[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7194F3WH9WJ89WNXDTN1WKF":3,"skill-same:01M7194F3WH9WJ89WNXDTN1WKF":68},{"skillNo":4,"name":5,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":8,"repositoryNo":11,"starsCount":12,"licenseSpdxId":13,"licenseFileUrl":6,"sourceHtmlUrl":14,"categories":15,"keywords":16,"publishedAt":29,"linkOnly":30,"markdownContent":31,"markdownBody":32,"frontMatter":33,"refName":40,"commitSha":41,"skillDirectory":42,"forksCount":43,"repositoryHtmlUrl":44,"repositoryHomepageUrl":45,"ownerHtmlUrl":46,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":47,"install":48,"resources":61},"01M7194F3WH9WJ89WNXDTN1WKF","evolve-evaluate",null,"openplaybooks-dev\u002Fconverge","openplaybooks-dev","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F270560357?v=4","01M7191G17AHW5SXN63HJ0BEZV","01M7191G29922BDNWN8CFVG4PD",3,"MIT","https:\u002F\u002Fgithub.com\u002Fopenplaybooks-dev\u002Fconverge\u002Fblob\u002Fmain\u002Fexamples\u002Fevolutionary-optimization\u002F.converge\u002Fplaybooks\u002Fskills\u002Fevolve-evaluate\u002FSKILL.md",[],[17,18,19,20,21,22,23,24,25,26,27,28],"evolve evaluate","evaluate","evolve","training","output","check","valid","format","node","fitness","configuration","js","2026-11-10 10:12:30",false,"---\nid: evolve-evaluate\ntitle: Evaluate Training Configuration\nchecks:\n  - id: score-format\n    cmd: node .\u002Fcheck.js\n    description: Scored output has valid fitness and scores\n---\n\n# Evaluate Training Configuration\n\nScore a single LLM training configuration against the training goal.\n\n## Process\n\n1. Read the candidate training configuration from the task body\n2. Evaluate architecture choices:\n   - Does the attention type suit the model scale? (GQA\u002FMQA preferred at 7B+)\n   - Is depth\u002Fwidth ratio appropriate? (typical: 32-40 layers for 7B)\n   - Is FFN expansion ratio efficient? (SwiGLU ~2.67x standard)\n   - Does context length match training goal requirements?\n3. Evaluate hyperparameters:\n   - Learning rate vs batch size relationship (linear scaling rule)\n   - Warmup proportion (typically 0.5-2% of total steps)\n   - Weight decay magnitude (0.01-0.1 typical)\n   - Optimizer suitability for the scale\n4. Evaluate data strategy:\n   - Token budget vs param count (Chinchilla: ~20 tokens\u002Fparam)\n   - Mix alignment with training goal (code model needs code-heavy mix)\n   - Filtering and dedup quality\n   - Curriculum design appropriateness\n5. Score on four criteria (each 0-1):\n   - **benchmarkPotential**: Expected performance on target benchmarks\n   - **trainingEfficiency**: FLOPs per quality point, tokens-per-param ratio\n   - **scalingProperties**: Adherence to known scaling laws\n   - **robustness**: Training stability, divergence risk\n6. Compute overall fitness as weighted average\n7. Identify specific strengths and weaknesses with technical rationale\n\n## Outputs\n\n- Scored candidate JSON file:\n  ```json\n  {\n    \"candidateId\": \"candidate-001\",\n    \"generation\": 0,\n    \"fitness\": 0.75,\n    \"scores\": {\n      \"benchmarkPotential\": 0.8,\n      \"trainingEfficiency\": 0.7,\n      \"scalingProperties\": 0.6,\n      \"robustness\": 0.9\n    },\n    \"strengths\": [\"GQA with 4 KV heads is memory-efficient for inference\", \"cosine schedule tuned for token budget\"],\n    \"weaknesses\": [\"LR 3e-4 may be too high for 4M batch size\", \"no math curriculum despite math benchmark targets\"]\n  }\n  ```\n\n## Quality Criteria\n\n- Fitness is a number between 0 and 1\n- All four score dimensions present\n- Strengths and weaknesses reference specific technical details (not generic)\n- Scoring is grounded in known scaling laws and training best practices\n","# Evaluate Training Configuration\n\nScore a single LLM training configuration against the training goal.\n\n## Process\n\n1. Read the candidate training configuration from the task body\n2. Evaluate architecture choices:\n   - Does the attention type suit the model scale? (GQA\u002FMQA preferred at 7B+)\n   - Is depth\u002Fwidth ratio appropriate? (typical: 32-40 layers for 7B)\n   - Is FFN expansion ratio efficient? (SwiGLU ~2.67x standard)\n   - Does context length match training goal requirements?\n3. Evaluate hyperparameters:\n   - Learning rate vs batch size relationship (linear scaling rule)\n   - Warmup proportion (typically 0.5-2% of total steps)\n   - Weight decay magnitude (0.01-0.1 typical)\n   - Optimizer suitability for the scale\n4. Evaluate data strategy:\n   - Token budget vs param count (Chinchilla: ~20 tokens\u002Fparam)\n   - Mix alignment with training goal (code model needs code-heavy mix)\n   - Filtering and dedup quality\n   - Curriculum design appropriateness\n5. Score on four criteria (each 0-1):\n   - **benchmarkPotential**: Expected performance on target benchmarks\n   - **trainingEfficiency**: FLOPs per quality point, tokens-per-param ratio\n   - **scalingProperties**: Adherence to known scaling laws\n   - **robustness**: Training stability, divergence risk\n6. Compute overall fitness as weighted average\n7. Identify specific strengths and weaknesses with technical rationale\n\n## Outputs\n\n- Scored candidate JSON file:\n  ```json\n  {\n    \"candidateId\": \"candidate-001\",\n    \"generation\": 0,\n    \"fitness\": 0.75,\n    \"scores\": {\n      \"benchmarkPotential\": 0.8,\n      \"trainingEfficiency\": 0.7,\n      \"scalingProperties\": 0.6,\n      \"robustness\": 0.9\n    },\n    \"strengths\": [\"GQA with 4 KV heads is memory-efficient for inference\", \"cosine schedule tuned for token budget\"],\n    \"weaknesses\": [\"LR 3e-4 may be too high for 4M batch size\", \"no math curriculum despite math benchmark targets\"]\n  }\n  ```\n\n## Quality Criteria\n\n- Fitness is a number between 0 and 1\n- All four score dimensions present\n- Strengths and weaknesses reference specific technical details (not generic)\n- Scoring is grounded in known scaling laws and training best practices\n",{"id":5,"title":34,"checks":35},"Evaluate Training Configuration",[36],{"id":37,"cmd":38,"description":39},"score-format","node .\u002Fcheck.js","Scored output has valid fitness and scores","main","7d618d8bc059f1a0974706d1d15ec25f73cf1e1d","examples\u002Fevolutionary-optimization\u002F.converge\u002Fplaybooks\u002Fskills\u002Fevolve-evaluate",1,"https:\u002F\u002Fgithub.com\u002Fopenplaybooks-dev\u002Fconverge","https:\u002F\u002Fplaybooks.to","https:\u002F\u002Fgithub.com\u002Fopenplaybooks-dev","2026-06-13 10:11:21",{"source":7,"skillName":5,"commands":49,"prompt":59,"zipAvailable":60},[50,53,56],{"tool":51,"command":52},"npx","npx skills add openplaybooks-dev\u002Fconverge --skill evolve-evaluate",{"tool":54,"command":55},"bunx","bunx skills add openplaybooks-dev\u002Fconverge --skill evolve-evaluate",{"tool":57,"command":58},"pnpm","pnpm dlx skills add openplaybooks-dev\u002Fconverge --skill evolve-evaluate","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fopenplaybooks-dev\u002Fconverge\u002Fblob\u002Fmain\u002Fexamples\u002Fevolutionary-optimization\u002F.converge\u002Fplaybooks\u002Fskills\u002Fevolve-evaluate\u002FSKILL.md",true,[62],{"path":63,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":66,"sourceHtmlUrl":67},"check.js","SCRIPT","脚本",2041,"https:\u002F\u002Fgithub.com\u002Fopenplaybooks-dev\u002Fconverge\u002Fblob\u002Fmain\u002Fexamples\u002Fevolutionary-optimization\u002F.converge\u002Fplaybooks\u002Fskills\u002Fevolve-evaluate\u002Fcheck.js",[69,96,105,127,152,180],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":8,"repositoryNo":11,"starsCount":12,"licenseSpdxId":13,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":80,"publishedAt":29},"01M71989MQF3GPZ7P5TX6S93PD","hello-checker","Use this skill when asked to \"verify\", \"check\", or \"validate\" a file's existence and content. It provides a reliable workflow for confirming files meet their spec.","https:\u002F\u002Fgithub.com\u002Fopenplaybooks-dev\u002Fconverge\u002Fblob\u002Fmain\u002Ftests\u002Ftest-codex-real\u002F.converge\u002Fskills\u002Fhello-checker\u002FSKILL.md",[75],{"categoryNo":76,"categoryName":77,"taxonomyType":78,"taxonomyTypeDesc":79},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[81,82,83,84,85,86,87,88,89,22,90,91,92,93,94,95],"checker","hello checker","hello","it","use","content","validate","spec","verify","file","workflow","when","Meet","files","for",{"skillNo":97,"name":98,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":8,"repositoryNo":11,"starsCount":12,"licenseSpdxId":13,"licenseFileUrl":6,"sourceHtmlUrl":99,"categories":100,"keywords":102,"publishedAt":29},"01M7198823W79B3S15TPV97TCB","ask-skill","https:\u002F\u002Fgithub.com\u002Fopenplaybooks-dev\u002Fconverge\u002Fblob\u002Fmain\u002Ftests\u002Ftest-ask-autonomous\u002F.converge\u002Fplaybooks\u002Fdefault\u002Fskills\u002Fask-skill\u002FSKILL.md",[101],{"categoryNo":76,"categoryName":77,"taxonomyType":78,"taxonomyTypeDesc":79},[103,104],"ask skill","ask",{"skillNo":106,"name":107,"nameZh":6,"description":108,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":8,"repositoryNo":11,"starsCount":12,"licenseSpdxId":13,"licenseFileUrl":6,"sourceHtmlUrl":109,"categories":110,"keywords":115,"publishedAt":29},"01M71986EJ6VCM8YKGEM2YEDQV","html-review-artifact","Framework-wide guidance for crafting polished human-review artifacts as markdown or HTML","https:\u002F\u002Fgithub.com\u002Fopenplaybooks-dev\u002Fconverge\u002Fblob\u002Fmain\u002Fskills\u002Fhtml-review-artifact\u002FSKILL.md",[111,114],{"categoryNo":112,"categoryName":113,"taxonomyType":78,"taxonomyTypeDesc":79},"01JBSEEDCAT000000000000009","前端开发",{"categoryNo":76,"categoryName":77,"taxonomyType":78,"taxonomyTypeDesc":79},[116,117,118,119,120,95,121,122,123,124,125,126],"html","artifact","review","html review artifact","as","framework","wide","guidance","markdown","human","artifacts",{"skillNo":128,"name":129,"nameZh":6,"description":130,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":8,"repositoryNo":11,"starsCount":12,"licenseSpdxId":13,"licenseFileUrl":6,"sourceHtmlUrl":131,"categories":132,"keywords":134,"publishedAt":29},"01M719833M2XWXG1VPBW30GW7J","converge-planning","Design a Converge playbook end-to-end. Use when the user says \"plan a project\", \"design a playbook\", \"decompose into tasks\", \"scaffold a Converge workflow\", \"onboard to Converge\", or asks how to write TASK.md \u002F playbook.yml. Two entry points: plan a new playbook or update an existing one. Hand off to `\u002Fconverge-control` after producing the playbook structure.","https:\u002F\u002Fgithub.com\u002Fopenplaybooks-dev\u002Fconverge\u002Fblob\u002Fmain\u002Fskills\u002Fconverge-planning\u002FSKILL.md",[133],{"categoryNo":76,"categoryName":77,"taxonomyType":78,"taxonomyTypeDesc":79},[135,136,137,138,91,139,85,140,141,142,143,144,145,146,147,148,149,150,151,92],"converge","converge planning","planning","structure","tasks","decompose","hand","write","plan","how","update","playbook","new","design","how to","end","project",{"skillNo":153,"name":154,"nameZh":6,"description":155,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":8,"repositoryNo":11,"starsCount":12,"licenseSpdxId":13,"licenseFileUrl":6,"sourceHtmlUrl":156,"categories":157,"keywords":164,"publishedAt":29},"01M7197XZT1T9GDYAJGVGX1J60","converge-control","Operate a Converge playbook end-to-end — start runs, monitor progress, narrow failures with `--select`, inspect runtime state, validate definitions, surgically reset subtrees, and stop stuck or orphaned runs. Use this skill whenever the user says \"run my playbook\", \"execute this\", \"monitor the run\", \"converge failed\", \"the run is stuck\", \"retry failures\", \"babysit this playbook\", \"what's the status?\", \"show me the gantt\", or asks how to use `converge run` \u002F `--select` \u002F `inspect` \u002F `doctor` \u002F `clean` \u002F `stop`. Also use it for `HTTP 401` \u002F `Invalid API key` failures on the first task — those are environment-not-playbook problems — and for any post-run cleanup work. Hand off to `converge-planning` if the user wants to design or restructure the playbook itself.","https:\u002F\u002Fgithub.com\u002Fopenplaybooks-dev\u002Fconverge\u002Fblob\u002Fmain\u002Fskills\u002Fconverge-control\u002FSKILL.md",[158,159],{"categoryNo":76,"categoryName":77,"taxonomyType":78,"taxonomyTypeDesc":79},{"categoryNo":160,"categoryName":161,"taxonomyType":162,"taxonomyTypeDesc":163},"01JBSEEDCAT000000000000030","开发工程师","PROFESSION","适用职业",[165,166,135,167,168,169,170,171,172,173,141,174,175,148,149,176,150,177,178,179],"converge control","control","operate","invalid","show","api","me","work","reset","show me","cleanup","clean","inspect","select","retry",{"skillNo":181,"name":182,"nameZh":6,"description":183,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":8,"repositoryNo":11,"starsCount":12,"licenseSpdxId":13,"licenseFileUrl":6,"sourceHtmlUrl":184,"categories":185,"keywords":187,"publishedAt":29},"01M7197SCVN0QTD3CM4M83WYJH","repair-taskts","Fix task.ts definition issues — Seed syntax, missing .build(), empty checks, import errors, TypeScript errors","https:\u002F\u002Fgithub.com\u002Fopenplaybooks-dev\u002Fconverge\u002Fblob\u002Fmain\u002Fpackages\u002Fcore\u002Fsrc\u002Fnavigator\u002Frepair\u002Fskill-templates\u002Frepair-taskts\u002FSKILL.md",[186],{"categoryNo":160,"categoryName":161,"taxonomyType":162,"taxonomyTypeDesc":163},[188,189,190,191,142,192,193,194,195,196,197,198,199,200,201,150,202,135,203,204],"repair taskts","repair","Bash","edit","read","line","at","spawn","code","analysis","number","errors","issues","fix","chain","definition","never"]