
          "skill_id": "trading-market-breadth-analyzer",
          "tokens": 1830
        },
        {
          "skill_id": "trading-edge-candidate-agent",
          "tokens": 1971
        },
        {
          "skill_id": "trading-us-market-bubble-detector",
          "tokens": 3062
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "gws-recipe-events-from-sheet",
          "tokens": 674
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Stock CSV Downloader. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"stock-csv-downloader\", \"weekly-stock-update\", \"daily-stock-check\", \"alphaear-stock\", \"alphaear-deepear-lite\", \"trading-sector-analyst\", \"alphaear-search\", \"alphaear-news\", \"trading-technical-analyst\", \"tab-stock-sync\", \"trading-market-environment-analysis\", \"alphaear-reporter\", \"trading-trade-hypothesis-ideator\", \"weekly-status-report\", \"ai-workflow-integrator\", \"trading-backtest-expert\", \"trading-us-stock-analysis\", \"trading-theme-detector\", \"tab-kiwoom\", \"trading-market-breadth-analyzer\", \"trading-edge-candidate-agent\", \"trading-us-market-bubble-detector\", \"tab-report-generate\", \"github-sprint-digest\", \"gws-recipe-events-from-sheet\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0159",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Developer Advocate Agent. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-developer-advocate",
        "agency-visual-storyteller",
        "agency-wechat-official-account-manager",
        "agency-content-creator",
        "agency-reddit-community-builder",
        "agency-twitter-engager",
        "agency-xiaohongshu-specialist",
        "agency-instagram-curator",
        "agency-ai-engineer",
        "agency-social-media-strategist",
        "agency-studio-operations",
        "agency-zhihu-strategist",
        "agency-xr-immersive-developer",
        "agency-growth-hacker",
        "agency-tiktok-strategist",
        "agency-experiment-tracker",
        "agency-mobile-app-builder",
        "agency-project-shepherd",
        "agency-trend-researcher",
        "agency-studio-producer",
        "agency-ux-architect",
        "agency-infrastructure-maintainer",
        "agency-legal-compliance-checker",
        "agency-data-engineer",
        "agency-feedback-synthesizer"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-social-media-strategist",
          "tokens": 2084
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-zhihu-strategist",
          "tokens": 3239
        },
        {
          "skill_id": "agency-xr-immersive-developer",
          "tokens": 697
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-mobile-app-builder",
          "tokens": 2595
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "agency-studio-producer",
          "tokens": 2953
        },
        {
          "skill_id": "agency-ux-architect",
          "tokens": 3119
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-data-engineer",
          "tokens": 3843
        },
        {
          "skill_id": "agency-feedback-synthesizer",
          "tokens": 2041
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Developer Advocate Agent. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-developer-advocate\", \"agency-visual-storyteller\", \"agency-wechat-official-account-manager\", \"agency-content-creator\", \"agency-reddit-community-builder\", \"agency-twitter-engager\", \"agency-xiaohongshu-specialist\", \"agency-instagram-curator\", \"agency-ai-engineer\", \"agency-social-media-strategist\", \"agency-studio-operations\", \"agency-zhihu-strategist\", \"agency-xr-immersive-developer\", \"agency-growth-hacker\", \"agency-tiktok-strategist\", \"agency-experiment-tracker\", \"agency-mobile-app-builder\", \"agency-project-shepherd\", \"agency-trend-researcher\", \"agency-studio-producer\", \"agency-ux-architect\", \"agency-infrastructure-maintainer\", \"agency-legal-compliance-checker\", \"agency-data-engineer\", \"agency-feedback-synthesizer\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0160",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Onboarding Accelerator. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "onboarding-accelerator",
        "docs-tutor-setup",
        "skill-guide-generator",
        "codebase-archaeologist",
        "full-stack-planner",
        "demo-forge",
        "setup-doctor",
        "visual-explainer",
        "role-developer",
        "infra-runbook-generator",
        "docs-tutor",
        "pipeline-builder",
        "release-commander",
        "incident-to-improvement",
        "critical-review",
        "agency-roster",
        "backend-expert",
        "technical-writer",
        "sre-devops-expert",
        "evals-skills",
        "issue-to-code",
        "refactor-simulator",
        "ai-workflow-integrator",
        "ai-quality-evaluator",
        "slack-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "onboarding-accelerator",
          "tokens": 2419
        },
        {
          "skill_id": "docs-tutor-setup",
          "tokens": 2288
        },
        {
          "skill_id": "skill-guide-generator",
          "tokens": 1858
        },
        {
          "skill_id": "codebase-archaeologist",
          "tokens": 1977
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        },
        {
          "skill_id": "demo-forge",
          "tokens": 2152
        },
        {
          "skill_id": "setup-doctor",
          "tokens": 2578
        },
        {
          "skill_id": "visual-explainer",
          "tokens": 5408
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "infra-runbook-generator",
          "tokens": 1623
        },
        {
          "skill_id": "docs-tutor",
          "tokens": 1759
        },
        {
          "skill_id": "pipeline-builder",
          "tokens": 2234
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "incident-to-improvement",
          "tokens": 2624
        },
        {
          "skill_id": "critical-review",
          "tokens": 3039
        },
        {
          "skill_id": "agency-roster",
          "tokens": 3411
        },
        {
          "skill_id": "backend-expert",
          "tokens": 1132
        },
        {
          "skill_id": "technical-writer",
          "tokens": 1250
        },
        {
          "skill_id": "sre-devops-expert",
          "tokens": 1450
        },
        {
          "skill_id": "evals-skills",
          "tokens": 1524
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "refactor-simulator",
          "tokens": 2044
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "ai-quality-evaluator",
          "tokens": 2228
        },
        {
          "skill_id": "slack-agent",
          "tokens": 2564
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Onboarding Accelerator. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"onboarding-accelerator\", \"docs-tutor-setup\", \"skill-guide-generator\", \"codebase-archaeologist\", \"full-stack-planner\", \"demo-forge\", \"setup-doctor\", \"visual-explainer\", \"role-developer\", \"infra-runbook-generator\", \"docs-tutor\", \"pipeline-builder\", \"release-commander\", \"incident-to-improvement\", \"critical-review\", \"agency-roster\", \"backend-expert\", \"technical-writer\", \"sre-devops-expert\", \"evals-skills\", \"issue-to-code\", \"refactor-simulator\", \"ai-workflow-integrator\", \"ai-quality-evaluator\", \"slack-agent\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0161",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Vendor Management. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-operations-vendor-management",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-operations-resource-planning",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-human-resources-interview-prep",
        "kwp-human-resources-people-analytics",
        "kwp-design-design-system-management",
        "kwp-design-design-critique",
        "kwp-engineering-system-design",
        "kwp-sales-call-prep",
        "kwp-product-management-feature-spec",
        "kwp-engineering-testing-strategy",
        "kwp-engineering-code-review",
        "kwp-engineering-incident-response",
        "kwp-engineering-tech-debt",
        "kwp-design-accessibility-review",
        "kwp-operations-risk-assessment",
        "kwp-brand-voice-discover-brand",
        "kwp-sales-daily-briefing",
        "kwp-sales-account-research",
        "kwp-finance-close-management",
        "kwp-legal-contract-review"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-recruiting-pipeline",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "kwp-design-design-critique",
          "tokens": 671
        },
        {
          "skill_id": "kwp-engineering-system-design",
          "tokens": 756
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-product-management-feature-spec",
          "tokens": 2614
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "kwp-engineering-code-review",
          "tokens": 610
        },
        {
          "skill_id": "kwp-engineering-incident-response",
          "tokens": 645
        },
        {
          "skill_id": "kwp-engineering-tech-debt",
          "tokens": 648
        },
        {
          "skill_id": "kwp-design-accessibility-review",
          "tokens": 679
        },
        {
          "skill_id": "kwp-operations-risk-assessment",
          "tokens": 808
        },
        {
          "skill_id": "kwp-brand-voice-discover-brand",
          "tokens": 1897
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "kwp-finance-close-management",
          "tokens": 2782
        },
        {
          "skill_id": "kwp-legal-contract-review",
          "tokens": 3137
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Vendor Management. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-operations-vendor-management\", \"kwp-operations-process-optimization\", \"kwp-operations-change-management\", \"kwp-operations-resource-planning\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-human-resources-interview-prep\", \"kwp-human-resources-people-analytics\", \"kwp-design-design-system-management\", \"kwp-design-design-critique\", \"kwp-engineering-system-design\", \"kwp-sales-call-prep\", \"kwp-product-management-feature-spec\", \"kwp-engineering-testing-strategy\", \"kwp-engineering-code-review\", \"kwp-engineering-incident-response\", \"kwp-engineering-tech-debt\", \"kwp-design-accessibility-review\", \"kwp-operations-risk-assessment\", \"kwp-brand-voice-discover-brand\", \"kwp-sales-daily-briefing\", \"kwp-sales-account-research\", \"kwp-finance-close-management\", \"kwp-legal-contract-review\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0162",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Subagent-Driven Development. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "sp-subagent-dev",
        "sp-git-worktrees",
        "kwp-design-design-handoff",
        "kwp-human-resources-interview-prep",
        "sp-executing-plans",
        "ship",
        "workflow-parallel",
        "design-architect",
        "test-suite",
        "kwp-cowork-create-plugin",
        "kwp-human-resources-people-analytics",
        "kwp-operations-vendor-management",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-operations-process-optimization",
        "kwp-engineering-testing-strategy",
        "kwp-operations-change-management",
        "kwp-operations-resource-planning",
        "hf-papers",
        "sp-brainstorming",
        "diagnose",
        "workflow-sequential",
        "ui-suite",
        "deep-review"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "sp-subagent-dev",
          "tokens": 2538
        },
        {
          "skill_id": "sp-git-worktrees",
          "tokens": 1452
        },
        {
          "skill_id": "kwp-design-design-handoff",
          "tokens": 649
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "sp-executing-plans",
          "tokens": 826
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "workflow-parallel",
          "tokens": 1734
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "kwp-cowork-create-plugin",
          "tokens": 3214
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-recruiting-pipeline",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "hf-papers",
          "tokens": 817
        },
        {
          "skill_id": "sp-brainstorming",
          "tokens": 1362
        },
        {
          "skill_id": "diagnose",
          "tokens": 1433
        },
        {
          "skill_id": "workflow-sequential",
          "tokens": 1485
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "deep-review",
          "tokens": 1923
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Subagent-Driven Development. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"sp-subagent-dev\", \"sp-git-worktrees\", \"kwp-design-design-handoff\", \"kwp-human-resources-interview-prep\", \"sp-executing-plans\", \"ship\", \"workflow-parallel\", \"design-architect\", \"test-suite\", \"kwp-cowork-create-plugin\", \"kwp-human-resources-people-analytics\", \"kwp-operations-vendor-management\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-operations-process-optimization\", \"kwp-engineering-testing-strategy\", \"kwp-operations-change-management\", \"kwp-operations-resource-planning\", \"hf-papers\", \"sp-brainstorming\", \"diagnose\", \"workflow-sequential\", \"ui-suite\", \"deep-review\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0163",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Daily Sales Briefing. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-sales-daily-briefing",
        "kwp-sales-call-prep",
        "kwp-sales-account-research",
        "ai-chief-of-staff",
        "kwp-sales-draft-outreach",
        "calendar-daily-briefing",
        "kwp-sales-competitive-intelligence",
        "kwp-operations-resource-planning",
        "kwp-finance-close-management",
        "morning-ship",
        "kwp-marketing-performance-analytics",
        "kwp-human-resources-people-analytics",
        "kwp-operations-vendor-management",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-human-resources-interview-prep",
        "kwp-common-room-weekly-prep-brief",
        "kwp-customer-support-escalation",
        "kwp-legal-meeting-briefing",
        "kwp-marketing-content-creation",
        "kwp-marketing-campaign-planning",
        "kwp-product-management-competitive-analysis"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "ai-chief-of-staff",
          "tokens": 1152
        },
        {
          "skill_id": "kwp-sales-draft-outreach",
          "tokens": 2782
        },
        {
          "skill_id": "calendar-daily-briefing",
          "tokens": 1042
        },
        {
          "skill_id": "kwp-sales-competitive-intelligence",
          "tokens": 3361
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-finance-close-management",
          "tokens": 2782
        },
        {
          "skill_id": "morning-ship",
          "tokens": 2943
        },
        {
          "skill_id": "kwp-marketing-performance-analytics",
          "tokens": 3966
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-recruiting-pipeline",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "kwp-common-room-weekly-prep-brief",
          "tokens": 1607
        },
        {
          "skill_id": "kwp-customer-support-escalation",
          "tokens": 2329
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "kwp-marketing-campaign-planning",
          "tokens": 2803
        },
        {
          "skill_id": "kwp-product-management-competitive-analysis",
          "tokens": 2899
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Daily Sales Briefing. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-sales-daily-briefing\", \"kwp-sales-call-prep\", \"kwp-sales-account-research\", \"ai-chief-of-staff\", \"kwp-sales-draft-outreach\", \"calendar-daily-briefing\", \"kwp-sales-competitive-intelligence\", \"kwp-operations-resource-planning\", \"kwp-finance-close-management\", \"morning-ship\", \"kwp-marketing-performance-analytics\", \"kwp-human-resources-people-analytics\", \"kwp-operations-vendor-management\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-operations-process-optimization\", \"kwp-operations-change-management\", \"kwp-human-resources-interview-prep\", \"kwp-common-room-weekly-prep-brief\", \"kwp-customer-support-escalation\", \"kwp-legal-meeting-briefing\", \"kwp-marketing-content-creation\", \"kwp-marketing-campaign-planning\", \"kwp-product-management-competitive-analysis\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0164",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: /search-first. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "ecc-search-first",
        "issue-to-code",
        "lead-lifecycle-manager",
        "hf-papers",
        "recall",
        "sp-git-worktrees",
        "sp-tdd",
        "docs-freshness-guardian",
        "alphaear-search",
        "ecc-agentic-engineering",
        "ecc-coding-standards",
        "frontend-expert",
        "ecc-strategic-compact",
        "diagnose",
        "ui-suite",
        "kwp-sales-account-research",
        "mission-control",
        "problem-definition",
        "kwp-product-management-feature-spec",
        "skill-optimizer",
        "ecc-configure",
        "tech-support-brain",
        "sp-verification",
        "sp-core",
        "paper-auto-classifier"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "ecc-search-first",
          "tokens": 1597
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "lead-lifecycle-manager",
          "tokens": 533
        },
        {
          "skill_id": "hf-papers",
          "tokens": 817
        },
        {
          "skill_id": "recall",
          "tokens": 1256
        },
        {
          "skill_id": "sp-git-worktrees",
          "tokens": 1452
        },
        {
          "skill_id": "sp-tdd",
          "tokens": 2511
        },
        {
          "skill_id": "docs-freshness-guardian",
          "tokens": 557
        },
        {
          "skill_id": "alphaear-search",
          "tokens": 672
        },
        {
          "skill_id": "ecc-agentic-engineering",
          "tokens": 690
        },
        {
          "skill_id": "ecc-coding-standards",
          "tokens": 1050
        },
        {
          "skill_id": "frontend-expert",
          "tokens": 1234
        },
        {
          "skill_id": "ecc-strategic-compact",
          "tokens": 1252
        },
        {
          "skill_id": "diagnose",
          "tokens": 1433
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "problem-definition",
          "tokens": 2601
        },
        {
          "skill_id": "kwp-product-management-feature-spec",
          "tokens": 2614
        },
        {
          "skill_id": "skill-optimizer",
          "tokens": 2940
        },
        {
          "skill_id": "ecc-configure",
          "tokens": 3534
        },
        {
          "skill_id": "tech-support-brain",
          "tokens": 620
        },
        {
          "skill_id": "sp-verification",
          "tokens": 1146
        },
        {
          "skill_id": "sp-core",
          "tokens": 1164
        },
        {
          "skill_id": "paper-auto-classifier",
          "tokens": 1262
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: /search-first. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"ecc-search-first\", \"issue-to-code\", \"lead-lifecycle-manager\", \"hf-papers\", \"recall\", \"sp-git-worktrees\", \"sp-tdd\", \"docs-freshness-guardian\", \"alphaear-search\", \"ecc-agentic-engineering\", \"ecc-coding-standards\", \"frontend-expert\", \"ecc-strategic-compact\", \"diagnose\", \"ui-suite\", \"kwp-sales-account-research\", \"mission-control\", \"problem-definition\", \"kwp-product-management-feature-spec\", \"skill-optimizer\", \"ecc-configure\", \"tech-support-brain\", \"sp-verification\", \"sp-core\", \"paper-auto-classifier\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0165",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Gmail Daily Triage. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "gmail-daily-triage",
        "gws-gmail",
        "proactive-meeting-scheduler",
        "email-auto-reply",
        "ai-chief-of-staff",
        "bespin-news-digest",
        "gws-recipe-personalized-emails",
        "gws-workflows",
        "gws-calendar",
        "gws-email-reply",
        "planning-meeting-sentinel",
        "calendar-daily-briefing",
        "feedback-meeting-scheduler",
        "executive-briefing",
        "notion-meeting-sync",
        "smart-meeting-scheduler",
        "gws-recipe-save-email-to-doc",
        "email-research-dispatcher",
        "daily-strategy-post",
        "gws-drive",
        "hf-collections",
        "sprint-orchestrator",
        "jira-linear-adapter",
        "portfolio-report-generator",
        "cursor-automations"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "gmail-daily-triage",
          "tokens": 2229
        },
        {
          "skill_id": "gws-gmail",
          "tokens": 1796
        },
        {
          "skill_id": "proactive-meeting-scheduler",
          "tokens": 1444
        },
        {
          "skill_id": "email-auto-reply",
          "tokens": 638
        },
        {
          "skill_id": "ai-chief-of-staff",
          "tokens": 1152
        },
        {
          "skill_id": "bespin-news-digest",
          "tokens": 3949
        },
        {
          "skill_id": "gws-recipe-personalized-emails",
          "tokens": 641
        },
        {
          "skill_id": "gws-workflows",
          "tokens": 952
        },
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "gws-email-reply",
          "tokens": 2222
        },
        {
          "skill_id": "planning-meeting-sentinel",
          "tokens": 604
        },
        {
          "skill_id": "calendar-daily-briefing",
          "tokens": 1042
        },
        {
          "skill_id": "feedback-meeting-scheduler",
          "tokens": 1238
        },
        {
          "skill_id": "executive-briefing",
          "tokens": 1419
        },
        {
          "skill_id": "notion-meeting-sync",
          "tokens": 4347
        },
        {
          "skill_id": "smart-meeting-scheduler",
          "tokens": 639
        },
        {
          "skill_id": "gws-recipe-save-email-to-doc",
          "tokens": 676
        },
        {
          "skill_id": "email-research-dispatcher",
          "tokens": 703
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "gws-drive",
          "tokens": 997
        },
        {
          "skill_id": "hf-collections",
          "tokens": 1239
        },
        {
          "skill_id": "sprint-orchestrator",
          "tokens": 1320
        },
        {
          "skill_id": "jira-linear-adapter",
          "tokens": 1365
        },
        {
          "skill_id": "portfolio-report-generator",
          "tokens": 1406
        },
        {
          "skill_id": "cursor-automations",
          "tokens": 2911
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Gmail Daily Triage. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"gmail-daily-triage\", \"gws-gmail\", \"proactive-meeting-scheduler\", \"email-auto-reply\", \"ai-chief-of-staff\", \"bespin-news-digest\", \"gws-recipe-personalized-emails\", \"gws-workflows\", \"gws-calendar\", \"gws-email-reply\", \"planning-meeting-sentinel\", \"calendar-daily-briefing\", \"feedback-meeting-scheduler\", \"executive-briefing\", \"notion-meeting-sync\", \"smart-meeting-scheduler\", \"gws-recipe-save-email-to-doc\", \"email-research-dispatcher\", \"daily-strategy-post\", \"gws-drive\", \"hf-collections\", \"sprint-orchestrator\", \"jira-linear-adapter\", \"portfolio-report-generator\", \"cursor-automations\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0166",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Market Breadth Analyzer Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "trading-market-breadth-analyzer",
        "trading-uptrend-analyzer",
        "trading-sector-analyst",
        "trading-theme-detector",
        "trading-position-sizer",
        "trading-market-top-detector",
        "alphaear-deepear-lite",
        "weekly-stock-update",
        "daily-stock-check",
        "trading-breadth-chart-analyst",
        "trading-trade-hypothesis-ideator",
        "stock-csv-downloader",
        "trading-market-environment-analysis",
        "trading-technical-analyst",
        "alphaear-sentiment",
        "mirofish-financial-sim",
        "trading-us-stock-analysis",
        "trading-us-market-bubble-detector",
        "alphaear-stock",
        "terraform-reviewer",
        "ai-quality-evaluator",
        "pipeline-builder",
        "trading-stanley-druckenmiller-investment",
        "trading-finviz-screener",
        "tab-dashboard-summary"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "trading-market-breadth-analyzer",
          "tokens": 1830
        },
        {
          "skill_id": "trading-uptrend-analyzer",
          "tokens": 1780
        },
        {
          "skill_id": "trading-sector-analyst",
          "tokens": 2657
        },
        {
          "skill_id": "trading-theme-detector",
          "tokens": 3487
        },
        {
          "skill_id": "trading-position-sizer",
          "tokens": 2013
        },
        {
          "skill_id": "trading-market-top-detector",
          "tokens": 2295
        },
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "trading-breadth-chart-analyst",
          "tokens": 3912
        },
        {
          "skill_id": "trading-trade-hypothesis-ideator",
          "tokens": 909
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "trading-technical-analyst",
          "tokens": 2868
        },
        {
          "skill_id": "alphaear-sentiment",
          "tokens": 928
        },
        {
          "skill_id": "mirofish-financial-sim",
          "tokens": 1330
        },
        {
          "skill_id": "trading-us-stock-analysis",
          "tokens": 2926
        },
        {
          "skill_id": "trading-us-market-bubble-detector",
          "tokens": 3062
        },
        {
          "skill_id": "alphaear-stock",
          "tokens": 820
        },
        {
          "skill_id": "terraform-reviewer",
          "tokens": 1457
        },
        {
          "skill_id": "ai-quality-evaluator",
          "tokens": 2228
        },
        {
          "skill_id": "pipeline-builder",
          "tokens": 2234
        },
        {
          "skill_id": "trading-stanley-druckenmiller-investment",
          "tokens": 2317
        },
        {
          "skill_id": "trading-finviz-screener",
          "tokens": 3681
        },
        {
          "skill_id": "tab-dashboard-summary",
          "tokens": 419
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Market Breadth Analyzer Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"trading-market-breadth-analyzer\", \"trading-uptrend-analyzer\", \"trading-sector-analyst\", \"trading-theme-detector\", \"trading-position-sizer\", \"trading-market-top-detector\", \"alphaear-deepear-lite\", \"weekly-stock-update\", \"daily-stock-check\", \"trading-breadth-chart-analyst\", \"trading-trade-hypothesis-ideator\", \"stock-csv-downloader\", \"trading-market-environment-analysis\", \"trading-technical-analyst\", \"alphaear-sentiment\", \"mirofish-financial-sim\", \"trading-us-stock-analysis\", \"trading-us-market-bubble-detector\", \"alphaear-stock\", \"terraform-reviewer\", \"ai-quality-evaluator\", \"pipeline-builder\", \"trading-stanley-druckenmiller-investment\", \"trading-finviz-screener\", \"tab-dashboard-summary\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0167",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Social Media Strategist Agent. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-social-media-strategist",
        "agency-content-creator",
        "agency-twitter-engager",
        "agency-zhihu-strategist",
        "agency-wechat-official-account-manager",
        "agency-developer-advocate",
        "agency-growth-hacker",
        "agency-xiaohongshu-specialist",
        "agency-instagram-curator",
        "agency-visual-storyteller",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "agency-whimsy-injector",
        "agency-mobile-app-builder",
        "agency-project-shepherd",
        "agency-brand-guardian",
        "agency-data-engineer",
        "agency-tool-evaluator",
        "agency-xr-interface-architect",
        "agency-xr-immersive-developer",
        "agency-cultural-intelligence-strategist",
        "agency-inclusive-visuals-specialist",
        "agency-feedback-synthesizer",
        "agency-ai-engineer",
        "agency-experiment-tracker"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-social-media-strategist",
          "tokens": 2084
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-zhihu-strategist",
          "tokens": 3239
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-mobile-app-builder",
          "tokens": 2595
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-brand-guardian",
          "tokens": 3135
        },
        {
          "skill_id": "agency-data-engineer",
          "tokens": 3843
        },
        {
          "skill_id": "agency-tool-evaluator",
          "tokens": 4549
        },
        {
          "skill_id": "agency-xr-interface-architect",
          "tokens": 657
        },
        {
          "skill_id": "agency-xr-immersive-developer",
          "tokens": 697
        },
        {
          "skill_id": "agency-cultural-intelligence-strategist",
          "tokens": 1892
        },
        {
          "skill_id": "agency-inclusive-visuals-specialist",
          "tokens": 1931
        },
        {
          "skill_id": "agency-feedback-synthesizer",
          "tokens": 2041
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Social Media Strategist Agent. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-social-media-strategist\", \"agency-content-creator\", \"agency-twitter-engager\", \"agency-zhihu-strategist\", \"agency-wechat-official-account-manager\", \"agency-developer-advocate\", \"agency-growth-hacker\", \"agency-xiaohongshu-specialist\", \"agency-instagram-curator\", \"agency-visual-storyteller\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"agency-whimsy-injector\", \"agency-mobile-app-builder\", \"agency-project-shepherd\", \"agency-brand-guardian\", \"agency-data-engineer\", \"agency-tool-evaluator\", \"agency-xr-interface-architect\", \"agency-xr-immersive-developer\", \"agency-cultural-intelligence-strategist\", \"agency-inclusive-visuals-specialist\", \"agency-feedback-synthesizer\", \"agency-ai-engineer\", \"agency-experiment-tracker\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0168",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Image Prompt Engineer Agent. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-image-prompt-engineer",
        "agency-visual-storyteller",
        "agency-data-engineer",
        "agency-ai-engineer",
        "agency-wechat-official-account-manager",
        "agency-growth-hacker",
        "agency-instagram-curator",
        "agency-whimsy-injector",
        "agency-experiment-tracker",
        "agency-xiaohongshu-specialist",
        "agency-studio-operations",
        "agency-trend-researcher",
        "agency-lsp-index-engineer",
        "agency-zhihu-strategist",
        "agency-test-results-analyzer",
        "agency-developer-advocate",
        "agency-data-consolidation-agent",
        "agency-feedback-synthesizer",
        "agency-social-media-strategist",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "agency-twitter-engager",
        "agency-project-shepherd",
        "agency-brand-guardian",
        "agency-finance-tracker"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-image-prompt-engineer",
          "tokens": 3027
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-data-engineer",
          "tokens": 3843
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "agency-lsp-index-engineer",
          "tokens": 3006
        },
        {
          "skill_id": "agency-zhihu-strategist",
          "tokens": 3239
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        },
        {
          "skill_id": "agency-feedback-synthesizer",
          "tokens": 2041
        },
        {
          "skill_id": "agency-social-media-strategist",
          "tokens": 2084
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-brand-guardian",
          "tokens": 3135
        },
        {
          "skill_id": "agency-finance-tracker",
          "tokens": 3177
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Image Prompt Engineer Agent. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-image-prompt-engineer\", \"agency-visual-storyteller\", \"agency-data-engineer\", \"agency-ai-engineer\", \"agency-wechat-official-account-manager\", \"agency-growth-hacker\", \"agency-instagram-curator\", \"agency-whimsy-injector\", \"agency-experiment-tracker\", \"agency-xiaohongshu-specialist\", \"agency-studio-operations\", \"agency-trend-researcher\", \"agency-lsp-index-engineer\", \"agency-zhihu-strategist\", \"agency-test-results-analyzer\", \"agency-developer-advocate\", \"agency-data-consolidation-agent\", \"agency-feedback-synthesizer\", \"agency-social-media-strategist\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"agency-twitter-engager\", \"agency-project-shepherd\", \"agency-brand-guardian\", \"agency-finance-tracker\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0169",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Marketing Twitter Engager. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-twitter-engager",
        "agency-zhihu-strategist",
        "agency-xiaohongshu-specialist",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "agency-social-media-strategist",
        "agency-instagram-curator",
        "agency-visual-storyteller",
        "agency-growth-hacker",
        "agency-whimsy-injector",
        "agency-wechat-official-account-manager",
        "agency-developer-advocate",
        "agency-experiment-tracker",
        "agency-app-store-optimizer",
        "agency-content-creator",
        "agency-ai-engineer",
        "agency-workflow-optimizer",
        "agency-lsp-index-engineer",
        "agency-infrastructure-maintainer",
        "agency-support-responder",
        "agency-api-tester",
        "agency-legal-compliance-checker",
        "agency-test-results-analyzer",
        "agency-tool-evaluator",
        "agency-inclusive-visuals-specialist"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-zhihu-strategist",
          "tokens": 3239
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-social-media-strategist",
          "tokens": 2084
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-app-store-optimizer",
          "tokens": 3512
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-lsp-index-engineer",
          "tokens": 3006
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-support-responder",
          "tokens": 3296
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "agency-tool-evaluator",
          "tokens": 4549
        },
        {
          "skill_id": "agency-inclusive-visuals-specialist",
          "tokens": 1931
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Marketing Twitter Engager. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-twitter-engager\", \"agency-zhihu-strategist\", \"agency-xiaohongshu-specialist\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"agency-social-media-strategist\", \"agency-instagram-curator\", \"agency-visual-storyteller\", \"agency-growth-hacker\", \"agency-whimsy-injector\", \"agency-wechat-official-account-manager\", \"agency-developer-advocate\", \"agency-experiment-tracker\", \"agency-app-store-optimizer\", \"agency-content-creator\", \"agency-ai-engineer\", \"agency-workflow-optimizer\", \"agency-lsp-index-engineer\", \"agency-infrastructure-maintainer\", \"agency-support-responder\", \"agency-api-tester\", \"agency-legal-compliance-checker\", \"agency-test-results-analyzer\", \"agency-tool-evaluator\", \"agency-inclusive-visuals-specialist\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0170",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Issue-to-Code Draft Generator. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "issue-to-code",
        "pr-to-issue-linker",
        "email-auto-reply",
        "ship",
        "role-developer",
        "commit-to-issue",
        "ui-suite",
        "test-suite",
        "pipeline-builder",
        "release-ship",
        "github-sprint-digest",
        "sprint-orchestrator",
        "ralph-loop",
        "kwp-brand-voice-guideline-generation",
        "demo-forge",
        "release-commander",
        "incident-to-improvement",
        "related-papers-scout",
        "sefo-orchestrator",
        "ppt-template-engine",
        "docx-template-engine",
        "paper-auto-classifier",
        "diagnose",
        "pr-review-captain",
        "hf-auto-finetune-planner"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "pr-to-issue-linker",
          "tokens": 1223
        },
        {
          "skill_id": "email-auto-reply",
          "tokens": 638
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "pipeline-builder",
          "tokens": 2234
        },
        {
          "skill_id": "release-ship",
          "tokens": 3056
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "sprint-orchestrator",
          "tokens": 1320
        },
        {
          "skill_id": "ralph-loop",
          "tokens": 1925
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "demo-forge",
          "tokens": 2152
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "incident-to-improvement",
          "tokens": 2624
        },
        {
          "skill_id": "related-papers-scout",
          "tokens": 3560
        },
        {
          "skill_id": "sefo-orchestrator",
          "tokens": 821
        },
        {
          "skill_id": "ppt-template-engine",
          "tokens": 1101
        },
        {
          "skill_id": "docx-template-engine",
          "tokens": 1193
        },
        {
          "skill_id": "paper-auto-classifier",
          "tokens": 1262
        },
        {
          "skill_id": "diagnose",
          "tokens": 1433
        },
        {
          "skill_id": "pr-review-captain",
          "tokens": 1496
        },
        {
          "skill_id": "hf-auto-finetune-planner",
          "tokens": 2091
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Issue-to-Code Draft Generator. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"issue-to-code\", \"pr-to-issue-linker\", \"email-auto-reply\", \"ship\", \"role-developer\", \"commit-to-issue\", \"ui-suite\", \"test-suite\", \"pipeline-builder\", \"release-ship\", \"github-sprint-digest\", \"sprint-orchestrator\", \"ralph-loop\", \"kwp-brand-voice-guideline-generation\", \"demo-forge\", \"release-commander\", \"incident-to-improvement\", \"related-papers-scout\", \"sefo-orchestrator\", \"ppt-template-engine\", \"docx-template-engine\", \"paper-auto-classifier\", \"diagnose\", \"pr-review-captain\", \"hf-auto-finetune-planner\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0171",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Write internal communications using company formats. Use for 3P updates (Progress, Plans, Problems), company newsletters, FAQ responses, status reports, leadership updates, project updates, incident r…. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "anthropic-internal-comms",
        "agency-executive-summary-generator",
        "pm-go-to-market",
        "kwp-product-management-stakeholder-comms",
        "sales-meeting-pipeline",
        "pm-toolkit",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "agency-twitter-engager",
        "slack-agent",
        "agency-xiaohongshu-specialist",
        "agency-project-shepherd",
        "agency-developer-advocate",
        "kwp-engineering-testing-strategy",
        "gws-recipe-audit-sharing",
        "gws-docs",
        "kwp-productivity-task-management",
        "pm-marketing-growth",
        "anthropic-theme-factory",
        "weekly-status-report",
        "agency-instagram-curator",
        "kwp-product-management-roadmap-management",
        "agency-rapid-prototyper",
        "kwp-marketing-content-creation",
        "problem-definition"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "anthropic-internal-comms",
          "tokens": 522
        },
        {
          "skill_id": "agency-executive-summary-generator",
          "tokens": 2465
        },
        {
          "skill_id": "pm-go-to-market",
          "tokens": 992
        },
        {
          "skill_id": "kwp-product-management-stakeholder-comms",
          "tokens": 2952
        },
        {
          "skill_id": "sales-meeting-pipeline",
          "tokens": 559
        },
        {
          "skill_id": "pm-toolkit",
          "tokens": 935
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "slack-agent",
          "tokens": 2564
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "gws-recipe-audit-sharing",
          "tokens": 645
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "kwp-productivity-task-management",
          "tokens": 776
        },
        {
          "skill_id": "pm-marketing-growth",
          "tokens": 954
        },
        {
          "skill_id": "anthropic-theme-factory",
          "tokens": 990
        },
        {
          "skill_id": "weekly-status-report",
          "tokens": 1282
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "kwp-product-management-roadmap-management",
          "tokens": 2467
        },
        {
          "skill_id": "agency-rapid-prototyper",
          "tokens": 2496
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "problem-definition",
          "tokens": 2601
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Write internal communications using company formats. Use for 3P updates (Progress, Plans, Problems), company newsletters, FAQ responses, status reports, leadership updates, project updates, incident r…. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-internal-comms\", \"agency-executive-summary-generator\", \"pm-go-to-market\", \"kwp-product-management-stakeholder-comms\", \"sales-meeting-pipeline\", \"pm-toolkit\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"agency-twitter-engager\", \"slack-agent\", \"agency-xiaohongshu-specialist\", \"agency-project-shepherd\", \"agency-developer-advocate\", \"kwp-engineering-testing-strategy\", \"gws-recipe-audit-sharing\", \"gws-docs\", \"kwp-productivity-task-management\", \"pm-marketing-growth\", \"anthropic-theme-factory\", \"weekly-status-report\", \"agency-instagram-curator\", \"kwp-product-management-roadmap-management\", \"agency-rapid-prototyper\", \"kwp-marketing-content-creation\", \"problem-definition\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0172",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AlphaEar Stock. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "alphaear-stock",
        "weekly-stock-update",
        "stock-csv-downloader",
        "daily-stock-check",
        "alphaear-search",
        "alphaear-deepear-lite",
        "trading-sector-analyst",
        "tab-stock-sync",
        "trading-technical-analyst",
        "trading-us-stock-analysis",
        "alphaear-predictor",
        "alphaear-news",
        "trading-market-environment-analysis",
        "kwp-engineering-tech-debt",
        "hf-papers",
        "hf-spaces",
        "hf-models",
        "weekly-status-report",
        "commit-to-issue",
        "dependency-auditor",
        "kwp-sync",
        "trading-position-sizer",
        "agency-feedback-synthesizer",
        "cognee",
        "trading-stanley-druckenmiller-investment"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "alphaear-stock",
          "tokens": 820
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "alphaear-search",
          "tokens": 672
        },
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        },
        {
          "skill_id": "trading-sector-analyst",
          "tokens": 2657
        },
        {
          "skill_id": "tab-stock-sync",
          "tokens": 345
        },
        {
          "skill_id": "trading-technical-analyst",
          "tokens": 2868
        },
        {
          "skill_id": "trading-us-stock-analysis",
          "tokens": 2926
        },
        {
          "skill_id": "alphaear-predictor",
          "tokens": 803
        },
        {
          "skill_id": "alphaear-news",
          "tokens": 873
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "kwp-engineering-tech-debt",
          "tokens": 648
        },
        {
          "skill_id": "hf-papers",
          "tokens": 817
        },
        {
          "skill_id": "hf-spaces",
          "tokens": 930
        },
        {
          "skill_id": "hf-models",
          "tokens": 1212
        },
        {
          "skill_id": "weekly-status-report",
          "tokens": 1282
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "dependency-auditor",
          "tokens": 1458
        },
        {
          "skill_id": "kwp-sync",
          "tokens": 1472
        },
        {
          "skill_id": "trading-position-sizer",
          "tokens": 2013
        },
        {
          "skill_id": "agency-feedback-synthesizer",
          "tokens": 2041
        },
        {
          "skill_id": "cognee",
          "tokens": 2070
        },
        {
          "skill_id": "trading-stanley-druckenmiller-investment",
          "tokens": 2317
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AlphaEar Stock. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"alphaear-stock\", \"weekly-stock-update\", \"stock-csv-downloader\", \"daily-stock-check\", \"alphaear-search\", \"alphaear-deepear-lite\", \"trading-sector-analyst\", \"tab-stock-sync\", \"trading-technical-analyst\", \"trading-us-stock-analysis\", \"alphaear-predictor\", \"alphaear-news\", \"trading-market-environment-analysis\", \"kwp-engineering-tech-debt\", \"hf-papers\", \"hf-spaces\", \"hf-models\", \"weekly-status-report\", \"commit-to-issue\", \"dependency-auditor\", \"kwp-sync\", \"trading-position-sizer\", \"agency-feedback-synthesizer\", \"cognee\", \"trading-stanley-druckenmiller-investment\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0173",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Paper Auto-Classifier. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "paper-auto-classifier",
        "hf-papers",
        "unified-intel-intake",
        "related-papers-scout",
        "email-research-dispatcher",
        "knowledge-to-paper",
        "hf-trending-intelligence",
        "auto-research",
        "alphaxiv-paper-lookup",
        "bespin-news-digest",
        "alphaear-search",
        "pr-to-issue-linker",
        "auto-research-distribute",
        "cognee",
        "tech-support-brain",
        "github-sprint-digest",
        "hf-hub",
        "role-ceo",
        "standup-digest",
        "ui-suite",
        "issue-to-code",
        "deep-review",
        "deep-research-pipeline",
        "simplify",
        "daily-am-orchestrator"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "paper-auto-classifier",
          "tokens": 1262
        },
        {
          "skill_id": "hf-papers",
          "tokens": 817
        },
        {
          "skill_id": "unified-intel-intake",
          "tokens": 1593
        },
        {
          "skill_id": "related-papers-scout",
          "tokens": 3560
        },
        {
          "skill_id": "email-research-dispatcher",
          "tokens": 703
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "hf-trending-intelligence",
          "tokens": 1832
        },
        {
          "skill_id": "auto-research",
          "tokens": 2139
        },
        {
          "skill_id": "alphaxiv-paper-lookup",
          "tokens": 1299
        },
        {
          "skill_id": "bespin-news-digest",
          "tokens": 3949
        },
        {
          "skill_id": "alphaear-search",
          "tokens": 672
        },
        {
          "skill_id": "pr-to-issue-linker",
          "tokens": 1223
        },
        {
          "skill_id": "auto-research-distribute",
          "tokens": 1771
        },
        {
          "skill_id": "cognee",
          "tokens": 2070
        },
        {
          "skill_id": "tech-support-brain",
          "tokens": 620
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "hf-hub",
          "tokens": 1151
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "standup-digest",
          "tokens": 1330
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "deep-review",
          "tokens": 1923
        },
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        },
        {
          "skill_id": "simplify",
          "tokens": 2596
        },
        {
          "skill_id": "daily-am-orchestrator",
          "tokens": 3613
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Paper Auto-Classifier. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"paper-auto-classifier\", \"hf-papers\", \"unified-intel-intake\", \"related-papers-scout\", \"email-research-dispatcher\", \"knowledge-to-paper\", \"hf-trending-intelligence\", \"auto-research\", \"alphaxiv-paper-lookup\", \"bespin-news-digest\", \"alphaear-search\", \"pr-to-issue-linker\", \"auto-research-distribute\", \"cognee\", \"tech-support-brain\", \"github-sprint-digest\", \"hf-hub\", \"role-ceo\", \"standup-digest\", \"ui-suite\", \"issue-to-code\", \"deep-review\", \"deep-research-pipeline\", \"simplify\", \"daily-am-orchestrator\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0174",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Agentic Engineering. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "ecc-agentic-engineering",
        "sefo-orchestrator",
        "plans",
        "agency-visionos-spatial-engineer",
        "issue-to-code",
        "ecc-search-first",
        "agency-ai-engineer",
        "agency-lsp-index-engineer",
        "agency-ux-architect",
        "agency-security-engineer",
        "ecc-autonomous-loops",
        "kwp-operations-resource-planning",
        "agency-xr-cockpit-interaction-specialist",
        "kwp-productivity-task-management",
        "ecc-coding-standards",
        "sp-finish-branch",
        "ecc-strategic-compact",
        "workflow-sequential",
        "paperclip-tasks",
        "agency-senior-project-manager",
        "paperclip-setup",
        "air-gap-orchestrator",
        "test-suite",
        "paperclip-agents",
        "ralph-loop"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "ecc-agentic-engineering",
          "tokens": 690
        },
        {
          "skill_id": "sefo-orchestrator",
          "tokens": 821
        },
        {
          "skill_id": "plans",
          "tokens": 2297
        },
        {
          "skill_id": "agency-visionos-spatial-engineer",
          "tokens": 1091
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "ecc-search-first",
          "tokens": 1597
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-lsp-index-engineer",
          "tokens": 3006
        },
        {
          "skill_id": "agency-ux-architect",
          "tokens": 3119
        },
        {
          "skill_id": "agency-security-engineer",
          "tokens": 3168
        },
        {
          "skill_id": "ecc-autonomous-loops",
          "tokens": 4660
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "agency-xr-cockpit-interaction-specialist",
          "tokens": 693
        },
        {
          "skill_id": "kwp-productivity-task-management",
          "tokens": 776
        },
        {
          "skill_id": "ecc-coding-standards",
          "tokens": 1050
        },
        {
          "skill_id": "sp-finish-branch",
          "tokens": 1167
        },
        {
          "skill_id": "ecc-strategic-compact",
          "tokens": 1252
        },
        {
          "skill_id": "workflow-sequential",
          "tokens": 1485
        },
        {
          "skill_id": "paperclip-tasks",
          "tokens": 1534
        },
        {
          "skill_id": "agency-senior-project-manager",
          "tokens": 1555
        },
        {
          "skill_id": "paperclip-setup",
          "tokens": 1708
        },
        {
          "skill_id": "air-gap-orchestrator",
          "tokens": 1788
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "paperclip-agents",
          "tokens": 1887
        },
        {
          "skill_id": "ralph-loop",
          "tokens": 1925
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Agentic Engineering. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"ecc-agentic-engineering\", \"sefo-orchestrator\", \"plans\", \"agency-visionos-spatial-engineer\", \"issue-to-code\", \"ecc-search-first\", \"agency-ai-engineer\", \"agency-lsp-index-engineer\", \"agency-ux-architect\", \"agency-security-engineer\", \"ecc-autonomous-loops\", \"kwp-operations-resource-planning\", \"agency-xr-cockpit-interaction-specialist\", \"kwp-productivity-task-management\", \"ecc-coding-standards\", \"sp-finish-branch\", \"ecc-strategic-compact\", \"workflow-sequential\", \"paperclip-tasks\", \"agency-senior-project-manager\", \"paperclip-setup\", \"air-gap-orchestrator\", \"test-suite\", \"paperclip-agents\", \"ralph-loop\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0175",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: KWP Sync. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-sync",
        "prompt-transformer",
        "skill-guide-generator",
        "skill-optimizer",
        "autoskill-extractor",
        "skill-seekers",
        "ecc-configure",
        "anthropic-skill-creator",
        "autoskill-merger",
        "agency-roster",
        "skill-composer",
        "autoskill-judge",
        "commit-to-issue",
        "weekly-stock-update",
        "ui-suite",
        "autoskill-evolve",
        "alphaear-stock",
        "ppt-template-engine",
        "i18n-sync",
        "stock-csv-downloader",
        "anthropic-pdf",
        "simplify",
        "kwp-cowork-create-cowork-plugin",
        "kwp-cowork-create-plugin",
        "cursor-sync"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-sync",
          "tokens": 1472
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "skill-guide-generator",
          "tokens": 1858
        },
        {
          "skill_id": "skill-optimizer",
          "tokens": 2940
        },
        {
          "skill_id": "autoskill-extractor",
          "tokens": 1235
        },
        {
          "skill_id": "skill-seekers",
          "tokens": 1568
        },
        {
          "skill_id": "ecc-configure",
          "tokens": 3534
        },
        {
          "skill_id": "anthropic-skill-creator",
          "tokens": 6862
        },
        {
          "skill_id": "autoskill-merger",
          "tokens": 1142
        },
        {
          "skill_id": "agency-roster",
          "tokens": 3411
        },
        {
          "skill_id": "skill-composer",
          "tokens": 599
        },
        {
          "skill_id": "autoskill-judge",
          "tokens": 1176
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "autoskill-evolve",
          "tokens": 1566
        },
        {
          "skill_id": "alphaear-stock",
          "tokens": 820
        },
        {
          "skill_id": "ppt-template-engine",
          "tokens": 1101
        },
        {
          "skill_id": "i18n-sync",
          "tokens": 1282
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "anthropic-pdf",
          "tokens": 2219
        },
        {
          "skill_id": "simplify",
          "tokens": 2596
        },
        {
          "skill_id": "kwp-cowork-create-cowork-plugin",
          "tokens": 3017
        },
        {
          "skill_id": "kwp-cowork-create-plugin",
          "tokens": 3214
        },
        {
          "skill_id": "cursor-sync",
          "tokens": 3267
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: KWP Sync. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-sync\", \"prompt-transformer\", \"skill-guide-generator\", \"skill-optimizer\", \"autoskill-extractor\", \"skill-seekers\", \"ecc-configure\", \"anthropic-skill-creator\", \"autoskill-merger\", \"agency-roster\", \"skill-composer\", \"autoskill-judge\", \"commit-to-issue\", \"weekly-stock-update\", \"ui-suite\", \"autoskill-evolve\", \"alphaear-stock\", \"ppt-template-engine\", \"i18n-sync\", \"stock-csv-downloader\", \"anthropic-pdf\", \"simplify\", \"kwp-cowork-create-cowork-plugin\", \"kwp-cowork-create-plugin\", \"cursor-sync\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0176",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Notion 문서 동기화. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "notion-docs-sync",
        "github-workflow-automation",
        "planning-weekly-pulse",
        "sales-meeting-pipeline",
        "planning-meeting-sentinel",
        "swagger-api-doc-generator",
        "dependency-radar",
        "lead-lifecycle-manager",
        "docs-freshness-guardian",
        "sales-weekly-dashboard",
        "feature-intake-router",
        "proposal-factory",
        "incident-lifecycle-manager",
        "tech-support-brain",
        "github-sprint-digest",
        "design-review",
        "figma-to-tds",
        "implement-screen",
        "fsd-development",
        "weekly-status-report",
        "jira-linear-adapter",
        "commit-to-issue",
        "kwp-sync",
        "md-enhance-publish",
        "auto-research-distribute"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "notion-docs-sync",
          "tokens": 857
        },
        {
          "skill_id": "github-workflow-automation",
          "tokens": 971
        },
        {
          "skill_id": "planning-weekly-pulse",
          "tokens": 528
        },
        {
          "skill_id": "sales-meeting-pipeline",
          "tokens": 559
        },
        {
          "skill_id": "planning-meeting-sentinel",
          "tokens": 604
        },
        {
          "skill_id": "swagger-api-doc-generator",
          "tokens": 784
        },
        {
          "skill_id": "dependency-radar",
          "tokens": 531
        },
        {
          "skill_id": "lead-lifecycle-manager",
          "tokens": 533
        },
        {
          "skill_id": "docs-freshness-guardian",
          "tokens": 557
        },
        {
          "skill_id": "sales-weekly-dashboard",
          "tokens": 558
        },
        {
          "skill_id": "feature-intake-router",
          "tokens": 568
        },
        {
          "skill_id": "proposal-factory",
          "tokens": 582
        },
        {
          "skill_id": "incident-lifecycle-manager",
          "tokens": 610
        },
        {
          "skill_id": "tech-support-brain",
          "tokens": 620
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "design-review",
          "tokens": 947
        },
        {
          "skill_id": "figma-to-tds",
          "tokens": 987
        },
        {
          "skill_id": "implement-screen",
          "tokens": 996
        },
        {
          "skill_id": "fsd-development",
          "tokens": 1234
        },
        {
          "skill_id": "weekly-status-report",
          "tokens": 1282
        },
        {
          "skill_id": "jira-linear-adapter",
          "tokens": 1365
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "kwp-sync",
          "tokens": 1472
        },
        {
          "skill_id": "md-enhance-publish",
          "tokens": 1740
        },
        {
          "skill_id": "auto-research-distribute",
          "tokens": 1771
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Notion 문서 동기화. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"notion-docs-sync\", \"github-workflow-automation\", \"planning-weekly-pulse\", \"sales-meeting-pipeline\", \"planning-meeting-sentinel\", \"swagger-api-doc-generator\", \"dependency-radar\", \"lead-lifecycle-manager\", \"docs-freshness-guardian\", \"sales-weekly-dashboard\", \"feature-intake-router\", \"proposal-factory\", \"incident-lifecycle-manager\", \"tech-support-brain\", \"github-sprint-digest\", \"design-review\", \"figma-to-tds\", \"implement-screen\", \"fsd-development\", \"weekly-status-report\", \"jira-linear-adapter\", \"commit-to-issue\", \"kwp-sync\", \"md-enhance-publish\", \"auto-research-distribute\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0177",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Plans. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "plans",
        "full-stack-planner",
        "sefo-orchestrator",
        "test-suite",
        "rtk",
        "design-architect",
        "ralph-loop",
        "prompt-architect",
        "pipeline-builder",
        "release-commander",
        "mission-control",
        "problem-definition",
        "release-ship",
        "skill-autoimprove",
        "prd-research-factory",
        "pm-execution",
        "video-compress",
        "workflow-sequential",
        "ui-suite",
        "issue-to-code",
        "scrapling",
        "prompt-transformer",
        "air-gap-orchestrator",
        "mission-control copy",
        "trading-edge-candidate-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "plans",
          "tokens": 2297
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        },
        {
          "skill_id": "sefo-orchestrator",
          "tokens": 821
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "rtk",
          "tokens": 1214
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "ralph-loop",
          "tokens": 1925
        },
        {
          "skill_id": "prompt-architect",
          "tokens": 1939
        },
        {
          "skill_id": "pipeline-builder",
          "tokens": 2234
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "problem-definition",
          "tokens": 2601
        },
        {
          "skill_id": "release-ship",
          "tokens": 3056
        },
        {
          "skill_id": "skill-autoimprove",
          "tokens": 3666
        },
        {
          "skill_id": "prd-research-factory",
          "tokens": 648
        },
        {
          "skill_id": "pm-execution",
          "tokens": 1178
        },
        {
          "skill_id": "video-compress",
          "tokens": 1230
        },
        {
          "skill_id": "workflow-sequential",
          "tokens": 1485
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "scrapling",
          "tokens": 1675
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "air-gap-orchestrator",
          "tokens": 1788
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "trading-edge-candidate-agent",
          "tokens": 1971
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Plans. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"plans\", \"full-stack-planner\", \"sefo-orchestrator\", \"test-suite\", \"rtk\", \"design-architect\", \"ralph-loop\", \"prompt-architect\", \"pipeline-builder\", \"release-commander\", \"mission-control\", \"problem-definition\", \"release-ship\", \"skill-autoimprove\", \"prd-research-factory\", \"pm-execution\", \"video-compress\", \"workflow-sequential\", \"ui-suite\", \"issue-to-code\", \"scrapling\", \"prompt-transformer\", \"air-gap-orchestrator\", \"mission-control copy\", \"trading-edge-candidate-agent\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0178",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PM Go-to-Market. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "pm-go-to-market",
        "pm-marketing-growth",
        "pm-market-research",
        "pm-toolkit",
        "pm-product-discovery",
        "pm-product-strategy",
        "pm-execution",
        "nlm-launch-kit",
        "role-cso",
        "pm-data-analytics",
        "agency-trend-researcher",
        "role-sales",
        "agency-tiktok-strategist",
        "agency-xiaohongshu-specialist",
        "kwp-marketing-competitive-analysis",
        "kwp-marketing-campaign-planning",
        "mission-control copy",
        "agency-reddit-community-builder",
        "agency-twitter-engager",
        "mission-control",
        "agency-executive-summary-generator",
        "kwp-engineering-testing-strategy",
        "prd-research-factory",
        "gws-recipe-doc-from-template",
        "kwp-human-resources-interview-prep"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "pm-go-to-market",
          "tokens": 992
        },
        {
          "skill_id": "pm-marketing-growth",
          "tokens": 954
        },
        {
          "skill_id": "pm-market-research",
          "tokens": 966
        },
        {
          "skill_id": "pm-toolkit",
          "tokens": 935
        },
        {
          "skill_id": "pm-product-discovery",
          "tokens": 1198
        },
        {
          "skill_id": "pm-product-strategy",
          "tokens": 1271
        },
        {
          "skill_id": "pm-execution",
          "tokens": 1178
        },
        {
          "skill_id": "nlm-launch-kit",
          "tokens": 2368
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        },
        {
          "skill_id": "pm-data-analytics",
          "tokens": 956
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "role-sales",
          "tokens": 1221
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "kwp-marketing-competitive-analysis",
          "tokens": 2723
        },
        {
          "skill_id": "kwp-marketing-campaign-planning",
          "tokens": 2803
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "agency-executive-summary-generator",
          "tokens": 2465
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "prd-research-factory",
          "tokens": 648
        },
        {
          "skill_id": "gws-recipe-doc-from-template",
          "tokens": 743
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: PM Go-to-Market. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"pm-go-to-market\", \"pm-marketing-growth\", \"pm-market-research\", \"pm-toolkit\", \"pm-product-discovery\", \"pm-product-strategy\", \"pm-execution\", \"nlm-launch-kit\", \"role-cso\", \"pm-data-analytics\", \"agency-trend-researcher\", \"role-sales\", \"agency-tiktok-strategist\", \"agency-xiaohongshu-specialist\", \"kwp-marketing-competitive-analysis\", \"kwp-marketing-campaign-planning\", \"mission-control copy\", \"agency-reddit-community-builder\", \"agency-twitter-engager\", \"mission-control\", \"agency-executive-summary-generator\", \"kwp-engineering-testing-strategy\", \"prd-research-factory\", \"gws-recipe-doc-from-template\", \"kwp-human-resources-interview-prep\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0179",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Account Research. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-sales-account-research",
        "kwp-sales-call-prep",
        "kwp-sales-daily-briefing",
        "kwp-sales-draft-outreach",
        "kwp-apollo-enrich-lead",
        "kwp-sales-competitive-intelligence",
        "kwp-human-resources-org-planning",
        "kwp-sales-create-an-asset",
        "kwp-operations-resource-planning",
        "kwp-common-room-contact-research",
        "kwp-common-room-account-research",
        "kwp-human-resources-people-analytics",
        "kwp-operations-vendor-management",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-operations-process-optimization",
        "kwp-operations-change-management",
        "kwp-human-resources-interview-prep",
        "kwp-apollo-prospect",
        "kwp-legal-meeting-briefing",
        "kwp-customer-support-customer-research",
        "kwp-legal-contract-review",
        "lead-lifecycle-manager",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-human-resources-employee-handbook",
        "kwp-design-design-handoff"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "kwp-sales-draft-outreach",
          "tokens": 2782
        },
        {
          "skill_id": "kwp-apollo-enrich-lead",
          "tokens": 837
        },
        {
          "skill_id": "kwp-sales-competitive-intelligence",
          "tokens": 3361
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-sales-create-an-asset",
          "tokens": 4079
        },
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-common-room-contact-research",
          "tokens": 1611
        },
        {
          "skill_id": "kwp-common-room-account-research",
          "tokens": 1784
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-recruiting-pipeline",
          "tokens": 578
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "kwp-apollo-prospect",
          "tokens": 1006
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-customer-support-customer-research",
          "tokens": 2529
        },
        {
          "skill_id": "kwp-legal-contract-review",
          "tokens": 3137
        },
        {
          "skill_id": "lead-lifecycle-manager",
          "tokens": 533
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-human-resources-employee-handbook",
          "tokens": 629
        },
        {
          "skill_id": "kwp-design-design-handoff",
          "tokens": 649
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Account Research. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-sales-account-research\", \"kwp-sales-call-prep\", \"kwp-sales-daily-briefing\", \"kwp-sales-draft-outreach\", \"kwp-apollo-enrich-lead\", \"kwp-sales-competitive-intelligence\", \"kwp-human-resources-org-planning\", \"kwp-sales-create-an-asset\", \"kwp-operations-resource-planning\", \"kwp-common-room-contact-research\", \"kwp-common-room-account-research\", \"kwp-human-resources-people-analytics\", \"kwp-operations-vendor-management\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-operations-process-optimization\", \"kwp-operations-change-management\", \"kwp-human-resources-interview-prep\", \"kwp-apollo-prospect\", \"kwp-legal-meeting-briefing\", \"kwp-customer-support-customer-research\", \"kwp-legal-contract-review\", \"lead-lifecycle-manager\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-human-resources-employee-handbook\", \"kwp-design-design-handoff\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0180",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Resource Planning. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-operations-resource-planning",
        "kwp-human-resources-org-planning",
        "kwp-operations-change-management",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-marketing-campaign-planning",
        "kwp-operations-vendor-management",
        "kwp-engineering-testing-strategy",
        "kwp-human-resources-interview-prep",
        "kwp-operations-process-optimization",
        "kwp-engineering-system-design",
        "kwp-human-resources-people-analytics",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-design-ux-writing",
        "kwp-design-user-research",
        "kwp-brand-voice-discover-brand",
        "kwp-sales-daily-briefing",
        "kwp-sales-account-research",
        "kwp-sales-call-prep",
        "kwp-marketing-performance-analytics",
        "kwp-engineering-tech-debt",
        "kwp-design-design-handoff",
        "kwp-engineering-documentation",
        "kwp-operations-compliance-tracking",
        "portfolio-report-generator",
        "kwp-customer-support-escalation"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-operations-resource-planning",
          "tokens": 649
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-marketing-campaign-planning",
          "tokens": 2803
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-engineering-testing-strategy",
          "tokens": 608
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "kwp-engineering-system-design",
          "tokens": 756
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "kwp-human-resources-recruiting-pipeline",
          "tokens": 578
        },
        {
          "skill_id": "kwp-design-ux-writing",
          "tokens": 675
        },
        {
          "skill_id": "kwp-design-user-research",
          "tokens": 718
        },
        {
          "skill_id": "kwp-brand-voice-discover-brand",
          "tokens": 1897
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-marketing-performance-analytics",
          "tokens": 3966
        },
        {
          "skill_id": "kwp-engineering-tech-debt",
          "tokens": 648
        },
        {
          "skill_id": "kwp-design-design-handoff",
          "tokens": 649
        },
        {
          "skill_id": "kwp-engineering-documentation",
          "tokens": 654
        },
        {
          "skill_id": "kwp-operations-compliance-tracking",
          "tokens": 686
        },
        {
          "skill_id": "portfolio-report-generator",
          "tokens": 1406
        },
        {
          "skill_id": "kwp-customer-support-escalation",
          "tokens": 2329
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Resource Planning. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-operations-resource-planning\", \"kwp-human-resources-org-planning\", \"kwp-operations-change-management\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-marketing-campaign-planning\", \"kwp-operations-vendor-management\", \"kwp-engineering-testing-strategy\", \"kwp-human-resources-interview-prep\", \"kwp-operations-process-optimization\", \"kwp-engineering-system-design\", \"kwp-human-resources-people-analytics\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-design-ux-writing\", \"kwp-design-user-research\", \"kwp-brand-voice-discover-brand\", \"kwp-sales-daily-briefing\", \"kwp-sales-account-research\", \"kwp-sales-call-prep\", \"kwp-marketing-performance-analytics\", \"kwp-engineering-tech-debt\", \"kwp-design-design-handoff\", \"kwp-engineering-documentation\", \"kwp-operations-compliance-tracking\", \"portfolio-report-generator\", \"kwp-customer-support-escalation\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0181",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: SEFO Federation. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "sefo-federation",
        "sefo-governance",
        "sefo-orchestrator",
        "jira-linear-adapter",
        "paperclip-control",
        "release-commander",
        "hf-hub",
        "paperclip-setup",
        "cursor-automations",
        "gws-workflows",
        "pr-to-issue-linker",
        "hf-buckets",
        "portfolio-report-generator",
        "paperclip-tasks",
        "air-gap-orchestrator",
        "mission-control copy",
        "paperclip-agents",
        "public-apis",
        "mission-control",
        "agency-project-shepherd",
        "cursor-sync",
        "intent-alignment-tracker",
        "kwp-design-design-system-management",
        "domain-commit",
        "ai-chief-of-staff"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "sefo-federation",
          "tokens": 756
        },
        {
          "skill_id": "sefo-governance",
          "tokens": 837
        },
        {
          "skill_id": "sefo-orchestrator",
          "tokens": 821
        },
        {
          "skill_id": "jira-linear-adapter",
          "tokens": 1365
        },
        {
          "skill_id": "paperclip-control",
          "tokens": 1407
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "hf-hub",
          "tokens": 1151
        },
        {
          "skill_id": "paperclip-setup",
          "tokens": 1708
        },
        {
          "skill_id": "cursor-automations",
          "tokens": 2911
        },
        {
          "skill_id": "gws-workflows",
          "tokens": 952
        },
        {
          "skill_id": "pr-to-issue-linker",
          "tokens": 1223
        },
        {
          "skill_id": "hf-buckets",
          "tokens": 1334
        },
        {
          "skill_id": "portfolio-report-generator",
          "tokens": 1406
        },
        {
          "skill_id": "paperclip-tasks",
          "tokens": 1534
        },
        {
          "skill_id": "air-gap-orchestrator",
          "tokens": 1788
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "paperclip-agents",
          "tokens": 1887
        },
        {
          "skill_id": "public-apis",
          "tokens": 2246
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "cursor-sync",
          "tokens": 3267
        },
        {
          "skill_id": "intent-alignment-tracker",
          "tokens": 558
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "domain-commit",
          "tokens": 915
        },
        {
          "skill_id": "ai-chief-of-staff",
          "tokens": 1152
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: SEFO Federation. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"sefo-federation\", \"sefo-governance\", \"sefo-orchestrator\", \"jira-linear-adapter\", \"paperclip-control\", \"release-commander\", \"hf-hub\", \"paperclip-setup\", \"cursor-automations\", \"gws-workflows\", \"pr-to-issue-linker\", \"hf-buckets\", \"portfolio-report-generator\", \"paperclip-tasks\", \"air-gap-orchestrator\", \"mission-control copy\", \"paperclip-agents\", \"public-apis\", \"mission-control\", \"agency-project-shepherd\", \"cursor-sync\", \"intent-alignment-tracker\", \"kwp-design-design-system-management\", \"domain-commit\", \"ai-chief-of-staff\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0182",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Competitive Analysis Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-marketing-competitive-analysis",
        "kwp-product-management-competitive-analysis",
        "kwp-marketing-performance-analytics",
        "kwp-marketing-campaign-planning",
        "kwp-sales-competitive-intelligence",
        "pm-go-to-market",
        "kwp-marketing-brand-voice",
        "role-cso",
        "kwp-finance-reconciliation",
        "kwp-legal-meeting-briefing",
        "kwp-customer-support-customer-research",
        "kwp-marketing-content-creation",
        "kwp-product-management-user-research-synthesis",
        "kwp-finance-variance-analysis",
        "kwp-product-management-metrics-tracking",
        "kwp-operations-vendor-management",
        "kwp-human-resources-compensation-benchmarking",
        "kwp-human-resources-interview-prep",
        "pm-marketing-growth",
        "kwp-sales-account-research",
        "kwp-sales-call-prep",
        "kwp-data-interactive-dashboard-builder",
        "kwp-data-data-exploration",
        "kwp-product-management-feature-spec",
        "kwp-data-data-validation"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-marketing-competitive-analysis",
          "tokens": 2723
        },
        {
          "skill_id": "kwp-product-management-competitive-analysis",
          "tokens": 2899
        },
        {
          "skill_id": "kwp-marketing-performance-analytics",
          "tokens": 3966
        },
        {
          "skill_id": "kwp-marketing-campaign-planning",
          "tokens": 2803
        },
        {
          "skill_id": "kwp-sales-competitive-intelligence",
          "tokens": 3361
        },
        {
          "skill_id": "pm-go-to-market",
          "tokens": 992
        },
        {
          "skill_id": "kwp-marketing-brand-voice",
          "tokens": 2392
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        },
        {
          "skill_id": "kwp-finance-reconciliation",
          "tokens": 2287
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-customer-support-customer-research",
          "tokens": 2529
        },
        {
          "skill_id": "kwp-marketing-content-creation",
          "tokens": 2596
        },
        {
          "skill_id": "kwp-product-management-user-research-synthesis",
          "tokens": 3071
        },
        {
          "skill_id": "kwp-finance-variance-analysis",
          "tokens": 3208
        },
        {
          "skill_id": "kwp-product-management-metrics-tracking",
          "tokens": 3606
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-compensation-benchmarking",
          "tokens": 585
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "pm-marketing-growth",
          "tokens": 954
        },
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-data-interactive-dashboard-builder",
          "tokens": 2350
        },
        {
          "skill_id": "kwp-data-data-exploration",
          "tokens": 2358
        },
        {
          "skill_id": "kwp-product-management-feature-spec",
          "tokens": 2614
        },
        {
          "skill_id": "kwp-data-data-validation",
          "tokens": 2650
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Competitive Analysis Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-marketing-competitive-analysis\", \"kwp-product-management-competitive-analysis\", \"kwp-marketing-performance-analytics\", \"kwp-marketing-campaign-planning\", \"kwp-sales-competitive-intelligence\", \"pm-go-to-market\", \"kwp-marketing-brand-voice\", \"role-cso\", \"kwp-finance-reconciliation\", \"kwp-legal-meeting-briefing\", \"kwp-customer-support-customer-research\", \"kwp-marketing-content-creation\", \"kwp-product-management-user-research-synthesis\", \"kwp-finance-variance-analysis\", \"kwp-product-management-metrics-tracking\", \"kwp-operations-vendor-management\", \"kwp-human-resources-compensation-benchmarking\", \"kwp-human-resources-interview-prep\", \"pm-marketing-growth\", \"kwp-sales-account-research\", \"kwp-sales-call-prep\", \"kwp-data-interactive-dashboard-builder\", \"kwp-data-data-exploration\", \"kwp-product-management-feature-spec\", \"kwp-data-data-validation\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0183",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Create distinctive, production-grade frontend interfaces with high design quality. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "anthropic-frontend-design",
        "anthropic-web-artifacts-builder",
        "ux-expert",
        "frontend-expert",
        "ui-suite",
        "design-architect",
        "kwp-design-design-system-management",
        "ecc-coding-standards",
        "agency-visual-storyteller",
        "agency-ai-engineer",
        "kwp-data-interactive-dashboard-builder",
        "agency-accessibility-auditor",
        "kwp-engineering-tech-debt",
        "agency-data-consolidation-agent",
        "ci-quality-gate",
        "backend-expert",
        "sre-devops-expert",
        "ship",
        "prompt-transformer",
        "performance-profiler",
        "deep-review",
        "agency-whimsy-injector",
        "code-review-all",
        "agency-frontend-developer",
        "agency-studio-producer"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "anthropic-frontend-design",
          "tokens": 1312
        },
        {
          "skill_id": "anthropic-web-artifacts-builder",
          "tokens": 968
        },
        {
          "skill_id": "ux-expert",
          "tokens": 1210
        },
        {
          "skill_id": "frontend-expert",
          "tokens": 1234
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "kwp-design-design-system-management",
          "tokens": 652
        },
        {
          "skill_id": "ecc-coding-standards",
          "tokens": 1050
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "kwp-data-interactive-dashboard-builder",
          "tokens": 2350
        },
        {
          "skill_id": "agency-accessibility-auditor",
          "tokens": 4031
        },
        {
          "skill_id": "kwp-engineering-tech-debt",
          "tokens": 648
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        },
        {
          "skill_id": "ci-quality-gate",
          "tokens": 1054
        },
        {
          "skill_id": "backend-expert",
          "tokens": 1132
        },
        {
          "skill_id": "sre-devops-expert",
          "tokens": 1450
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "performance-profiler",
          "tokens": 1738
        },
        {
          "skill_id": "deep-review",
          "tokens": 1923
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "code-review-all",
          "tokens": 2524
        },
        {
          "skill_id": "agency-frontend-developer",
          "tokens": 2533
        },
        {
          "skill_id": "agency-studio-producer",
          "tokens": 2953
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Create distinctive, production-grade frontend interfaces with high design quality. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"anthropic-frontend-design\", \"anthropic-web-artifacts-builder\", \"ux-expert\", \"frontend-expert\", \"ui-suite\", \"design-architect\", \"kwp-design-design-system-management\", \"ecc-coding-standards\", \"agency-visual-storyteller\", \"agency-ai-engineer\", \"kwp-data-interactive-dashboard-builder\", \"agency-accessibility-auditor\", \"kwp-engineering-tech-debt\", \"agency-data-consolidation-agent\", \"ci-quality-gate\", \"backend-expert\", \"sre-devops-expert\", \"ship\", \"prompt-transformer\", \"performance-profiler\", \"deep-review\", \"agency-whimsy-injector\", \"code-review-all\", \"agency-frontend-developer\", \"agency-studio-producer\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0184",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Weekly Status Report Generator. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "weekly-status-report",
        "standup-digest",
        "portfolio-report-generator",
        "github-sprint-digest",
        "sprint-orchestrator",
        "hf-trending-intelligence",
        "notion-meeting-sync",
        "gws-workflows",
        "stock-csv-downloader",
        "knowledge-daily-aggregator",
        "daily-stock-check",
        "deep-research-pipeline",
        "ai-workflow-integrator",
        "bespin-news-digest",
        "tab-report-generate",
        "daily-strategy-post",
        "trading-trade-hypothesis-ideator",
        "jira-linear-adapter",
        "executive-briefing",
        "weekly-stock-update",
        "morning",
        "ai-quality-evaluator",
        "pipeline-builder",
        "morning-ship",
        "planning-weekly-pulse"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "weekly-status-report",
          "tokens": 1282
        },
        {
          "skill_id": "standup-digest",
          "tokens": 1330
        },
        {
          "skill_id": "portfolio-report-generator",
          "tokens": 1406
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "sprint-orchestrator",
          "tokens": 1320
        },
        {
          "skill_id": "hf-trending-intelligence",
          "tokens": 1832
        },
        {
          "skill_id": "notion-meeting-sync",
          "tokens": 4347
        },
        {
          "skill_id": "gws-workflows",
          "tokens": 952
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "knowledge-daily-aggregator",
          "tokens": 1425
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        },
        {
          "skill_id": "ai-workflow-integrator",
          "tokens": 2227
        },
        {
          "skill_id": "bespin-news-digest",
          "tokens": 3949
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "trading-trade-hypothesis-ideator",
          "tokens": 909
        },
        {
          "skill_id": "jira-linear-adapter",
          "tokens": 1365
        },
        {
          "skill_id": "executive-briefing",
          "tokens": 1419
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "morning",
          "tokens": 2088
        },
        {
          "skill_id": "ai-quality-evaluator",
          "tokens": 2228
        },
        {
          "skill_id": "pipeline-builder",
          "tokens": 2234
        },
        {
          "skill_id": "morning-ship",
          "tokens": 2943
        },
        {
          "skill_id": "planning-weekly-pulse",
          "tokens": 528
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Weekly Status Report Generator. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"weekly-status-report\", \"standup-digest\", \"portfolio-report-generator\", \"github-sprint-digest\", \"sprint-orchestrator\", \"hf-trending-intelligence\", \"notion-meeting-sync\", \"gws-workflows\", \"stock-csv-downloader\", \"knowledge-daily-aggregator\", \"daily-stock-check\", \"deep-research-pipeline\", \"ai-workflow-integrator\", \"bespin-news-digest\", \"tab-report-generate\", \"daily-strategy-post\", \"trading-trade-hypothesis-ideator\", \"jira-linear-adapter\", \"executive-briefing\", \"weekly-stock-update\", \"morning\", \"ai-quality-evaluator\", \"pipeline-builder\", \"morning-ship\", \"planning-weekly-pulse\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0185",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Prospecting. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-common-room-prospect",
        "kwp-common-room-compose-outreach",
        "kwp-common-room-call-prep",
        "kwp-common-room-account-research",
        "kwp-common-room-contact-research",
        "kwp-common-room-weekly-prep-brief",
        "kwp-data-interactive-dashboard-builder",
        "kwp-apollo-enrich-lead",
        "kwp-sales-account-research",
        "kwp-sales-call-prep",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-customer-support-customer-research",
        "kwp-sales-draft-outreach",
        "kwp-data-data-visualization",
        "kwp-legal-legal-risk-assessment",
        "kwp-human-resources-recruiting-pipeline",
        "kwp-operations-change-management",
        "kwp-apollo-sequence-load",
        "kwp-slack-slack-search",
        "mission-control copy",
        "kwp-data-data-context-extractor",
        "kwp-brand-voice-discover-brand",
        "kwp-enterprise-search-search-strategy",
        "kwp-finance-journal-entry-prep",
        "mission-control"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-common-room-prospect",
          "tokens": 1803
        },
        {
          "skill_id": "kwp-common-room-compose-outreach",
          "tokens": 1627
        },
        {
          "skill_id": "kwp-common-room-call-prep",
          "tokens": 1729
        },
        {
          "skill_id": "kwp-common-room-account-research",
          "tokens": 1784
        },
        {
          "skill_id": "kwp-common-room-contact-research",
          "tokens": 1611
        },
        {
          "skill_id": "kwp-common-room-weekly-prep-brief",
          "tokens": 1607
        },
        {
          "skill_id": "kwp-data-interactive-dashboard-builder",
          "tokens": 2350
        },
        {
          "skill_id": "kwp-apollo-enrich-lead",
          "tokens": 837
        },
        {
          "skill_id": "kwp-sales-account-research",
          "tokens": 2104
        },
        {
          "skill_id": "kwp-sales-call-prep",
          "tokens": 2230
        },
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "kwp-customer-support-customer-research",
          "tokens": 2529
        },
        {
          "skill_id": "kwp-sales-draft-outreach",
          "tokens": 2782
        },
        {
          "skill_id": "kwp-data-data-visualization",
          "tokens": 3020
        },
        {
          "skill_id": "kwp-legal-legal-risk-assessment",
          "tokens": 3406
        },
        {
          "skill_id": "kwp-human-resources-recruiting-pipeline",
          "tokens": 578
        },
        {
          "skill_id": "kwp-operations-change-management",
          "tokens": 622
        },
        {
          "skill_id": "kwp-apollo-sequence-load",
          "tokens": 1107
        },
        {
          "skill_id": "kwp-slack-slack-search",
          "tokens": 1186
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "kwp-data-data-context-extractor",
          "tokens": 1883
        },
        {
          "skill_id": "kwp-brand-voice-discover-brand",
          "tokens": 1897
        },
        {
          "skill_id": "kwp-enterprise-search-search-strategy",
          "tokens": 2009
        },
        {
          "skill_id": "kwp-finance-journal-entry-prep",
          "tokens": 2215
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Prospecting. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-common-room-prospect\", \"kwp-common-room-compose-outreach\", \"kwp-common-room-call-prep\", \"kwp-common-room-account-research\", \"kwp-common-room-contact-research\", \"kwp-common-room-weekly-prep-brief\", \"kwp-data-interactive-dashboard-builder\", \"kwp-apollo-enrich-lead\", \"kwp-sales-account-research\", \"kwp-sales-call-prep\", \"kwp-bio-research-single-cell-rna-qc\", \"kwp-customer-support-customer-research\", \"kwp-sales-draft-outreach\", \"kwp-data-data-visualization\", \"kwp-legal-legal-risk-assessment\", \"kwp-human-resources-recruiting-pipeline\", \"kwp-operations-change-management\", \"kwp-apollo-sequence-load\", \"kwp-slack-slack-search\", \"mission-control copy\", \"kwp-data-data-context-extractor\", \"kwp-brand-voice-discover-brand\", \"kwp-enterprise-search-search-strategy\", \"kwp-finance-journal-entry-prep\", \"mission-control\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0186",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Release Ship. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "release-ship",
        "release-commander",
        "domain-commit",
        "commit-to-issue",
        "ship",
        "pr-to-issue-linker",
        "ui-suite",
        "sod-ship",
        "issue-to-code",
        "test-suite",
        "github-sprint-digest",
        "iac-review-agent",
        "sprint-orchestrator",
        "deep-review",
        "ralph-loop",
        "mission-control",
        "eod-ship",
        "critical-review",
        "autoskill-pipeline",
        "mission-control copy",
        "demo-forge",
        "plans",
        "simplify",
        "morning-ship",
        "semantic-guard"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "release-ship",
          "tokens": 3056
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "domain-commit",
          "tokens": 915
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "pr-to-issue-linker",
          "tokens": 1223
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "sod-ship",
          "tokens": 4151
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "test-suite",
          "tokens": 1869
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "iac-review-agent",
          "tokens": 1312
        },
        {
          "skill_id": "sprint-orchestrator",
          "tokens": 1320
        },
        {
          "skill_id": "deep-review",
          "tokens": 1923
        },
        {
          "skill_id": "ralph-loop",
          "tokens": 1925
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "eod-ship",
          "tokens": 2420
        },
        {
          "skill_id": "critical-review",
          "tokens": 3039
        },
        {
          "skill_id": "autoskill-pipeline",
          "tokens": 1508
        },
        {
          "skill_id": "mission-control copy",
          "tokens": 1882
        },
        {
          "skill_id": "demo-forge",
          "tokens": 2152
        },
        {
          "skill_id": "plans",
          "tokens": 2297
        },
        {
          "skill_id": "simplify",
          "tokens": 2596
        },
        {
          "skill_id": "morning-ship",
          "tokens": 2943
        },
        {
          "skill_id": "semantic-guard",
          "tokens": 571
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Release Ship. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"release-ship\", \"release-commander\", \"domain-commit\", \"commit-to-issue\", \"ship\", \"pr-to-issue-linker\", \"ui-suite\", \"sod-ship\", \"issue-to-code\", \"test-suite\", \"github-sprint-digest\", \"iac-review-agent\", \"sprint-orchestrator\", \"deep-review\", \"ralph-loop\", \"mission-control\", \"eod-ship\", \"critical-review\", \"autoskill-pipeline\", \"mission-control copy\", \"demo-forge\", \"plans\", \"simplify\", \"morning-ship\", \"semantic-guard\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0187",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Competitive Analysis Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "kwp-product-management-competitive-analysis",
        "kwp-marketing-competitive-analysis",
        "role-ceo",
        "kwp-customer-support-escalation",
        "kwp-finance-financial-statements",
        "kwp-sales-competitive-intelligence",
        "kwp-human-resources-people-analytics",
        "role-cso",
        "kwp-bio-research-single-cell-rna-qc",
        "kwp-data-data-exploration",
        "kwp-product-management-feature-spec",
        "kwp-data-data-validation",
        "kwp-marketing-campaign-planning",
        "kwp-finance-variance-analysis",
        "kwp-product-management-metrics-tracking",
        "kwp-marketing-performance-analytics",
        "kwp-operations-process-optimization",
        "kwp-human-resources-interview-prep",
        "kwp-sales-daily-briefing",
        "kwp-legal-meeting-briefing",
        "kwp-legal-legal-risk-assessment",
        "kwp-finance-audit-support",
        "kwp-operations-vendor-management",
        "kwp-human-resources-org-planning",
        "kwp-human-resources-recruiting-pipeline"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "kwp-product-management-competitive-analysis",
          "tokens": 2899
        },
        {
          "skill_id": "kwp-marketing-competitive-analysis",
          "tokens": 2723
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "kwp-customer-support-escalation",
          "tokens": 2329
        },
        {
          "skill_id": "kwp-finance-financial-statements",
          "tokens": 2768
        },
        {
          "skill_id": "kwp-sales-competitive-intelligence",
          "tokens": 3361
        },
        {
          "skill_id": "kwp-human-resources-people-analytics",
          "tokens": 574
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        },
        {
          "skill_id": "kwp-bio-research-single-cell-rna-qc",
          "tokens": 2261
        },
        {
          "skill_id": "kwp-data-data-exploration",
          "tokens": 2358
        },
        {
          "skill_id": "kwp-product-management-feature-spec",
          "tokens": 2614
        },
        {
          "skill_id": "kwp-data-data-validation",
          "tokens": 2650
        },
        {
          "skill_id": "kwp-marketing-campaign-planning",
          "tokens": 2803
        },
        {
          "skill_id": "kwp-finance-variance-analysis",
          "tokens": 3208
        },
        {
          "skill_id": "kwp-product-management-metrics-tracking",
          "tokens": 3606
        },
        {
          "skill_id": "kwp-marketing-performance-analytics",
          "tokens": 3966
        },
        {
          "skill_id": "kwp-operations-process-optimization",
          "tokens": 595
        },
        {
          "skill_id": "kwp-human-resources-interview-prep",
          "tokens": 763
        },
        {
          "skill_id": "kwp-sales-daily-briefing",
          "tokens": 1954
        },
        {
          "skill_id": "kwp-legal-meeting-briefing",
          "tokens": 2514
        },
        {
          "skill_id": "kwp-legal-legal-risk-assessment",
          "tokens": 3406
        },
        {
          "skill_id": "kwp-finance-audit-support",
          "tokens": 4426
        },
        {
          "skill_id": "kwp-operations-vendor-management",
          "tokens": 576
        },
        {
          "skill_id": "kwp-human-resources-org-planning",
          "tokens": 578
        },
        {
          "skill_id": "kwp-human-resources-recruiting-pipeline",
          "tokens": 578
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Competitive Analysis Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"kwp-product-management-competitive-analysis\", \"kwp-marketing-competitive-analysis\", \"role-ceo\", \"kwp-customer-support-escalation\", \"kwp-finance-financial-statements\", \"kwp-sales-competitive-intelligence\", \"kwp-human-resources-people-analytics\", \"role-cso\", \"kwp-bio-research-single-cell-rna-qc\", \"kwp-data-data-exploration\", \"kwp-product-management-feature-spec\", \"kwp-data-data-validation\", \"kwp-marketing-campaign-planning\", \"kwp-finance-variance-analysis\", \"kwp-product-management-metrics-tracking\", \"kwp-marketing-performance-analytics\", \"kwp-operations-process-optimization\", \"kwp-human-resources-interview-prep\", \"kwp-sales-daily-briefing\", \"kwp-legal-meeting-briefing\", \"kwp-legal-legal-risk-assessment\", \"kwp-finance-audit-support\", \"kwp-operations-vendor-management\", \"kwp-human-resources-org-planning\", \"kwp-human-resources-recruiting-pipeline\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0188",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Google Workspace CLI. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "gws-workspace",
        "gws-workflows",
        "gws-gmail",
        "gws-drive",
        "gws-calendar",
        "hf-hub",
        "gws-recipe-personalized-emails",
        "gws-docs",
        "gws-recipe-doc-from-template",
        "gws-chat",
        "gws-sheets",
        "gws-recipe-events-from-sheet",
        "gws-recipe-save-email-to-doc",
        "ai-chief-of-staff",
        "local-dev-runner",
        "gws-email-reply",
        "smart-meeting-scheduler",
        "gws-recipe-audit-sharing",
        "calendar-daily-briefing",
        "email-auto-reply",
        "knowledge-to-paper",
        "hf-buckets",
        "proactive-meeting-scheduler",
        "paperclip-setup",
        "notebooklm-studio"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "gws-workspace",
          "tokens": 1186
        },
        {
          "skill_id": "gws-workflows",
          "tokens": 952
        },
        {
          "skill_id": "gws-gmail",
          "tokens": 1796
        },
        {
          "skill_id": "gws-drive",
          "tokens": 997
        },
        {
          "skill_id": "gws-calendar",
          "tokens": 1075
        },
        {
          "skill_id": "hf-hub",
          "tokens": 1151
        },
        {
          "skill_id": "gws-recipe-personalized-emails",
          "tokens": 641
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "gws-recipe-doc-from-template",
          "tokens": 743
        },
        {
          "skill_id": "gws-chat",
          "tokens": 751
        },
        {
          "skill_id": "gws-sheets",
          "tokens": 918
        },
        {
          "skill_id": "gws-recipe-events-from-sheet",
          "tokens": 674
        },
        {
          "skill_id": "gws-recipe-save-email-to-doc",
          "tokens": 676
        },
        {
          "skill_id": "ai-chief-of-staff",
          "tokens": 1152
        },
        {
          "skill_id": "local-dev-runner",
          "tokens": 1419
        },
        {
          "skill_id": "gws-email-reply",
          "tokens": 2222
        },
        {
          "skill_id": "smart-meeting-scheduler",
          "tokens": 639
        },
        {
          "skill_id": "gws-recipe-audit-sharing",
          "tokens": 645
        },
        {
          "skill_id": "calendar-daily-briefing",
          "tokens": 1042
        },
        {
          "skill_id": "email-auto-reply",
          "tokens": 638
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "hf-buckets",
          "tokens": 1334
        },
        {
          "skill_id": "proactive-meeting-scheduler",
          "tokens": 1444
        },
        {
          "skill_id": "paperclip-setup",
          "tokens": 1708
        },
        {
          "skill_id": "notebooklm-studio",
          "tokens": 1812
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Google Workspace CLI. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"gws-workspace\", \"gws-workflows\", \"gws-gmail\", \"gws-drive\", \"gws-calendar\", \"hf-hub\", \"gws-recipe-personalized-emails\", \"gws-docs\", \"gws-recipe-doc-from-template\", \"gws-chat\", \"gws-sheets\", \"gws-recipe-events-from-sheet\", \"gws-recipe-save-email-to-doc\", \"ai-chief-of-staff\", \"local-dev-runner\", \"gws-email-reply\", \"smart-meeting-scheduler\", \"gws-recipe-audit-sharing\", \"calendar-daily-briefing\", \"email-auto-reply\", \"knowledge-to-paper\", \"hf-buckets\", \"proactive-meeting-scheduler\", \"paperclip-setup\", \"notebooklm-studio\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0189",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Options Theta Trading. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "trading-options-theta",
        "trading-options-strategy-advisor",
        "trading-position-sizer",
        "mirofish-financial-sim",
        "trading-sector-analyst",
        "weekly-stock-update",
        "trading-market-environment-analysis",
        "daily-stock-check",
        "trading-market-top-detector",
        "cursor-automations",
        "trading-theme-detector",
        "daily-am-orchestrator",
        "kwp-productivity-task-management",
        "trading-trade-hypothesis-ideator",
        "domain-commit",
        "stock-csv-downloader",
        "mirofish-opinion-sim",
        "diagnose",
        "terraform-reviewer",
        "trading-market-news-analyst",
        "prompt-transformer",
        "trading-uptrend-analyzer",
        "trading-market-breadth-analyzer",
        "prompt-architect",
        "trading-edge-candidate-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "trading-options-theta",
          "tokens": 1885
        },
        {
          "skill_id": "trading-options-strategy-advisor",
          "tokens": 1657
        },
        {
          "skill_id": "trading-position-sizer",
          "tokens": 2013
        },
        {
          "skill_id": "mirofish-financial-sim",
          "tokens": 1330
        },
        {
          "skill_id": "trading-sector-analyst",
          "tokens": 2657
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "trading-market-environment-analysis",
          "tokens": 1581
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "trading-market-top-detector",
          "tokens": 2295
        },
        {
          "skill_id": "cursor-automations",
          "tokens": 2911
        },
        {
          "skill_id": "trading-theme-detector",
          "tokens": 3487
        },
        {
          "skill_id": "daily-am-orchestrator",
          "tokens": 3613
        },
        {
          "skill_id": "kwp-productivity-task-management",
          "tokens": 776
        },
        {
          "skill_id": "trading-trade-hypothesis-ideator",
          "tokens": 909
        },
        {
          "skill_id": "domain-commit",
          "tokens": 915
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "mirofish-opinion-sim",
          "tokens": 1383
        },
        {
          "skill_id": "diagnose",
          "tokens": 1433
        },
        {
          "skill_id": "terraform-reviewer",
          "tokens": 1457
        },
        {
          "skill_id": "trading-market-news-analyst",
          "tokens": 1467
        },
        {
          "skill_id": "prompt-transformer",
          "tokens": 1720
        },
        {
          "skill_id": "trading-uptrend-analyzer",
          "tokens": 1780
        },
        {
          "skill_id": "trading-market-breadth-analyzer",
          "tokens": 1830
        },
        {
          "skill_id": "prompt-architect",
          "tokens": 1939
        },
        {
          "skill_id": "trading-edge-candidate-agent",
          "tokens": 1971
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Options Theta Trading. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"trading-options-theta\", \"trading-options-strategy-advisor\", \"trading-position-sizer\", \"mirofish-financial-sim\", \"trading-sector-analyst\", \"weekly-stock-update\", \"trading-market-environment-analysis\", \"daily-stock-check\", \"trading-market-top-detector\", \"cursor-automations\", \"trading-theme-detector\", \"daily-am-orchestrator\", \"kwp-productivity-task-management\", \"trading-trade-hypothesis-ideator\", \"domain-commit\", \"stock-csv-downloader\", \"mirofish-opinion-sim\", \"diagnose\", \"terraform-reviewer\", \"trading-market-news-analyst\", \"prompt-transformer\", \"trading-uptrend-analyzer\", \"trading-market-breadth-analyzer\", \"prompt-architect\", \"trading-edge-candidate-agent\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0190",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: QA Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-evidence-collector",
        "agency-reality-checker",
        "agency-instagram-curator",
        "agency-visual-storyteller",
        "agency-xr-cockpit-interaction-specialist",
        "agency-terminal-integration-specialist",
        "agency-behavioral-nudge-engine",
        "agency-senior-developer",
        "agency-cultural-intelligence-strategist",
        "agency-inclusive-visuals-specialist",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "agency-twitter-engager",
        "agency-whimsy-injector",
        "agency-rapid-prototyper",
        "agency-xiaohongshu-specialist",
        "agency-workflow-optimizer",
        "agency-lsp-index-engineer",
        "agency-image-prompt-engineer",
        "agency-ux-architect",
        "agency-macos-spatial-metal-engineer",
        "agency-zhihu-strategist",
        "agency-infrastructure-maintainer",
        "agency-support-responder",
        "agency-api-tester"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-evidence-collector",
          "tokens": 2300
        },
        {
          "skill_id": "agency-reality-checker",
          "tokens": 2764
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-xr-cockpit-interaction-specialist",
          "tokens": 693
        },
        {
          "skill_id": "agency-terminal-integration-specialist",
          "tokens": 1273
        },
        {
          "skill_id": "agency-behavioral-nudge-engine",
          "tokens": 1553
        },
        {
          "skill_id": "agency-senior-developer",
          "tokens": 1781
        },
        {
          "skill_id": "agency-cultural-intelligence-strategist",
          "tokens": 1892
        },
        {
          "skill_id": "agency-inclusive-visuals-specialist",
          "tokens": 1931
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-rapid-prototyper",
          "tokens": 2496
        },
        {
          "skill_id": "agency-xiaohongshu-specialist",
          "tokens": 2644
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-lsp-index-engineer",
          "tokens": 3006
        },
        {
          "skill_id": "agency-image-prompt-engineer",
          "tokens": 3027
        },
        {
          "skill_id": "agency-ux-architect",
          "tokens": 3119
        },
        {
          "skill_id": "agency-macos-spatial-metal-engineer",
          "tokens": 3229
        },
        {
          "skill_id": "agency-zhihu-strategist",
          "tokens": 3239
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-support-responder",
          "tokens": 3296
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: QA Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-evidence-collector\", \"agency-reality-checker\", \"agency-instagram-curator\", \"agency-visual-storyteller\", \"agency-xr-cockpit-interaction-specialist\", \"agency-terminal-integration-specialist\", \"agency-behavioral-nudge-engine\", \"agency-senior-developer\", \"agency-cultural-intelligence-strategist\", \"agency-inclusive-visuals-specialist\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"agency-twitter-engager\", \"agency-whimsy-injector\", \"agency-rapid-prototyper\", \"agency-xiaohongshu-specialist\", \"agency-workflow-optimizer\", \"agency-lsp-index-engineer\", \"agency-image-prompt-engineer\", \"agency-ux-architect\", \"agency-macos-spatial-metal-engineer\", \"agency-zhihu-strategist\", \"agency-infrastructure-maintainer\", \"agency-support-responder\", \"agency-api-tester\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0191",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Studio Producer Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-studio-producer",
        "agency-project-shepherd",
        "agency-studio-operations",
        "agency-infrastructure-maintainer",
        "agency-finance-tracker",
        "agency-legal-compliance-checker",
        "agency-visual-storyteller",
        "agency-ai-engineer",
        "agency-experiment-tracker",
        "agency-developer-advocate",
        "agency-growth-hacker",
        "agency-senior-project-manager",
        "agency-instagram-curator",
        "agency-whimsy-injector",
        "agency-mobile-app-builder",
        "agency-wechat-official-account-manager",
        "agency-workflow-optimizer",
        "agency-trend-researcher",
        "agency-ux-architect",
        "agency-brand-guardian",
        "agency-macos-spatial-metal-engineer",
        "agency-tool-evaluator",
        "agency-xr-immersive-developer",
        "agency-data-consolidation-agent",
        "agency-content-creator"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-studio-producer",
          "tokens": 2953
        },
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-finance-tracker",
          "tokens": 3177
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-senior-project-manager",
          "tokens": 1555
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        },
        {
          "skill_id": "agency-mobile-app-builder",
          "tokens": 2595
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "agency-ux-architect",
          "tokens": 3119
        },
        {
          "skill_id": "agency-brand-guardian",
          "tokens": 3135
        },
        {
          "skill_id": "agency-macos-spatial-metal-engineer",
          "tokens": 3229
        },
        {
          "skill_id": "agency-tool-evaluator",
          "tokens": 4549
        },
        {
          "skill_id": "agency-xr-immersive-developer",
          "tokens": 697
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Studio Producer Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-studio-producer\", \"agency-project-shepherd\", \"agency-studio-operations\", \"agency-infrastructure-maintainer\", \"agency-finance-tracker\", \"agency-legal-compliance-checker\", \"agency-visual-storyteller\", \"agency-ai-engineer\", \"agency-experiment-tracker\", \"agency-developer-advocate\", \"agency-growth-hacker\", \"agency-senior-project-manager\", \"agency-instagram-curator\", \"agency-whimsy-injector\", \"agency-mobile-app-builder\", \"agency-wechat-official-account-manager\", \"agency-workflow-optimizer\", \"agency-trend-researcher\", \"agency-ux-architect\", \"agency-brand-guardian\", \"agency-macos-spatial-metal-engineer\", \"agency-tool-evaluator\", \"agency-xr-immersive-developer\", \"agency-data-consolidation-agent\", \"agency-content-creator\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0192",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Sales Perspective Analyzer. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "role-sales",
        "role-ux-designer",
        "role-hr",
        "role-cso",
        "role-finance",
        "role-developer",
        "role-ceo",
        "role-pm",
        "role-security-engineer",
        "role-cto",
        "role-data-scientist",
        "role-trading-expert",
        "deep-research-pipeline",
        "executive-briefing",
        "paper-review",
        "full-stack-planner",
        "role-dispatcher",
        "prd-research-factory",
        "pm-market-research",
        "pm-go-to-market",
        "kwp-brand-voice-guideline-generation",
        "nlm-launch-kit",
        "meeting-digest",
        "kwp-engineering-documentation",
        "agency-data-consolidation-agent"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "role-sales",
          "tokens": 1221
        },
        {
          "skill_id": "role-ux-designer",
          "tokens": 1268
        },
        {
          "skill_id": "role-hr",
          "tokens": 1350
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        },
        {
          "skill_id": "role-finance",
          "tokens": 1353
        },
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "role-pm",
          "tokens": 1304
        },
        {
          "skill_id": "role-security-engineer",
          "tokens": 1406
        },
        {
          "skill_id": "role-cto",
          "tokens": 1413
        },
        {
          "skill_id": "role-data-scientist",
          "tokens": 1695
        },
        {
          "skill_id": "role-trading-expert",
          "tokens": 1791
        },
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        },
        {
          "skill_id": "executive-briefing",
          "tokens": 1419
        },
        {
          "skill_id": "paper-review",
          "tokens": 5815
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        },
        {
          "skill_id": "role-dispatcher",
          "tokens": 2075
        },
        {
          "skill_id": "prd-research-factory",
          "tokens": 648
        },
        {
          "skill_id": "pm-market-research",
          "tokens": 966
        },
        {
          "skill_id": "pm-go-to-market",
          "tokens": 992
        },
        {
          "skill_id": "kwp-brand-voice-guideline-generation",
          "tokens": 2021
        },
        {
          "skill_id": "nlm-launch-kit",
          "tokens": 2368
        },
        {
          "skill_id": "meeting-digest",
          "tokens": 4491
        },
        {
          "skill_id": "kwp-engineering-documentation",
          "tokens": 654
        },
        {
          "skill_id": "agency-data-consolidation-agent",
          "tokens": 843
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Sales Perspective Analyzer. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"role-sales\", \"role-ux-designer\", \"role-hr\", \"role-cso\", \"role-finance\", \"role-developer\", \"role-ceo\", \"role-pm\", \"role-security-engineer\", \"role-cto\", \"role-data-scientist\", \"role-trading-expert\", \"deep-research-pipeline\", \"executive-briefing\", \"paper-review\", \"full-stack-planner\", \"role-dispatcher\", \"prd-research-factory\", \"pm-market-research\", \"pm-go-to-market\", \"kwp-brand-voice-guideline-generation\", \"nlm-launch-kit\", \"meeting-digest\", \"kwp-engineering-documentation\", \"agency-data-consolidation-agent\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0193",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Strategic Compact Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "ecc-strategic-compact",
        "lead-lifecycle-manager",
        "ralph-loop",
        "ecc-autonomous-loops",
        "planning-meeting-sentinel",
        "email-auto-reply",
        "recall",
        "ecc-search-first",
        "feature-intake-router",
        "prd-research-factory",
        "ecc-agentic-engineering",
        "ecc-verification-loop",
        "diagnose",
        "ui-suite",
        "auto-research-distribute",
        "ecc-eval-harness",
        "mission-control",
        "simplify",
        "problem-definition",
        "anthropic-doc-coauthoring",
        "docs-freshness-guardian",
        "sales-weekly-dashboard",
        "tech-support-brain",
        "kwp-design-user-research",
        "kwp-engineering-system-design"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "ecc-strategic-compact",
          "tokens": 1252
        },
        {
          "skill_id": "lead-lifecycle-manager",
          "tokens": 533
        },
        {
          "skill_id": "ralph-loop",
          "tokens": 1925
        },
        {
          "skill_id": "ecc-autonomous-loops",
          "tokens": 4660
        },
        {
          "skill_id": "planning-meeting-sentinel",
          "tokens": 604
        },
        {
          "skill_id": "email-auto-reply",
          "tokens": 638
        },
        {
          "skill_id": "recall",
          "tokens": 1256
        },
        {
          "skill_id": "ecc-search-first",
          "tokens": 1597
        },
        {
          "skill_id": "feature-intake-router",
          "tokens": 568
        },
        {
          "skill_id": "prd-research-factory",
          "tokens": 648
        },
        {
          "skill_id": "ecc-agentic-engineering",
          "tokens": 690
        },
        {
          "skill_id": "ecc-verification-loop",
          "tokens": 886
        },
        {
          "skill_id": "diagnose",
          "tokens": 1433
        },
        {
          "skill_id": "ui-suite",
          "tokens": 1545
        },
        {
          "skill_id": "auto-research-distribute",
          "tokens": 1771
        },
        {
          "skill_id": "ecc-eval-harness",
          "tokens": 1810
        },
        {
          "skill_id": "mission-control",
          "tokens": 2365
        },
        {
          "skill_id": "simplify",
          "tokens": 2596
        },
        {
          "skill_id": "problem-definition",
          "tokens": 2601
        },
        {
          "skill_id": "anthropic-doc-coauthoring",
          "tokens": 4138
        },
        {
          "skill_id": "docs-freshness-guardian",
          "tokens": 557
        },
        {
          "skill_id": "sales-weekly-dashboard",
          "tokens": 558
        },
        {
          "skill_id": "tech-support-brain",
          "tokens": 620
        },
        {
          "skill_id": "kwp-design-user-research",
          "tokens": 718
        },
        {
          "skill_id": "kwp-engineering-system-design",
          "tokens": 756
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Strategic Compact Skill. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"ecc-strategic-compact\", \"lead-lifecycle-manager\", \"ralph-loop\", \"ecc-autonomous-loops\", \"planning-meeting-sentinel\", \"email-auto-reply\", \"recall\", \"ecc-search-first\", \"feature-intake-router\", \"prd-research-factory\", \"ecc-agentic-engineering\", \"ecc-verification-loop\", \"diagnose\", \"ui-suite\", \"auto-research-distribute\", \"ecc-eval-harness\", \"mission-control\", \"simplify\", \"problem-definition\", \"anthropic-doc-coauthoring\", \"docs-freshness-guardian\", \"sales-weekly-dashboard\", \"tech-support-brain\", \"kwp-design-user-research\", \"kwp-engineering-system-design\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0194",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AI Platform FE - FSD 개발 및 마이그레이션. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "fsd-development",
        "design-review",
        "figma-to-tds",
        "implement-screen",
        "fe-pipeline",
        "swagger-api-doc-generator",
        "github-workflow-automation",
        "overlay-layout-patterns",
        "screen-description",
        "notion-docs-sync",
        "anthropic-frontend-design",
        "design-architect",
        "tab-genai-features",
        "tab-llm-agents",
        "tab-screening",
        "planning-weekly-pulse",
        "dependency-radar",
        "lead-lifecycle-manager",
        "docs-freshness-guardian",
        "sales-weekly-dashboard",
        "sales-meeting-pipeline",
        "feature-intake-router",
        "proposal-factory",
        "planning-meeting-sentinel",
        "incident-lifecycle-manager"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "fsd-development",
          "tokens": 1234
        },
        {
          "skill_id": "design-review",
          "tokens": 947
        },
        {
          "skill_id": "figma-to-tds",
          "tokens": 987
        },
        {
          "skill_id": "implement-screen",
          "tokens": 996
        },
        {
          "skill_id": "fe-pipeline",
          "tokens": 2354
        },
        {
          "skill_id": "swagger-api-doc-generator",
          "tokens": 784
        },
        {
          "skill_id": "github-workflow-automation",
          "tokens": 971
        },
        {
          "skill_id": "overlay-layout-patterns",
          "tokens": 1047
        },
        {
          "skill_id": "screen-description",
          "tokens": 836
        },
        {
          "skill_id": "notion-docs-sync",
          "tokens": 857
        },
        {
          "skill_id": "anthropic-frontend-design",
          "tokens": 1312
        },
        {
          "skill_id": "design-architect",
          "tokens": 1827
        },
        {
          "skill_id": "tab-genai-features",
          "tokens": 354
        },
        {
          "skill_id": "tab-llm-agents",
          "tokens": 382
        },
        {
          "skill_id": "tab-screening",
          "tokens": 410
        },
        {
          "skill_id": "planning-weekly-pulse",
          "tokens": 528
        },
        {
          "skill_id": "dependency-radar",
          "tokens": 531
        },
        {
          "skill_id": "lead-lifecycle-manager",
          "tokens": 533
        },
        {
          "skill_id": "docs-freshness-guardian",
          "tokens": 557
        },
        {
          "skill_id": "sales-weekly-dashboard",
          "tokens": 558
        },
        {
          "skill_id": "sales-meeting-pipeline",
          "tokens": 559
        },
        {
          "skill_id": "feature-intake-router",
          "tokens": 568
        },
        {
          "skill_id": "proposal-factory",
          "tokens": 582
        },
        {
          "skill_id": "planning-meeting-sentinel",
          "tokens": 604
        },
        {
          "skill_id": "incident-lifecycle-manager",
          "tokens": 610
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: AI Platform FE - FSD 개발 및 마이그레이션. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"fsd-development\", \"design-review\", \"figma-to-tds\", \"implement-screen\", \"fe-pipeline\", \"swagger-api-doc-generator\", \"github-workflow-automation\", \"overlay-layout-patterns\", \"screen-description\", \"notion-docs-sync\", \"anthropic-frontend-design\", \"design-architect\", \"tab-genai-features\", \"tab-llm-agents\", \"tab-screening\", \"planning-weekly-pulse\", \"dependency-radar\", \"lead-lifecycle-manager\", \"docs-freshness-guardian\", \"sales-weekly-dashboard\", \"sales-meeting-pipeline\", \"feature-intake-router\", \"proposal-factory\", \"planning-meeting-sentinel\", \"incident-lifecycle-manager\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0195",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Execute HuggingFace Hub operations via the hf CLI. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "hf-cli",
        "hf-models",
        "hf-model-trainer",
        "hf-buckets",
        "hf-datasets",
        "hf-repos",
        "hf-discussions",
        "hf-collections",
        "hf-dataset-viewer",
        "hf-jobs",
        "hf-evaluation",
        "hf-hub",
        "hf-endpoints",
        "gws-drive",
        "hf-trackio",
        "weekly-stock-update",
        "daily-stock-check",
        "hf-papers",
        "alphaear-sentiment",
        "hf-spaces",
        "stock-csv-downloader",
        "gws-gmail",
        "tab-glossary",
        "gws-docs",
        "alphaear-deepear-lite"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "hf-cli",
          "tokens": 2275
        },
        {
          "skill_id": "hf-models",
          "tokens": 1212
        },
        {
          "skill_id": "hf-model-trainer",
          "tokens": 5150
        },
        {
          "skill_id": "hf-buckets",
          "tokens": 1334
        },
        {
          "skill_id": "hf-datasets",
          "tokens": 3241
        },
        {
          "skill_id": "hf-repos",
          "tokens": 1106
        },
        {
          "skill_id": "hf-discussions",
          "tokens": 1118
        },
        {
          "skill_id": "hf-collections",
          "tokens": 1239
        },
        {
          "skill_id": "hf-dataset-viewer",
          "tokens": 1387
        },
        {
          "skill_id": "hf-jobs",
          "tokens": 4467
        },
        {
          "skill_id": "hf-evaluation",
          "tokens": 4549
        },
        {
          "skill_id": "hf-hub",
          "tokens": 1151
        },
        {
          "skill_id": "hf-endpoints",
          "tokens": 1177
        },
        {
          "skill_id": "gws-drive",
          "tokens": 997
        },
        {
          "skill_id": "hf-trackio",
          "tokens": 1424
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "hf-papers",
          "tokens": 817
        },
        {
          "skill_id": "alphaear-sentiment",
          "tokens": 928
        },
        {
          "skill_id": "hf-spaces",
          "tokens": 930
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        },
        {
          "skill_id": "gws-gmail",
          "tokens": 1796
        },
        {
          "skill_id": "tab-glossary",
          "tokens": 538
        },
        {
          "skill_id": "gws-docs",
          "tokens": 683
        },
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Execute HuggingFace Hub operations via the hf CLI. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"hf-cli\", \"hf-models\", \"hf-model-trainer\", \"hf-buckets\", \"hf-datasets\", \"hf-repos\", \"hf-discussions\", \"hf-collections\", \"hf-dataset-viewer\", \"hf-jobs\", \"hf-evaluation\", \"hf-hub\", \"hf-endpoints\", \"gws-drive\", \"hf-trackio\", \"weekly-stock-update\", \"daily-stock-check\", \"hf-papers\", \"alphaear-sentiment\", \"hf-spaces\", \"stock-csv-downloader\", \"gws-gmail\", \"tab-glossary\", \"gws-docs\", \"alphaear-deepear-lite\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0196",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-technical-analysis. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "tab-technical-analysis",
        "tab-turtle-refresh",
        "tab-bollinger-refresh",
        "tab-market-breadth",
        "tab-strategy-comparison",
        "tab-glossary",
        "tab-stock-sync",
        "tab-genai-features",
        "tab-dualma-refresh",
        "tab-screening",
        "tab-dashboard-summary",
        "tab-news-fetch",
        "tab-market-environment",
        "tab-report-generate",
        "weekly-stock-update",
        "daily-stock-check",
        "morning-ship",
        "tab-event-detect",
        "tab-fundamental-sync",
        "tab-llm-agents",
        "tab-patterns",
        "daily-strategy-post",
        "tab-hot-stock-discovery",
        "tab-analysis-run",
        "email-research-dispatcher"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "tab-technical-analysis",
          "tokens": 414
        },
        {
          "skill_id": "tab-turtle-refresh",
          "tokens": 371
        },
        {
          "skill_id": "tab-bollinger-refresh",
          "tokens": 383
        },
        {
          "skill_id": "tab-market-breadth",
          "tokens": 504
        },
        {
          "skill_id": "tab-strategy-comparison",
          "tokens": 571
        },
        {
          "skill_id": "tab-glossary",
          "tokens": 538
        },
        {
          "skill_id": "tab-stock-sync",
          "tokens": 345
        },
        {
          "skill_id": "tab-genai-features",
          "tokens": 354
        },
        {
          "skill_id": "tab-dualma-refresh",
          "tokens": 391
        },
        {
          "skill_id": "tab-screening",
          "tokens": 410
        },
        {
          "skill_id": "tab-dashboard-summary",
          "tokens": 419
        },
        {
          "skill_id": "tab-news-fetch",
          "tokens": 474
        },
        {
          "skill_id": "tab-market-environment",
          "tokens": 566
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "daily-stock-check",
          "tokens": 1739
        },
        {
          "skill_id": "morning-ship",
          "tokens": 2943
        },
        {
          "skill_id": "tab-event-detect",
          "tokens": 319
        },
        {
          "skill_id": "tab-fundamental-sync",
          "tokens": 331
        },
        {
          "skill_id": "tab-llm-agents",
          "tokens": 382
        },
        {
          "skill_id": "tab-patterns",
          "tokens": 509
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "tab-hot-stock-discovery",
          "tokens": 359
        },
        {
          "skill_id": "tab-analysis-run",
          "tokens": 404
        },
        {
          "skill_id": "email-research-dispatcher",
          "tokens": 703
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: tab-technical-analysis. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"tab-technical-analysis\", \"tab-turtle-refresh\", \"tab-bollinger-refresh\", \"tab-market-breadth\", \"tab-strategy-comparison\", \"tab-glossary\", \"tab-stock-sync\", \"tab-genai-features\", \"tab-dualma-refresh\", \"tab-screening\", \"tab-dashboard-summary\", \"tab-news-fetch\", \"tab-market-environment\", \"tab-report-generate\", \"weekly-stock-update\", \"daily-stock-check\", \"morning-ship\", \"tab-event-detect\", \"tab-fundamental-sync\", \"tab-llm-agents\", \"tab-patterns\", \"daily-strategy-post\", \"tab-hot-stock-discovery\", \"tab-analysis-run\", \"email-research-dispatcher\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0197",
      "setting": "budget_constrained",
      "query": "Within a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Overview. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "hf-evaluation",
        "hf-trackio",
        "hf-model-trainer",
        "hf-cli",
        "hf-jobs",
        "ecc-eval-harness",
        "hf-datasets",
        "hf-dataset-viewer",
        "alphaear-deepear-lite",
        "alphaear-signal-tracker",
        "alphaear-sentiment",
        "hf-models",
        "alphaear-predictor",
        "alphaear-news",
        "hf-endpoints",
        "weekly-stock-update",
        "ai-quality-evaluator",
        "trading-us-stock-analysis",
        "agency-test-results-analyzer",
        "anthropic-skill-creator",
        "tab-kiwoom",
        "tab-report-generate",
        "alphaear-reporter",
        "alphaear-stock",
        "stock-csv-downloader"
      ],
      "distractor_skills": [],
      "budget_tokens": 1000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "hf-evaluation",
          "tokens": 4549
        },
        {
          "skill_id": "hf-trackio",
          "tokens": 1424
        },
        {
          "skill_id": "hf-model-trainer",
          "tokens": 5150
        },
        {
          "skill_id": "hf-cli",
          "tokens": 2275
        },
        {
          "skill_id": "hf-jobs",
          "tokens": 4467
        },
        {
          "skill_id": "ecc-eval-harness",
          "tokens": 1810
        },
        {
          "skill_id": "hf-datasets",
          "tokens": 3241
        },
        {
          "skill_id": "hf-dataset-viewer",
          "tokens": 1387
        },
        {
          "skill_id": "alphaear-deepear-lite",
          "tokens": 727
        },
        {
          "skill_id": "alphaear-signal-tracker",
          "tokens": 854
        },
        {
          "skill_id": "alphaear-sentiment",
          "tokens": 928
        },
        {
          "skill_id": "hf-models",
          "tokens": 1212
        },
        {
          "skill_id": "alphaear-predictor",
          "tokens": 803
        },
        {
          "skill_id": "alphaear-news",
          "tokens": 873
        },
        {
          "skill_id": "hf-endpoints",
          "tokens": 1177
        },
        {
          "skill_id": "weekly-stock-update",
          "tokens": 1497
        },
        {
          "skill_id": "ai-quality-evaluator",
          "tokens": 2228
        },
        {
          "skill_id": "trading-us-stock-analysis",
          "tokens": 2926
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "anthropic-skill-creator",
          "tokens": 6862
        },
        {
          "skill_id": "tab-kiwoom",
          "tokens": 452
        },
        {
          "skill_id": "tab-report-generate",
          "tokens": 575
        },
        {
          "skill_id": "alphaear-reporter",
          "tokens": 810
        },
        {
          "skill_id": "alphaear-stock",
          "tokens": 820
        },
        {
          "skill_id": "stock-csv-downloader",
          "tokens": 1309
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 1000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Overview. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"hf-evaluation\", \"hf-trackio\", \"hf-model-trainer\", \"hf-cli\", \"hf-jobs\", \"ecc-eval-harness\", \"hf-datasets\", \"hf-dataset-viewer\", \"alphaear-deepear-lite\", \"alphaear-signal-tracker\", \"alphaear-sentiment\", \"hf-models\", \"alphaear-predictor\", \"alphaear-news\", \"hf-endpoints\", \"weekly-stock-update\", \"ai-quality-evaluator\", \"trading-us-stock-analysis\", \"agency-test-results-analyzer\", \"anthropic-skill-creator\", \"tab-kiwoom\", \"tab-report-generate\", \"alphaear-reporter\", \"alphaear-stock\", \"stock-csv-downloader\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0198",
      "setting": "budget_constrained",
      "query": "Within a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: SOD Ship. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "sod-ship",
        "morning-ship",
        "eod-ship",
        "morning",
        "cursor-sync",
        "domain-commit",
        "release-ship",
        "daily-am-orchestrator",
        "commit-to-issue",
        "ship",
        "jira-linear-adapter",
        "portfolio-report-generator",
        "release-commander",
        "daily-strategy-post",
        "knowledge-daily-aggregator",
        "trading-data-quality-checker",
        "bespin-news-digest",
        "github-sprint-digest",
        "knowledge-to-paper",
        "ai-chief-of-staff",
        "pr-to-issue-linker",
        "sprint-orchestrator",
        "hf-buckets",
        "compliance-gate",
        "proactive-meeting-scheduler"
      ],
      "distractor_skills": [],
      "budget_tokens": 2000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "sod-ship",
          "tokens": 4151
        },
        {
          "skill_id": "morning-ship",
          "tokens": 2943
        },
        {
          "skill_id": "eod-ship",
          "tokens": 2420
        },
        {
          "skill_id": "morning",
          "tokens": 2088
        },
        {
          "skill_id": "cursor-sync",
          "tokens": 3267
        },
        {
          "skill_id": "domain-commit",
          "tokens": 915
        },
        {
          "skill_id": "release-ship",
          "tokens": 3056
        },
        {
          "skill_id": "daily-am-orchestrator",
          "tokens": 3613
        },
        {
          "skill_id": "commit-to-issue",
          "tokens": 1424
        },
        {
          "skill_id": "ship",
          "tokens": 1608
        },
        {
          "skill_id": "jira-linear-adapter",
          "tokens": 1365
        },
        {
          "skill_id": "portfolio-report-generator",
          "tokens": 1406
        },
        {
          "skill_id": "release-commander",
          "tokens": 2245
        },
        {
          "skill_id": "daily-strategy-post",
          "tokens": 711
        },
        {
          "skill_id": "knowledge-daily-aggregator",
          "tokens": 1425
        },
        {
          "skill_id": "trading-data-quality-checker",
          "tokens": 1962
        },
        {
          "skill_id": "bespin-news-digest",
          "tokens": 3949
        },
        {
          "skill_id": "github-sprint-digest",
          "tokens": 671
        },
        {
          "skill_id": "knowledge-to-paper",
          "tokens": 755
        },
        {
          "skill_id": "ai-chief-of-staff",
          "tokens": 1152
        },
        {
          "skill_id": "pr-to-issue-linker",
          "tokens": 1223
        },
        {
          "skill_id": "sprint-orchestrator",
          "tokens": 1320
        },
        {
          "skill_id": "hf-buckets",
          "tokens": 1334
        },
        {
          "skill_id": "compliance-gate",
          "tokens": 1376
        },
        {
          "skill_id": "proactive-meeting-scheduler",
          "tokens": 1444
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 2000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: SOD Ship. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"sod-ship\", \"morning-ship\", \"eod-ship\", \"morning\", \"cursor-sync\", \"domain-commit\", \"release-ship\", \"daily-am-orchestrator\", \"commit-to-issue\", \"ship\", \"jira-linear-adapter\", \"portfolio-report-generator\", \"release-commander\", \"daily-strategy-post\", \"knowledge-daily-aggregator\", \"trading-data-quality-checker\", \"bespin-news-digest\", \"github-sprint-digest\", \"knowledge-to-paper\", \"ai-chief-of-staff\", \"pr-to-issue-linker\", \"sprint-orchestrator\", \"hf-buckets\", \"compliance-gate\", \"proactive-meeting-scheduler\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0199",
      "setting": "budget_constrained",
      "query": "Within a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Project Shepherd Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "agency-project-shepherd",
        "agency-studio-producer",
        "agency-studio-operations",
        "agency-experiment-tracker",
        "agency-senior-project-manager",
        "agency-feedback-synthesizer",
        "agency-ai-engineer",
        "agency-wechat-official-account-manager",
        "agency-workflow-optimizer",
        "agency-trend-researcher",
        "agency-finance-tracker",
        "agency-infrastructure-maintainer",
        "agency-api-tester",
        "agency-legal-compliance-checker",
        "agency-test-results-analyzer",
        "agency-developer-advocate",
        "agency-growth-hacker",
        "agency-content-creator",
        "agency-instagram-curator",
        "agency-visual-storyteller",
        "agency-social-media-strategist",
        "agency-reddit-community-builder",
        "agency-tiktok-strategist",
        "agency-twitter-engager",
        "agency-whimsy-injector"
      ],
      "distractor_skills": [],
      "budget_tokens": 4000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "agency-project-shepherd",
          "tokens": 2675
        },
        {
          "skill_id": "agency-studio-producer",
          "tokens": 2953
        },
        {
          "skill_id": "agency-studio-operations",
          "tokens": 2697
        },
        {
          "skill_id": "agency-experiment-tracker",
          "tokens": 2587
        },
        {
          "skill_id": "agency-senior-project-manager",
          "tokens": 1555
        },
        {
          "skill_id": "agency-feedback-synthesizer",
          "tokens": 2041
        },
        {
          "skill_id": "agency-ai-engineer",
          "tokens": 2083
        },
        {
          "skill_id": "agency-wechat-official-account-manager",
          "tokens": 2703
        },
        {
          "skill_id": "agency-workflow-optimizer",
          "tokens": 2705
        },
        {
          "skill_id": "agency-trend-researcher",
          "tokens": 2852
        },
        {
          "skill_id": "agency-finance-tracker",
          "tokens": 3177
        },
        {
          "skill_id": "agency-infrastructure-maintainer",
          "tokens": 3266
        },
        {
          "skill_id": "agency-api-tester",
          "tokens": 3360
        },
        {
          "skill_id": "agency-legal-compliance-checker",
          "tokens": 3578
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        },
        {
          "skill_id": "agency-growth-hacker",
          "tokens": 998
        },
        {
          "skill_id": "agency-content-creator",
          "tokens": 1020
        },
        {
          "skill_id": "agency-instagram-curator",
          "tokens": 1879
        },
        {
          "skill_id": "agency-visual-storyteller",
          "tokens": 2082
        },
        {
          "skill_id": "agency-social-media-strategist",
          "tokens": 2084
        },
        {
          "skill_id": "agency-reddit-community-builder",
          "tokens": 2119
        },
        {
          "skill_id": "agency-tiktok-strategist",
          "tokens": 2127
        },
        {
          "skill_id": "agency-twitter-engager",
          "tokens": 2188
        },
        {
          "skill_id": "agency-whimsy-injector",
          "tokens": 2428
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 4000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Project Shepherd Agent Personality. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"agency-project-shepherd\", \"agency-studio-producer\", \"agency-studio-operations\", \"agency-experiment-tracker\", \"agency-senior-project-manager\", \"agency-feedback-synthesizer\", \"agency-ai-engineer\", \"agency-wechat-official-account-manager\", \"agency-workflow-optimizer\", \"agency-trend-researcher\", \"agency-finance-tracker\", \"agency-infrastructure-maintainer\", \"agency-api-tester\", \"agency-legal-compliance-checker\", \"agency-test-results-analyzer\", \"agency-developer-advocate\", \"agency-growth-hacker\", \"agency-content-creator\", \"agency-instagram-curator\", \"agency-visual-storyteller\", \"agency-social-media-strategist\", \"agency-reddit-community-builder\", \"agency-tiktok-strategist\", \"agency-twitter-engager\", \"agency-whimsy-injector\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    },
    {
      "query_id": "BC-0200",
      "setting": "budget_constrained",
      "query": "Within a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Developer Perspective Analyzer. Assume each candidate skill consumes its full SKILL.md token_count when packed.",
      "gold_skills": [
        "role-developer",
        "role-cto",
        "role-hr",
        "role-finance",
        "role-sales",
        "role-ux-designer",
        "role-security-engineer",
        "role-ceo",
        "role-pm",
        "role-cso",
        "role-data-scientist",
        "role-trading-expert",
        "deep-research-pipeline",
        "executive-briefing",
        "refactor-simulator",
        "problem-definition",
        "kwp-engineering-documentation",
        "diagnose",
        "issue-to-code",
        "infra-runbook-generator",
        "full-stack-planner",
        "incident-to-improvement",
        "agency-ux-architect",
        "agency-test-results-analyzer",
        "agency-developer-advocate"
      ],
      "distractor_skills": [],
      "budget_tokens": 8000,
      "difficulty": "hard",
      "source": "synthetic",
      "ranked_skill_token_costs": [
        {
          "skill_id": "role-developer",
          "tokens": 1386
        },
        {
          "skill_id": "role-cto",
          "tokens": 1413
        },
        {
          "skill_id": "role-hr",
          "tokens": 1350
        },
        {
          "skill_id": "role-finance",
          "tokens": 1353
        },
        {
          "skill_id": "role-sales",
          "tokens": 1221
        },
        {
          "skill_id": "role-ux-designer",
          "tokens": 1268
        },
        {
          "skill_id": "role-security-engineer",
          "tokens": 1406
        },
        {
          "skill_id": "role-ceo",
          "tokens": 1188
        },
        {
          "skill_id": "role-pm",
          "tokens": 1304
        },
        {
          "skill_id": "role-cso",
          "tokens": 1312
        },
        {
          "skill_id": "role-data-scientist",
          "tokens": 1695
        },
        {
          "skill_id": "role-trading-expert",
          "tokens": 1791
        },
        {
          "skill_id": "deep-research-pipeline",
          "tokens": 2220
        },
        {
          "skill_id": "executive-briefing",
          "tokens": 1419
        },
        {
          "skill_id": "refactor-simulator",
          "tokens": 2044
        },
        {
          "skill_id": "problem-definition",
          "tokens": 2601
        },
        {
          "skill_id": "kwp-engineering-documentation",
          "tokens": 654
        },
        {
          "skill_id": "diagnose",
          "tokens": 1433
        },
        {
          "skill_id": "issue-to-code",
          "tokens": 1576
        },
        {
          "skill_id": "infra-runbook-generator",
          "tokens": 1623
        },
        {
          "skill_id": "full-stack-planner",
          "tokens": 2101
        },
        {
          "skill_id": "incident-to-improvement",
          "tokens": 2624
        },
        {
          "skill_id": "agency-ux-architect",
          "tokens": 3119
        },
        {
          "skill_id": "agency-test-results-analyzer",
          "tokens": 3833
        },
        {
          "skill_id": "agency-developer-advocate",
          "tokens": 3906
        }
      ],
      "judge_verdict": "uncertain",
      "judge_confidence": 0.71,
      "judge_rationale": "stubbed heuristic judge (no LLM call)",
      "needs_human_review": true,
      "judge_prompt_user": "Query:\nWithin a 8000-token retrieval budget over the whole 501-skill library, list skills most relevant to workflows similar to: Developer Perspective Analyzer. Assume each candidate skill consumes its full SKILL.md token_count when packed.\n\nProposed gold_skills (ordered if composition):\n[\"role-developer\", \"role-cto\", \"role-hr\", \"role-finance\", \"role-sales\", \"role-ux-designer\", \"role-security-engineer\", \"role-ceo\", \"role-pm\", \"role-cso\", \"role-data-scientist\", \"role-trading-expert\", \"deep-research-pipeline\", \"executive-briefing\", \"refactor-simulator\", \"problem-definition\", \"kwp-engineering-documentation\", \"diagnose\", \"issue-to-code\", \"infra-runbook-generator\", \"full-stack-planner\", \"incident-to-improvement\", \"agency-ux-architect\", \"agency-test-results-analyzer\", \"agency-developer-advocate\"]\n\nSetting: budget_constrained\nDistractors (if any, first 15 shown):\n[]\n\nDoes the gold set match the query intent for ThakiCloud Cursor skills?\n"
    }
  ]
}