[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M77XM2PC5G28HNA1QSC7WZX4":3,"skill-same:01M77XM2PC5G28HNA1QSC7WZX4":77},{"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":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":30,"install":31,"resources":44},"01M77XM2PC5G28HNA1QSC7WZX4","vs-eval",null,"Use when asked to write PathGrade cases, static pins, live evals, exclusive contracts, or fixture-backed skill tests. Use when the user types \u002Fvs-eval.","vltansky\u002Fvs","vltansky","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F5851280?v=4","01M77XJAE230PB81N52XJHT5A1","Vlad Tansky","01M77XJAET1CMBAZ2TWTT4ZQT8",0,"MIT","https:\u002F\u002Fgithub.com\u002Fvltansky\u002Fvs\u002Fblob\u002Fmaster\u002Fskills\u002Fvs-eval\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: vs-eval\ndescription: \"Use when asked to write PathGrade cases, static pins, live evals, exclusive contracts, or fixture-backed skill tests. Use when the user types \u002Fvs-eval.\"\n---\n\n# Eval\n\nWrite PathGrade cases that fail a skill which copies slogans and omits the\nbehavior. Building block.\n\n## Flow Contract\n\n- **Kind:** Building block\n- **Inputs:** A skill change, a hypothesized contract, and the existing eval style in this repo\n- **Outputs:** `*.static.eval.ts` and\u002For `*.eval.ts` plus fixtures that can fail; a `scripts\u002Freject-*.mjs` workspace scorer when the contract is fixture content; status of the pin\n- **Status:** PINNED, LIVE, WARN, or BLOCKED\n- **Consumers:** Skill authors; `\u002Fvs-tdd`, `\u002Fvs-shape-it`, `\u002Fvs-build-it` compose only when the unit is a vs skill PathGrade contract (not default phases); `\u002Fvs-try-skill` dogfood; `\u002Fvs-ship-it` before a skill PR\n- **Skip conditions:** Skip when the change is not a skill or eval contract (plain app code with ordinary unit tests)\n\n## Workshop first\n\nAsk 1-3 clarifying questions. Draft one scenario. Write eval code only\nafter the user accepts.\n\nThe live prompt tests skill behavior. It is a problem a real user would\ntype, not implementation steps.\n\nBad: \"read SKILL.md then run X\".\nGood: the request a user would type to trigger the skill.\n\n## Workspace scorer, not slogan pins\n\nA writer following this skill emits `scripts\u002Freject-*.mjs` that reads\nfixture files and scores their content. Same CLI contract as `\u002Fvs-write`\nand `\u002Fvs-deslop` (examples only; do not edit those skills here):\n\n```bash\nnode scripts\u002Freject-*.mjs \u003Cfile-or-pair>\n```\n\nExit 0 clean. Exit 1 reject. Exit 2 cannot check. Treat 2 as not\nchecked, not a pass.\n\nThe static CASE runs that script on fixtures. Do not make a slogan\n`toMatch` the contract.\n\n## Static pin vs live CASE\n\n- Static pin (`*.static.eval.ts`): assert wording the skill must keep,\n  **or** run the workspace scorer on fixtures. No agent and no tokens.\n- Live CASE (`*.eval.ts`): score a transcript with PathGrade\n  `createAgent` (from `vs-internal-shared\u002Ftest\u002Fpathgrade-agent`) and\n  `evaluate` (from `@wix\u002Fpathgrade`). Scope a live file when needed.\n\nDo not write a live CASE for a phrase that SKILL.md can assert. Do not\nwrite a static pin for a first-turn question, a tool call, or a\ntranscript shape. Do not invent a new PathGrade runner.\n\n### Gold, fail-closed, isolation, APIs\n\nDo not invent APIs. Gold stays true if the skill internals change. Score outcome (lint, tests,\nbehavior), not filenames or methods the skill might rename. If a check\nlooks for a file, the user prompt must name it.\n\nEach live CASE needs at least one `check()` or `score()`. Optional\n`judge()` at weight 0.2-0.5. `toolUsage` only when the workflow is mandatory.\nScore outcome first.\n\nAPI: `createAgent`, `prompt` \u002F `runConversation`, `evaluate`, `check`,\n`score`, `judge`, `toolUsage` from `@wix\u002Fpathgrade`. This repo wraps\n`createAgent` via `vs-internal-shared\u002Ftest\u002Fpathgrade-agent`.\n\nIsolation is `createAgent({ skillDir, workspace })` copying the skill\nand fixture into an isolated workspace and HOME. Score that workspace.\nFail-closed: missing evidence fails; `onScorerError: 'zero'` or\n`'fail'`, not `'skip'`. Multi-turn skills use `runConversation`\nplus reactions, not a flattened one-shot.\n\n## Exclusive order\n\n1. Name the required behavior.\n2. Pair every copied slogan with `not.toMatch` on that slogan.\n3. Add the exclusive alternative a slogan-only skill fails (run the\n   reject script on a bad fixture).\n\nA fixture can fail good-mention plus bad-behavior. Put the bad tell in\nthe fixture; put the reject rule in `scripts\u002Freject-*.mjs`. Keep fixture\ncanaries out of SKILL.md.\n\nMatch text across wraps with `\\s+`. Do not relax an assertion to make\nit pass.\n\nWhen a skill already pins `toBeLessThanOrEqual(N)` on SKILL.md line\ncount, stay at or under N. `\u002Fvs-write` uses 240. Do not invent a\nceiling a skill does not already pin.\n\n## PathGrade commands - do not edit PathGrade\n\nPrefer pathgrade run over bare vitest. Do not install PathGrade; vs already has @wix\u002Fpathgrade. Keep the vs wrappers.\n\n```bash\npathgrade analyze --dir skills\u002Fthat-skill\npathgrade validate skills\u002Fthat-skill\u002Ftest\u002Fthat.eval.ts\npathgrade run\nnpm run eval:static\nnpm run eval\nnpm run eval -- skills\u002Fvs-shape-it\u002Ftest\u002Fshape-it.eval.ts\nnpm run eval:preview\n```\n\n`npm run eval:static` is the edit loop. `npm run eval` spends tokens on live\nagents; scope it.\nDo not edit PathGrade: not wix-private\u002Fpathgrade, not `@wix\u002Fpathgrade`, not\nnode_modules\u002F@wix\u002Fpathgrade source, not .pathgrade internals. Change only skill text, fixtures, `scripts\u002Freject-*.mjs`, and\n`*.eval.ts` \u002F `*.static.eval.ts` beside the skill.\n\n## Output style\n\nApply the shared output style at ..\u002Fvs-internal-shared\u002Freferences\u002Foutput-style.md to every user-facing message.\n\n## Workflow\n\nDirect: emit **Next** only. Composed: return to caller.\n\n**Prev:** skill change | failing pin\n**Next:** done\n**Relevant:** `\u002Fvs-tdd` | `\u002Fvs-ship-it`\n\n","# Eval\n\nWrite PathGrade cases that fail a skill which copies slogans and omits the\nbehavior. Building block.\n\n## Flow Contract\n\n- **Kind:** Building block\n- **Inputs:** A skill change, a hypothesized contract, and the existing eval style in this repo\n- **Outputs:** `*.static.eval.ts` and\u002For `*.eval.ts` plus fixtures that can fail; a `scripts\u002Freject-*.mjs` workspace scorer when the contract is fixture content; status of the pin\n- **Status:** PINNED, LIVE, WARN, or BLOCKED\n- **Consumers:** Skill authors; `\u002Fvs-tdd`, `\u002Fvs-shape-it`, `\u002Fvs-build-it` compose only when the unit is a vs skill PathGrade contract (not default phases); `\u002Fvs-try-skill` dogfood; `\u002Fvs-ship-it` before a skill PR\n- **Skip conditions:** Skip when the change is not a skill or eval contract (plain app code with ordinary unit tests)\n\n## Workshop first\n\nAsk 1-3 clarifying questions. Draft one scenario. Write eval code only\nafter the user accepts.\n\nThe live prompt tests skill behavior. It is a problem a real user would\ntype, not implementation steps.\n\nBad: \"read SKILL.md then run X\".\nGood: the request a user would type to trigger the skill.\n\n## Workspace scorer, not slogan pins\n\nA writer following this skill emits `scripts\u002Freject-*.mjs` that reads\nfixture files and scores their content. Same CLI contract as `\u002Fvs-write`\nand `\u002Fvs-deslop` (examples only; do not edit those skills here):\n\n```bash\nnode scripts\u002Freject-*.mjs \u003Cfile-or-pair>\n```\n\nExit 0 clean. Exit 1 reject. Exit 2 cannot check. Treat 2 as not\nchecked, not a pass.\n\nThe static CASE runs that script on fixtures. Do not make a slogan\n`toMatch` the contract.\n\n## Static pin vs live CASE\n\n- Static pin (`*.static.eval.ts`): assert wording the skill must keep,\n  **or** run the workspace scorer on fixtures. No agent and no tokens.\n- Live CASE (`*.eval.ts`): score a transcript with PathGrade\n  `createAgent` (from `vs-internal-shared\u002Ftest\u002Fpathgrade-agent`) and\n  `evaluate` (from `@wix\u002Fpathgrade`). Scope a live file when needed.\n\nDo not write a live CASE for a phrase that SKILL.md can assert. Do not\nwrite a static pin for a first-turn question, a tool call, or a\ntranscript shape. Do not invent a new PathGrade runner.\n\n### Gold, fail-closed, isolation, APIs\n\nDo not invent APIs. Gold stays true if the skill internals change. Score outcome (lint, tests,\nbehavior), not filenames or methods the skill might rename. If a check\nlooks for a file, the user prompt must name it.\n\nEach live CASE needs at least one `check()` or `score()`. Optional\n`judge()` at weight 0.2-0.5. `toolUsage` only when the workflow is mandatory.\nScore outcome first.\n\nAPI: `createAgent`, `prompt` \u002F `runConversation`, `evaluate`, `check`,\n`score`, `judge`, `toolUsage` from `@wix\u002Fpathgrade`. This repo wraps\n`createAgent` via `vs-internal-shared\u002Ftest\u002Fpathgrade-agent`.\n\nIsolation is `createAgent({ skillDir, workspace })` copying the skill\nand fixture into an isolated workspace and HOME. Score that workspace.\nFail-closed: missing evidence fails; `onScorerError: 'zero'` or\n`'fail'`, not `'skip'`. Multi-turn skills use `runConversation`\nplus reactions, not a flattened one-shot.\n\n## Exclusive order\n\n1. Name the required behavior.\n2. Pair every copied slogan with `not.toMatch` on that slogan.\n3. Add the exclusive alternative a slogan-only skill fails (run the\n   reject script on a bad fixture).\n\nA fixture can fail good-mention plus bad-behavior. Put the bad tell in\nthe fixture; put the reject rule in `scripts\u002Freject-*.mjs`. Keep fixture\ncanaries out of SKILL.md.\n\nMatch text across wraps with `\\s+`. Do not relax an assertion to make\nit pass.\n\nWhen a skill already pins `toBeLessThanOrEqual(N)` on SKILL.md line\ncount, stay at or under N. `\u002Fvs-write` uses 240. Do not invent a\nceiling a skill does not already pin.\n\n## PathGrade commands - do not edit PathGrade\n\nPrefer pathgrade run over bare vitest. Do not install PathGrade; vs already has @wix\u002Fpathgrade. Keep the vs wrappers.\n\n```bash\npathgrade analyze --dir skills\u002Fthat-skill\npathgrade validate skills\u002Fthat-skill\u002Ftest\u002Fthat.eval.ts\npathgrade run\nnpm run eval:static\nnpm run eval\nnpm run eval -- skills\u002Fvs-shape-it\u002Ftest\u002Fshape-it.eval.ts\nnpm run eval:preview\n```\n\n`npm run eval:static` is the edit loop. `npm run eval` spends tokens on live\nagents; scope it.\nDo not edit PathGrade: not wix-private\u002Fpathgrade, not `@wix\u002Fpathgrade`, not\nnode_modules\u002F@wix\u002Fpathgrade source, not .pathgrade internals. Change only skill text, fixtures, `scripts\u002Freject-*.mjs`, and\n`*.eval.ts` \u002F `*.static.eval.ts` beside the skill.\n\n## Output style\n\nApply the shared output style at ..\u002Fvs-internal-shared\u002Freferences\u002Foutput-style.md to every user-facing message.\n\n## Workflow\n\nDirect: emit **Next** only. Composed: return to caller.\n\n**Prev:** skill change | failing pin\n**Next:** done\n**Relevant:** `\u002Fvs-tdd` | `\u002Fvs-ship-it`\n\n",{"name":5,"description":7},"master","72448991571c9e5f0e76cccf189b8522d442e6eb","skills\u002Fvs-eval",1,"https:\u002F\u002Fgithub.com\u002Fvltansky\u002Fvs","https:\u002F\u002Fgithub.com\u002Fvltansky","2026-11-06 18:47:54",{"source":8,"skillName":5,"commands":32,"prompt":42,"zipAvailable":43},[33,36,39],{"tool":34,"command":35},"npx","npx skills add vltansky\u002Fvs --skill vs-eval",{"tool":37,"command":38},"bunx","bunx skills add vltansky\u002Fvs --skill vs-eval",{"tool":40,"command":41},"pnpm","pnpm dlx skills add vltansky\u002Fvs --skill vs-eval","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fvltansky\u002Fvs\u002Fblob\u002Fmaster\u002Fskills\u002Fvs-eval\u002FSKILL.md",true,[45,51,57,61,65,69,73],{"path":46,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":49,"sourceHtmlUrl":50},"agents\u002Fopenai.yaml","REFERENCE","参考文档",173,"https:\u002F\u002Fgithub.com\u002Fvltansky\u002Fvs\u002Fblob\u002Fmaster\u002Fskills\u002Fvs-eval\u002Fagents\u002Fopenai.yaml",{"path":52,"resourceType":53,"resourceTypeDesc":54,"sizeBytes":55,"sourceHtmlUrl":56},"scripts\u002Freject-mention-only.mjs","SCRIPT","脚本",2672,"https:\u002F\u002Fgithub.com\u002Fvltansky\u002Fvs\u002Fblob\u002Fmaster\u002Fskills\u002Fvs-eval\u002Fscripts\u002Freject-mention-only.mjs",{"path":58,"resourceType":53,"resourceTypeDesc":54,"sizeBytes":59,"sourceHtmlUrl":60},"test\u002Ffixtures\u002Fclean-exclusive.static.eval.ts",518,"https:\u002F\u002Fgithub.com\u002Fvltansky\u002Fvs\u002Fblob\u002Fmaster\u002Fskills\u002Fvs-eval\u002Ftest\u002Ffixtures\u002Fclean-exclusive.static.eval.ts",{"path":62,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":63,"sourceHtmlUrl":64},"test\u002Ffixtures\u002Fclean-skill.md",645,"https:\u002F\u002Fgithub.com\u002Fvltansky\u002Fvs\u002Fblob\u002Fmaster\u002Fskills\u002Fvs-eval\u002Ftest\u002Ffixtures\u002Fclean-skill.md",{"path":66,"resourceType":53,"resourceTypeDesc":54,"sizeBytes":67,"sourceHtmlUrl":68},"test\u002Ffixtures\u002Fmention-only.static.eval.ts",244,"https:\u002F\u002Fgithub.com\u002Fvltansky\u002Fvs\u002Fblob\u002Fmaster\u002Fskills\u002Fvs-eval\u002Ftest\u002Ffixtures\u002Fmention-only.static.eval.ts",{"path":70,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":71,"sourceHtmlUrl":72},"test\u002Ffixtures\u002Fslogan-only-skill.md",396,"https:\u002F\u002Fgithub.com\u002Fvltansky\u002Fvs\u002Fblob\u002Fmaster\u002Fskills\u002Fvs-eval\u002Ftest\u002Ffixtures\u002Fslogan-only-skill.md",{"path":74,"resourceType":53,"resourceTypeDesc":54,"sizeBytes":75,"sourceHtmlUrl":76},"test\u002Fvs-eval.static.eval.ts",3219,"https:\u002F\u002Fgithub.com\u002Fvltansky\u002Fvs\u002Fblob\u002Fmaster\u002Fskills\u002Fvs-eval\u002Ftest\u002Fvs-eval.static.eval.ts",[78,85,91,96,103,108],{"skillNo":79,"name":80,"nameZh":6,"description":81,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":19},"01M77XSKWFMPKFXM0311QKEJHH","vs-write","Use when asked to write, rewrite, edit, tighten, simplify, or polish prose for clarity, or to put text in STE, ASD-STE100, or controlled language. Produces plain, direct, accessible copy without losing substance or precision.","https:\u002F\u002Fgithub.com\u002Fvltansky\u002Fvs\u002Fblob\u002Fmaster\u002Fskills\u002Fvs-write\u002FSKILL.md",[],[],{"skillNo":86,"name":87,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":88,"categories":89,"keywords":90,"publishedAt":19},"01M77XSJNX2V560HSKQKQD67KP","vs-fake-verify","https:\u002F\u002Fgithub.com\u002Fvltansky\u002Fvs\u002Fblob\u002Fmaster\u002Fskills\u002Fvs-verify\u002Ftest\u002Ffixtures\u002Fpath-end-state\u002Fstub-rejector\u002FSKILL.md",[],[],{"skillNo":92,"name":87,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":93,"categories":94,"keywords":95,"publishedAt":19},"01M77XSHHBEMMYZ6B4PD384NZH","https:\u002F\u002Fgithub.com\u002Fvltansky\u002Fvs\u002Fblob\u002Fmaster\u002Fskills\u002Fvs-verify\u002Ftest\u002Ffixtures\u002Fpath-end-state\u002Fstructure-paste\u002FSKILL.md",[],[],{"skillNo":97,"name":98,"nameZh":6,"description":99,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":100,"categories":101,"keywords":102,"publishedAt":19},"01M77XSE2Y7MW0KP3DTFNEE2E8","vs-verify","Use when asked to verify, prove it works, check completion, or before claiming done. Produces evidence-backed PASS\u002FWARN\u002FFAIL status.","https:\u002F\u002Fgithub.com\u002Fvltansky\u002Fvs\u002Fblob\u002Fmaster\u002Fskills\u002Fvs-verify\u002Ftest\u002Ffixtures\u002Fpath-end-state\u002Fpublished-pair\u002FSKILL.md",[],[],{"skillNo":104,"name":87,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":105,"categories":106,"keywords":107,"publishedAt":19},"01M77XSAXFQXNE6VCRC7275Z0X","https:\u002F\u002Fgithub.com\u002Fvltansky\u002Fvs\u002Fblob\u002Fmaster\u002Fskills\u002Fvs-verify\u002Ftest\u002Ffixtures\u002Fpath-end-state\u002Fphrase-complete-beside\u002FSKILL.md",[],[],{"skillNo":109,"name":98,"nameZh":6,"description":99,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":110,"categories":111,"keywords":112,"publishedAt":19},"01M77XS6R0E9X1E5KKSS00WYSS","https:\u002F\u002Fgithub.com\u002Fvltansky\u002Fvs\u002Fblob\u002Fmaster\u002Fskills\u002Fvs-verify\u002FSKILL.md",[],[]]