[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6NQ2K3AMGNH0VZX8TQ5VE23":3,"skill-same:01M6NQ2K3AMGNH0VZX8TQ5VE23":71},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":23,"publishedAt":44,"linkOnly":45,"markdownContent":46,"markdownBody":47,"frontMatter":48,"refName":50,"commitSha":51,"skillDirectory":52,"forksCount":14,"repositoryHtmlUrl":53,"repositoryHomepageUrl":6,"ownerHtmlUrl":54,"ownerBlogUrl":55,"ownerDescription":6,"updatedAt":56,"install":57,"resources":70},"01M6NQ2K3AMGNH0VZX8TQ5VE23","agent-skill-kit-description-optimizer",null,"Optimize a SKILL.md description for trigger quality — generate synthetic positive and negative queries, score the current description by hit-rate vs false-positive, and propose three ranked rewrites. Use when the user asks to auto-tune, improve, or rewrite a skill description, fix a skill that never fires, or measure description quality.","carlosmarte\u002Fai-agent-artifacts","carlosmarte","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F3478576?v=4","01M6NPRTEXZG86XJXMGQ05A24X","Carlos Marte","01M6NPRTK4A6A4Q8CMWMH1ZY7W",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fagent-skill-kit-description-optimizer\u002FSKILL.md",[18],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43],"optimizer","agent skill kit description optimizer","kit","agent skill","use","auto","never","queries","optimize","three","cli","fix","app","false","trigger","synthetic","rewrite","improve","for","generate","2026-11-10 10:12:30",false,"---\nname: agent-skill-kit-description-optimizer\ndescription: Optimize a SKILL.md description for trigger quality — generate synthetic positive and negative queries, score the current description by hit-rate vs false-positive, and propose three ranked rewrites. Use when the user asks to auto-tune, improve, or rewrite a skill description, fix a skill that never fires, or measure description quality.\ntier: app\nlicense: Apache-2.0\ncompatibility: requires `skills-ref` CLI\n---\n\n# Agent Skill Description Optimizer\n\nWraps `skills-ref optimize`. Generates trigger queries, scores the current description against the F04 rubric, and proposes three deterministic rewrites. The `--apply N` flag commits the chosen variant back to disk with byte-preservation of every other line.\n\n## When to use\n\n- \"Improve this description\"\n- \"Why doesn't my skill ever fire?\"\n- \"Optimize the SKILL.md descriptions in this tree\"\n- \"Rewrite this description to score higher\"\n- \"Measure description quality before shipping\"\n\n## Where LLM-quality variants come from\n\nThe CLI ships **deterministic** variants only — three template-based transforms, \u003C 200 ms, no network, no API keys, no provider config. This is intentional: when the skill runs inside an AI harness (Claude Code, an SDK agent, etc.), the calling harness *is itself* an LLM with provider credentials already attached. Asking the CLI to hold a separate API key duplicates auth surface for no benefit.\n\nWhen higher-quality rewrites are wanted, the harness does the work:\n\n1. Run `skills-ref optimize \u003Cdir> --baseline-only --format json` to get the current score + breakdown.\n2. The harness reads the SKILL.md and generates its own three improved descriptions, using whatever model\u002Fauth it already has.\n3. The harness writes the chosen description directly into the YAML frontmatter (Edit \u002F Write).\n4. Run `skills-ref validate \u003Cdir>` to confirm the post-edit file passes (the same gate `--apply` uses internally).\n\n`--apply N` remains useful for the deterministic case — it preserves bytes and auto-rolls back on validation failure.\n\n## Invocation\n\n```bash\n# Score the current description only (no rewrites)\nskills-ref optimize \u002Fpath\u002Fto\u002Fskill-dir --baseline-only\n\n# Score + propose three ranked variants (sorted by F04 score desc)\nskills-ref optimize \u002Fpath\u002Fto\u002Fskill-dir\n\n# Apply variant 0 in-place (with re-validation safety + auto-rollback)\nskills-ref optimize \u002Fpath\u002Fto\u002Fskill-dir --apply 0\n\n# Machine-readable output for the CI gate\nskills-ref optimize \u002Fpath\u002Fto\u002Fskill-dir --format json\n```\n\n## Output shape\n\nHuman format reports the baseline and each variant on its own block:\n\n```\n\u002Fabs\u002Fpath\u002FSKILL.md\n  provider: none\n  baseline: score=98\u002F100 hit=6\u002F8 false=1\u002F8\n  variants:\n    [0] score=99\u002F100 hit=5\u002F8 false=1\u002F8\n        rationale: lead-with-verb + append trigger phrase\n        text: \u003Crewritten description>\n```\n\nJSON format is canonical sorted-key output containing `baseline`, `variants`, `queries`, `provider`, `path`. Both runtimes emit byte-identical JSON (parity-tested against the frozen `.agents\u002Fskills\u002Fagent-skill-kit-commons\u002Ffixtures\u002Foptimize\u002Fexpected\u002Flow-score-vague.json` snapshot).\n\n## Two scores per variant\n\n- **`score`** — F04's description-quality score (0–100). The deterministic ranker sorts variants by this.\n- **`hit_rate.score`** — `(positive_hits \u002F positive_total - false_positives \u002F false_total) * 100`, clamped to 0–100. Surfaced for CI gating without re-running F04's scorer.\n\n## Safety contract for `--apply`\n\n- The `description:` line in the YAML frontmatter is replaced; every other line is byte-preserved.\n- A `.bak` file is created before write.\n- After write, the validator runs against the same directory.\n- On validation failure, the backup restores the original and the `.bak` is removed; the CLI exits 1.\n- On success, the `.bak` is deleted; the CLI exits 0.\n\n## Limits\n\n- The deterministic rewriter is intentionally conservative — three template-based transforms, no LLM. For higher-quality variants, see \"Where LLM-quality variants come from\" above; the harness generates them, the CLI does not.\n- The `chars \u002F 4` token estimate from the prompt emitter is not used here; the optimizer's input is bounded by the body excerpt (first 1000 chars) and the query JSON.\n- The trigger-phrase regex re-uses F04's; both share the same edge cases (e.g. \"for \" inside \"helpful for\" matches as a trigger). The lint pipeline is the canonical filter; this skill defers to it.\n\n## Reference\n\nFull specification: see the canonical spec at [..\u002F..\u002F..\u002Fdocs\u002Fagentskills.io.md](..\u002F..\u002F..\u002Fdocs\u002Fagentskills.io.md).\n","# Agent Skill Description Optimizer\n\nWraps `skills-ref optimize`. Generates trigger queries, scores the current description against the F04 rubric, and proposes three deterministic rewrites. The `--apply N` flag commits the chosen variant back to disk with byte-preservation of every other line.\n\n## When to use\n\n- \"Improve this description\"\n- \"Why doesn't my skill ever fire?\"\n- \"Optimize the SKILL.md descriptions in this tree\"\n- \"Rewrite this description to score higher\"\n- \"Measure description quality before shipping\"\n\n## Where LLM-quality variants come from\n\nThe CLI ships **deterministic** variants only — three template-based transforms, \u003C 200 ms, no network, no API keys, no provider config. This is intentional: when the skill runs inside an AI harness (Claude Code, an SDK agent, etc.), the calling harness *is itself* an LLM with provider credentials already attached. Asking the CLI to hold a separate API key duplicates auth surface for no benefit.\n\nWhen higher-quality rewrites are wanted, the harness does the work:\n\n1. Run `skills-ref optimize \u003Cdir> --baseline-only --format json` to get the current score + breakdown.\n2. The harness reads the SKILL.md and generates its own three improved descriptions, using whatever model\u002Fauth it already has.\n3. The harness writes the chosen description directly into the YAML frontmatter (Edit \u002F Write).\n4. Run `skills-ref validate \u003Cdir>` to confirm the post-edit file passes (the same gate `--apply` uses internally).\n\n`--apply N` remains useful for the deterministic case — it preserves bytes and auto-rolls back on validation failure.\n\n## Invocation\n\n```bash\n# Score the current description only (no rewrites)\nskills-ref optimize \u002Fpath\u002Fto\u002Fskill-dir --baseline-only\n\n# Score + propose three ranked variants (sorted by F04 score desc)\nskills-ref optimize \u002Fpath\u002Fto\u002Fskill-dir\n\n# Apply variant 0 in-place (with re-validation safety + auto-rollback)\nskills-ref optimize \u002Fpath\u002Fto\u002Fskill-dir --apply 0\n\n# Machine-readable output for the CI gate\nskills-ref optimize \u002Fpath\u002Fto\u002Fskill-dir --format json\n```\n\n## Output shape\n\nHuman format reports the baseline and each variant on its own block:\n\n```\n\u002Fabs\u002Fpath\u002FSKILL.md\n  provider: none\n  baseline: score=98\u002F100 hit=6\u002F8 false=1\u002F8\n  variants:\n    [0] score=99\u002F100 hit=5\u002F8 false=1\u002F8\n        rationale: lead-with-verb + append trigger phrase\n        text: \u003Crewritten description>\n```\n\nJSON format is canonical sorted-key output containing `baseline`, `variants`, `queries`, `provider`, `path`. Both runtimes emit byte-identical JSON (parity-tested against the frozen `.agents\u002Fskills\u002Fagent-skill-kit-commons\u002Ffixtures\u002Foptimize\u002Fexpected\u002Flow-score-vague.json` snapshot).\n\n## Two scores per variant\n\n- **`score`** — F04's description-quality score (0–100). The deterministic ranker sorts variants by this.\n- **`hit_rate.score`** — `(positive_hits \u002F positive_total - false_positives \u002F false_total) * 100`, clamped to 0–100. Surfaced for CI gating without re-running F04's scorer.\n\n## Safety contract for `--apply`\n\n- The `description:` line in the YAML frontmatter is replaced; every other line is byte-preserved.\n- A `.bak` file is created before write.\n- After write, the validator runs against the same directory.\n- On validation failure, the backup restores the original and the `.bak` is removed; the CLI exits 1.\n- On success, the `.bak` is deleted; the CLI exits 0.\n\n## Limits\n\n- The deterministic rewriter is intentionally conservative — three template-based transforms, no LLM. For higher-quality variants, see \"Where LLM-quality variants come from\" above; the harness generates them, the CLI does not.\n- The `chars \u002F 4` token estimate from the prompt emitter is not used here; the optimizer's input is bounded by the body excerpt (first 1000 chars) and the query JSON.\n- The trigger-phrase regex re-uses F04's; both share the same edge cases (e.g. \"for \" inside \"helpful for\" matches as a trigger). The lint pipeline is the canonical filter; this skill defers to it.\n\n## Reference\n\nFull specification: see the canonical spec at [..\u002F..\u002F..\u002Fdocs\u002Fagentskills.io.md](..\u002F..\u002F..\u002Fdocs\u002Fagentskills.io.md).\n",{"name":5,"tier":36,"license":15,"description":7,"compatibility":49},"requires `skills-ref` CLI","main","b78f8878258c912b66be9d4b56b316515e44c728",".claude\u002Fskills\u002Fagent-skill-kit-description-optimizer","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts","https:\u002F\u002Fgithub.com\u002Fcarlosmarte","http:\u002F\u002Fcarlosmarte.me\u002F","2026-08-25 00:53:46",{"source":8,"skillName":5,"commands":58,"prompt":68,"zipAvailable":69},[59,62,65],{"tool":60,"command":61},"npx","npx skills add carlosmarte\u002Fai-agent-artifacts --skill agent-skill-kit-description-optimizer",{"tool":63,"command":64},"bunx","bunx skills add carlosmarte\u002Fai-agent-artifacts --skill agent-skill-kit-description-optimizer",{"tool":66,"command":67},"pnpm","pnpm dlx skills add carlosmarte\u002Fai-agent-artifacts --skill agent-skill-kit-description-optimizer","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fagent-skill-kit-description-optimizer\u002FSKILL.md",true,[],[72,100,126,161,186,208],{"skillNo":73,"name":74,"nameZh":6,"description":75,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":76,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":44},"01M6NQ7MN7FKWQE3CKRWKVAC9K","validation_handoff","Forces the agent to conclude its response by generating a strict validation checklist and outlining pending actions, focusing heavily on defensive execution and edge-case testing.","MIT","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fvalidation_handoff\u002FSKILL.md",[],[80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99],"handoff","validation","validation handoff","testing","architecture","edge","actions","refactor","execution","system","implementation","checklist","feature implementation","feature","end","case","defensive","hook","at","pending",{"skillNo":101,"name":102,"nameZh":6,"description":103,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":76,"licenseFileUrl":6,"sourceHtmlUrl":104,"categories":105,"keywords":106,"publishedAt":44},"01M6NQ74C8EKMSABDNN4T6TFBD","project-ployglot-project-scafford","Scaffold a polyglot library project as TypeScript + Python twin packages under `packages\u002Fts\u002F` + `packages\u002Fpy\u002F`, with a shared `SPEC.md`, cross-language `tests\u002Fparity\u002Ffixtures.json`, side-by-side `examples\u002Fsdk\u002F` + `examples\u002Fapi\u002F` documentation, per-package `Makefile`, and a root orchestrator that exposes `make ci`. Open-source friendly — no organization-specific paths or scopes baked in. Use when starting a new polyglot library, porting a single-language library to a twin, or auditing\u002Fregularizing an existing two-language repo against the canonical layout.","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fproject-ployglot-project-scafford\u002FSKILL.md",[],[107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,37,124,125],"project","project ployglot project scafford","documentation","packages","run","ci","api","open","makefile","Bash","read","orchestrator","only","organization","shared","layout","dry","scope","\u003Cslug>",{"skillNo":127,"name":128,"nameZh":6,"description":129,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":76,"licenseFileUrl":6,"sourceHtmlUrl":130,"categories":131,"keywords":146,"publishedAt":44},"01M6NQ6VJVPM37YA73T8BCD3VF","playwright-ui-locator","Auto-heal Playwright tests after a frontend developer changes a div's class, id, or other selector. Catches a failing Playwright test, captures the old vs new DOM tree, finds the element that drifted, derives the most stable replacement locator, and proposes the fix as a GitHub pull request (confident heals) or files a GitHub issue (ambiguous \u002F element-removed cases) — never editing the working tree directly. Trigger when Playwright tests fail with \"locator resolved to 0 elements\", \"Timeout waiting for selector\", or after a UI refactor \u002F class rename \u002F Tailwind cleanup broke previously-passing tests.","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002FSKILL.md",[132,135,138,141],{"categoryNo":133,"categoryName":134,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000007","代码审查",{"categoryNo":136,"categoryName":137,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000008","测试与质量",{"categoryNo":139,"categoryName":140,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000009","前端开发",{"categoryNo":142,"categoryName":143,"taxonomyType":144,"taxonomyTypeDesc":145},"01JBSEEDCAT000000000000032","设计师","PROFESSION","适用职业",[147,148,149,150,151,152,116,117,153,154,35,155,29,30,156,157,158,37,159,160],"playwright","ui","playwright ui locator","frontend","cleanup","Tailwind","write","edit","snapshot","issue","files","rename","element","request",{"skillNo":162,"name":163,"nameZh":6,"description":164,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":76,"licenseFileUrl":6,"sourceHtmlUrl":165,"categories":166,"keywords":168,"publishedAt":44},"01M6NQ6BCW05C6Y6VRV0RBQD1V","org-learn-context-window","Distill the current conversation into a reusable command or skill artifact. Use when the user wants to capture a workflow they just executed (create a reusable slash command) or formalize a resolution\u002Fmistake\u002Fcorrection into a reusable skill so the same class of issue never recurs.","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Forg-learn-context-window\u002FSKILL.md",[167],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[169,170,171,172,173,174,175,176,177,178,179,180,28,30,156,181,182,183,184,185],"learn","context","org","window","org learn context window","workflow","create","just","capture","distill","command","user","reusable","conversation","artifact","resolution","skill artifact",{"skillNo":187,"name":188,"nameZh":6,"description":189,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":76,"licenseFileUrl":6,"sourceHtmlUrl":190,"categories":191,"keywords":193,"publishedAt":44},"01M6NQ66NFPQM1PDBHP1B2G82X","org-agent-skill-creator","Scaffold and validate scope-aware agent skills (company, enterprise, org, team, project, application). Use when the user asks to create a new agent skill, generate a SKILL.md file, scaffold a skill directory, or validate that existing skill directories meet SKILL_SPEC.md (frontmatter, naming rules, structure, scope\u002Ftier).","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Forg-agent-skill-creator\u002FSKILL.md",[192],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[194,27,171,195,196,197,175,198,199,200,201,28,202,203,107,124,204,205,206,207],"skill creator","agent skill creator","creator","org agent skill creator","file","Enterprise","team","Rules","structure","new agent","directories","Existing","company","Meet",{"skillNo":209,"name":210,"nameZh":6,"description":211,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":76,"licenseFileUrl":6,"sourceHtmlUrl":212,"categories":213,"keywords":217,"publishedAt":44},"01M6NQ5VSH5M682YMYDF9H9KFE","app-workshop-session-pdf-recap-next","Extract content from a workshop session PDF and produce a structured recap focused on (1) what was learned, (2) next steps, and (3) possible enhancements. Ingests a single PDF file path, parses session material (slides, transcripts, exercises, Q&A), and emits a Markdown report with Learning Recap, Next Steps, and Enhancement Opportunities sections with page-level traceability. Use when the user asks to \"summarize a workshop\", \"recap a training PDF\", \"pull next steps from a session deck\", or provides a workshop\u002Fseminar\u002Ftutorial PDF and wants a learn-and-act summary.","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fapp-workshop-session-pdf-recap-next\u002FSKILL.md",[214],{"categoryNo":215,"categoryName":216,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000002","文档处理",[218,219,220,221,36,222,223,169,224,225,226,227,228,229,28,230,37,231,232,233],"PDF","recap","session","next","app workshop session pdf recap next","training","pull","produce","learning","slides","page","tutorial","deck","material","traceability","next steps"]