[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7594AXE3AGRVYQDBBA50MFS":3,"skill-same:01M7594AXE3AGRVYQDBBA50MFS":54},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":13,"repositoryHtmlUrl":27,"repositoryHomepageUrl":6,"ownerHtmlUrl":28,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":29,"install":30,"resources":43},"01M7594AXE3AGRVYQDBBA50MFS","healing-test-failures",null,"Diagnose a failing test before touching production code — reproduce, classify, minimize, fix with a strict two-cycle cap per error signature, and escalate with evidence when the cap is hit.","JackSmack1971\u002Ffreeforge-chat","JackSmack1971","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F128181597?v=4","01M6KS1FH32N6WNWZ68JC9FFDJ","01M7593JRZ0MV9B0VVTRGWZ0WC",0,"MIT","https:\u002F\u002Fgithub.com\u002FJackSmack1971\u002Ffreeforge-chat\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fhealing-test-failures\u002FSKILL.md",[],[],"2026-11-12 13:55:43",false,"---\nname: healing-test-failures\ndescription: Diagnose a failing test before touching production code — reproduce, classify, minimize, fix with a strict two-cycle cap per error signature, and escalate with evidence when the cap is hit.\ncompatibility: Requires Node.js 22+ and the repository's `node:test` suite under `tests\u002Fsecurity\u002F`.\n---\n\n# Healing Test Failures\n\n## Core principle\n\n**Diagnose before changing production code.** A test failure is evidence, not an\ninstruction to edit the nearest file. Do not touch `freeforge\u002Fsrc\u002F` until you know\nwhether the defect is in that source, in the test itself, or in the environment\nrunning it.\n\n## Required workflow\n\nComplete each step in order. Do not skip ahead to remediation.\n\n### 1. Reproduce the exact failure\n\nRun the failing command exactly as reported — do not paraphrase it or swap in a\nbroader command yet.\n\n```bash\nnode --test tests\u002Fsecurity\u002F\u003Cname>.test.mjs\n```\n\nCapture the raw stdout\u002Fstderr verbatim. Record the current commit (`git rev-parse\nHEAD`) as the baseline SHA before making any change. If the failure does not\nreproduce on the first try, run it 2-3 more times before concluding it is flaky —\n`node:test` failures in this repo are almost always deterministic (see\n`tests\u002FAGENTS.md`); an inconsistent result is itself a finding (see step 2,\nenvironment\u002Ftooling).\n\n### 2. Classify the failure\n\nDecide which bucket the failure belongs to before writing any fix:\n\n- **Production defect** — `freeforge\u002Fsrc\u002F*` produces behavior that contradicts its\n  documented\u002Ftested contract (e.g. a DOM mutation, storage key, or sanitize step\n  described in `AGENTS.md` \u002F `tests\u002FAGENTS.md` is missing or wrong).\n- **Test defect** — the test asserts something the contract does not actually\n  require, uses a stale fixture, or misuses `tests\u002Fhelpers\u002Fmock-dom.mjs` (e.g.\n  missing `restore()`, missing `resetState(S)`, asserting on a global that was\n  never installed).\n- **Environment\u002Ftooling defect** — failure depends on Node version, run order,\n  leaked global state, or timing, not on the test's assertions being wrong. In\n  this codebase this almost always traces back to one of the invariants in\n  `tests\u002FAGENTS.md` §\"Invariants Every Test Must Preserve\": unrestored globals,\n  stale `state.S`, an unresolved microtask, or a stray network\u002Fstorage call.\n\nOnly \"production defect\" justifies editing `freeforge\u002Fsrc\u002F`. \"Test defect\" is\nfixed in `tests\u002Fsecurity\u002F`. \"Environment\u002Ftooling\" is fixed in the test's setup\n(`installGlobals`\u002F`restore`\u002F`resetState`\u002Fawaited microtask), not in application\ncode.\n\n### 3. Minimize to the smallest failing case\n\nNarrow to the single failing `test()` block, not the whole file:\n\n```bash\nnode --test --test-name-pattern=\"\u003Cexact test name>\" tests\u002Fsecurity\u002F\u003Cname>.test.mjs\n```\n\nIf the failure only appears when the full suite runs (`npm --prefix freeforge\ntest`) but not in isolation, that itself is diagnostic: it points at shared\nstate leaking across tests (unrestored global, un-reset `state.S`, or import\ncaching — see `importFresh` vs `importShared` in `tests\u002FAGENTS.md`), i.e. an\nenvironment\u002Ftooling defect, not the isolated test's logic.\n\n### 4. Inspect the relevant source and test contract\n\nBefore writing a fix, read:\n\n- The `freeforge\u002Fsrc\u002F` module(s) the test exercises.\n- The test file itself, especially its `installGlobals(...)` call, whether it\n  wraps in `try\u002Ffinally { restore(); }`, whether it calls `resetState(S)` before\n  `importFresh('freeforge\u002Fsrc\u002Fapp.js')`, and whether it awaits a microtask\n  (`await new Promise(r => setTimeout(r, 0))`) after dispatching an event that\n  triggers async work (stream completion, model load).\n- `tests\u002Fhelpers\u002Fmock-dom.mjs` for the exact shape of `MockDocument`,\n  `MemoryStorage`, and `makeFetchResponse` if the failure touches DOM, storage,\n  or fetch mocking.\n- The nearest `AGENTS.md` (root, then `tests\u002FAGENTS.md`) for the invariant the\n  test is meant to enforce (no real network, no real storage, sanitize-before-\n  `innerHTML`, sessionStorage-only key storage, etc.).\n\nDo not guess the contract from the test name alone.\n\n### 5. Remediate — at most two targeted cycles per error signature\n\nAn \"error signature\" is the specific assertion message \u002F stack frame \u002F thrown\nerror, not just \"the test is red.\" For a given error signature:\n\n- **Cycle 1**: form one hypothesis from steps 1-4, make the smallest change that\n  tests it (one file, one root cause — no bundled cleanup), rerun the narrowed\n  command from step 3.\n- **Cycle 2** (only if cycle 1's rerun still shows the *same* error signature):\n  form a new hypothesis informed by cycle 1's result, make one more targeted\n  change, rerun.\n- If the *same* error signature appears a third time: **stop**. Do not attempt a\n  third remediation. Produce the escalation payload (below) instead.\n\nA different error signature after a fix is progress, not a repeat — it starts a\nfresh two-cycle budget for the new signature, but note in your report that the\nsymptom changed.\n\n### 6. Rerun focused validation after each cycle\n\nAfter every remediation attempt, rerun only the narrowed command from step 3\nfirst. Do not broaden scope while the focused case is still red.\n\n### 7. Broaden only after focused success\n\nOnce the narrowed test\u002Ftest-name-pattern passes, broaden in this order and stop\nif any step fails (treat a new failure as its own diagnosis, not a rubber-stamp\nretry):\n\n```bash\nnode --test tests\u002Fsecurity\u002F\u003Cname>.test.mjs      # whole file\nnpm --prefix freeforge test                      # full suite\nnpx --yes @biomejs\u002Fbiome@1.9.4 check freeforge\u002Fsrc tests\u002Fsecurity tests\u002Fhelpers\n```\n\nBoth commands are listed as the repository's \"Full pre-merge verification\" in\n`AGENTS.md` — run them before calling the fix done.\n\n### 8. Stop on repeated identical failure — produce the escalation payload\n\nIf step 5's cap is hit (same error signature survives two targeted remediation\ncycles), stop editing and assemble the escalation payload below instead of\nattempting a third fix.\n\n## Escalation payload\n\nProduce exactly these fields, backed by evidence gathered in steps 1-7 — do not\nsummarize from memory:\n\n- **Baseline SHA** — `git rev-parse HEAD` captured in step 1, before any edits.\n- **Exact command** — the precise command from step 1 or step 3 that fails.\n- **Raw stdout\u002Fstderr** — verbatim output (or a faithful excerpt if very long —\n  keep the assertion diff and stack trace, trim repeated boilerplate).\n- **Error signature** — the specific assertion\u002Fexception that repeated\n  identically across both remediation cycles.\n- **Files changed** — every file touched across both cycles (`git diff\n  --stat` since the baseline SHA), even ones later reverted.\n- **Hypothesis** — current best explanation of the root cause, and why cycles 1\n  and 2 did not resolve it.\n- **Attempted remediations** — what each of the two cycles changed and the\n  resulting (still-identical) error signature.\n- **Remaining evidence needed** — what would disambiguate the hypothesis (e.g.\n  \"needs confirmation whether `state.S` is shared with another in-flight test\n  file run in the same process\").\n\n### Reporting the escalation\n\nState the payload to the user\u002Fcaller directly. **Do not open or update a GitHub\nissue automatically.** Only use an issue-filing capability if one is available in\nthis session *and* its use for this action has been explicitly authorized —\nthis skill can request that external effect, but it cannot grant itself the\nauthority to perform it. If no such capability or approval exists, stop after\npresenting the payload and wait for direction.\n\n## What NOT to do\n\n- Do not edit `freeforge\u002Fsrc\u002F` to make a test pass without first classifying the\n  failure as a production defect (step 2).\n- Do not weaken, delete, or skip an assertion to make a test go green — that is\n  never a valid remediation for any classification.\n- Do not attempt a third remediation cycle for the same error signature; escalate\n  instead.\n- Do not broaden to the full suite or Biome while the narrowed case is still\n  failing.\n- Do not fabricate or paraphrase stdout\u002Fstderr in the escalation payload — quote\n  it.\n","# Healing Test Failures\n\n## Core principle\n\n**Diagnose before changing production code.** A test failure is evidence, not an\ninstruction to edit the nearest file. Do not touch `freeforge\u002Fsrc\u002F` until you know\nwhether the defect is in that source, in the test itself, or in the environment\nrunning it.\n\n## Required workflow\n\nComplete each step in order. Do not skip ahead to remediation.\n\n### 1. Reproduce the exact failure\n\nRun the failing command exactly as reported — do not paraphrase it or swap in a\nbroader command yet.\n\n```bash\nnode --test tests\u002Fsecurity\u002F\u003Cname>.test.mjs\n```\n\nCapture the raw stdout\u002Fstderr verbatim. Record the current commit (`git rev-parse\nHEAD`) as the baseline SHA before making any change. If the failure does not\nreproduce on the first try, run it 2-3 more times before concluding it is flaky —\n`node:test` failures in this repo are almost always deterministic (see\n`tests\u002FAGENTS.md`); an inconsistent result is itself a finding (see step 2,\nenvironment\u002Ftooling).\n\n### 2. Classify the failure\n\nDecide which bucket the failure belongs to before writing any fix:\n\n- **Production defect** — `freeforge\u002Fsrc\u002F*` produces behavior that contradicts its\n  documented\u002Ftested contract (e.g. a DOM mutation, storage key, or sanitize step\n  described in `AGENTS.md` \u002F `tests\u002FAGENTS.md` is missing or wrong).\n- **Test defect** — the test asserts something the contract does not actually\n  require, uses a stale fixture, or misuses `tests\u002Fhelpers\u002Fmock-dom.mjs` (e.g.\n  missing `restore()`, missing `resetState(S)`, asserting on a global that was\n  never installed).\n- **Environment\u002Ftooling defect** — failure depends on Node version, run order,\n  leaked global state, or timing, not on the test's assertions being wrong. In\n  this codebase this almost always traces back to one of the invariants in\n  `tests\u002FAGENTS.md` §\"Invariants Every Test Must Preserve\": unrestored globals,\n  stale `state.S`, an unresolved microtask, or a stray network\u002Fstorage call.\n\nOnly \"production defect\" justifies editing `freeforge\u002Fsrc\u002F`. \"Test defect\" is\nfixed in `tests\u002Fsecurity\u002F`. \"Environment\u002Ftooling\" is fixed in the test's setup\n(`installGlobals`\u002F`restore`\u002F`resetState`\u002Fawaited microtask), not in application\ncode.\n\n### 3. Minimize to the smallest failing case\n\nNarrow to the single failing `test()` block, not the whole file:\n\n```bash\nnode --test --test-name-pattern=\"\u003Cexact test name>\" tests\u002Fsecurity\u002F\u003Cname>.test.mjs\n```\n\nIf the failure only appears when the full suite runs (`npm --prefix freeforge\ntest`) but not in isolation, that itself is diagnostic: it points at shared\nstate leaking across tests (unrestored global, un-reset `state.S`, or import\ncaching — see `importFresh` vs `importShared` in `tests\u002FAGENTS.md`), i.e. an\nenvironment\u002Ftooling defect, not the isolated test's logic.\n\n### 4. Inspect the relevant source and test contract\n\nBefore writing a fix, read:\n\n- The `freeforge\u002Fsrc\u002F` module(s) the test exercises.\n- The test file itself, especially its `installGlobals(...)` call, whether it\n  wraps in `try\u002Ffinally { restore(); }`, whether it calls `resetState(S)` before\n  `importFresh('freeforge\u002Fsrc\u002Fapp.js')`, and whether it awaits a microtask\n  (`await new Promise(r => setTimeout(r, 0))`) after dispatching an event that\n  triggers async work (stream completion, model load).\n- `tests\u002Fhelpers\u002Fmock-dom.mjs` for the exact shape of `MockDocument`,\n  `MemoryStorage`, and `makeFetchResponse` if the failure touches DOM, storage,\n  or fetch mocking.\n- The nearest `AGENTS.md` (root, then `tests\u002FAGENTS.md`) for the invariant the\n  test is meant to enforce (no real network, no real storage, sanitize-before-\n  `innerHTML`, sessionStorage-only key storage, etc.).\n\nDo not guess the contract from the test name alone.\n\n### 5. Remediate — at most two targeted cycles per error signature\n\nAn \"error signature\" is the specific assertion message \u002F stack frame \u002F thrown\nerror, not just \"the test is red.\" For a given error signature:\n\n- **Cycle 1**: form one hypothesis from steps 1-4, make the smallest change that\n  tests it (one file, one root cause — no bundled cleanup), rerun the narrowed\n  command from step 3.\n- **Cycle 2** (only if cycle 1's rerun still shows the *same* error signature):\n  form a new hypothesis informed by cycle 1's result, make one more targeted\n  change, rerun.\n- If the *same* error signature appears a third time: **stop**. Do not attempt a\n  third remediation. Produce the escalation payload (below) instead.\n\nA different error signature after a fix is progress, not a repeat — it starts a\nfresh two-cycle budget for the new signature, but note in your report that the\nsymptom changed.\n\n### 6. Rerun focused validation after each cycle\n\nAfter every remediation attempt, rerun only the narrowed command from step 3\nfirst. Do not broaden scope while the focused case is still red.\n\n### 7. Broaden only after focused success\n\nOnce the narrowed test\u002Ftest-name-pattern passes, broaden in this order and stop\nif any step fails (treat a new failure as its own diagnosis, not a rubber-stamp\nretry):\n\n```bash\nnode --test tests\u002Fsecurity\u002F\u003Cname>.test.mjs      # whole file\nnpm --prefix freeforge test                      # full suite\nnpx --yes @biomejs\u002Fbiome@1.9.4 check freeforge\u002Fsrc tests\u002Fsecurity tests\u002Fhelpers\n```\n\nBoth commands are listed as the repository's \"Full pre-merge verification\" in\n`AGENTS.md` — run them before calling the fix done.\n\n### 8. Stop on repeated identical failure — produce the escalation payload\n\nIf step 5's cap is hit (same error signature survives two targeted remediation\ncycles), stop editing and assemble the escalation payload below instead of\nattempting a third fix.\n\n## Escalation payload\n\nProduce exactly these fields, backed by evidence gathered in steps 1-7 — do not\nsummarize from memory:\n\n- **Baseline SHA** — `git rev-parse HEAD` captured in step 1, before any edits.\n- **Exact command** — the precise command from step 1 or step 3 that fails.\n- **Raw stdout\u002Fstderr** — verbatim output (or a faithful excerpt if very long —\n  keep the assertion diff and stack trace, trim repeated boilerplate).\n- **Error signature** — the specific assertion\u002Fexception that repeated\n  identically across both remediation cycles.\n- **Files changed** — every file touched across both cycles (`git diff\n  --stat` since the baseline SHA), even ones later reverted.\n- **Hypothesis** — current best explanation of the root cause, and why cycles 1\n  and 2 did not resolve it.\n- **Attempted remediations** — what each of the two cycles changed and the\n  resulting (still-identical) error signature.\n- **Remaining evidence needed** — what would disambiguate the hypothesis (e.g.\n  \"needs confirmation whether `state.S` is shared with another in-flight test\n  file run in the same process\").\n\n### Reporting the escalation\n\nState the payload to the user\u002Fcaller directly. **Do not open or update a GitHub\nissue automatically.** Only use an issue-filing capability if one is available in\nthis session *and* its use for this action has been explicitly authorized —\nthis skill can request that external effect, but it cannot grant itself the\nauthority to perform it. If no such capability or approval exists, stop after\npresenting the payload and wait for direction.\n\n## What NOT to do\n\n- Do not edit `freeforge\u002Fsrc\u002F` to make a test pass without first classifying the\n  failure as a production defect (step 2).\n- Do not weaken, delete, or skip an assertion to make a test go green — that is\n  never a valid remediation for any classification.\n- Do not attempt a third remediation cycle for the same error signature; escalate\n  instead.\n- Do not broaden to the full suite or Biome while the narrowed case is still\n  failing.\n- Do not fabricate or paraphrase stdout\u002Fstderr in the escalation payload — quote\n  it.\n",{"name":5,"description":7,"compatibility":23},"Requires Node.js 22+ and the repository's `node:test` suite under `tests\u002Fsecurity\u002F`.","main","a5bed13dcca926e006d2681d548e722fffa499ab",".agents\u002Fskills\u002Fhealing-test-failures","https:\u002F\u002Fgithub.com\u002FJackSmack1971\u002Ffreeforge-chat","https:\u002F\u002Fgithub.com\u002FJackSmack1971","2026-10-02 05:58:57",{"source":8,"skillName":5,"commands":31,"prompt":41,"zipAvailable":42},[32,35,38],{"tool":33,"command":34},"npx","npx skills add JackSmack1971\u002Ffreeforge-chat --skill healing-test-failures",{"tool":36,"command":37},"bunx","bunx skills add JackSmack1971\u002Ffreeforge-chat --skill healing-test-failures",{"tool":39,"command":40},"pnpm","pnpm dlx skills add JackSmack1971\u002Ffreeforge-chat --skill healing-test-failures","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FJackSmack1971\u002Ffreeforge-chat\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fhealing-test-failures\u002FSKILL.md",true,[44,50],{"path":45,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":48,"sourceHtmlUrl":49},"tests\u002Fevaluation-cases.md","REFERENCE","参考文档",1170,"https:\u002F\u002Fgithub.com\u002FJackSmack1971\u002Ffreeforge-chat\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fhealing-test-failures\u002Ftests\u002Fevaluation-cases.md",{"path":51,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":52,"sourceHtmlUrl":53},"VERIFICATION.md",4464,"https:\u002F\u002Fgithub.com\u002FJackSmack1971\u002Ffreeforge-chat\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fhealing-test-failures\u002FVERIFICATION.md",[55,62,68,73,78,83],{"skillNo":56,"name":57,"nameZh":6,"description":58,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":59,"categories":60,"keywords":61,"publishedAt":18},"01M7596K2F59EBMTV0GEA7G1DF","sample","Fixture skill.","https:\u002F\u002Fgithub.com\u002FJackSmack1971\u002Ffreeforge-chat\u002Fblob\u002Fmain\u002Ftests\u002Fcontrol-plane\u002Ffixtures\u002Fsnapshot-good\u002F.agents\u002Fskills\u002Fsample\u002FSKILL.md",[],[],{"skillNo":63,"name":57,"nameZh":6,"description":64,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":65,"categories":66,"keywords":67,"publishedAt":18},"01M7596HMJGJCK6C1BBBP6YFWM","A minimal fixture skill used only to exercise the verifier.","https:\u002F\u002Fgithub.com\u002FJackSmack1971\u002Ffreeforge-chat\u002Fblob\u002Fmain\u002Ftests\u002Fcontrol-plane\u002Ffixtures\u002Fmalformed\u002Fwriter-configured-read-only\u002F.agents\u002Fskills\u002Fsample\u002FSKILL.md",[],[],{"skillNo":69,"name":57,"nameZh":6,"description":64,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":70,"categories":71,"keywords":72,"publishedAt":18},"01M7596FVEYPJ99D4XJZZ07HB1","https:\u002F\u002Fgithub.com\u002FJackSmack1971\u002Ffreeforge-chat\u002Fblob\u002Fmain\u002Ftests\u002Fcontrol-plane\u002Ffixtures\u002Fmalformed\u002Fstale-identifier-under-codex\u002F.agents\u002Fskills\u002Fsample\u002FSKILL.md",[],[],{"skillNo":74,"name":57,"nameZh":6,"description":64,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":75,"categories":76,"keywords":77,"publishedAt":18},"01M7596ECKJNJE0G09SFTENYTW","https:\u002F\u002Fgithub.com\u002FJackSmack1971\u002Ffreeforge-chat\u002Fblob\u002Fmain\u002Ftests\u002Fcontrol-plane\u002Ffixtures\u002Fmalformed\u002Freviewer-configured-write-enabled\u002F.agents\u002Fskills\u002Fsample\u002FSKILL.md",[],[],{"skillNo":79,"name":57,"nameZh":6,"description":64,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":18},"01M7596BMRAMMD4P3MW4N25ZAV","https:\u002F\u002Fgithub.com\u002FJackSmack1971\u002Ffreeforge-chat\u002Fblob\u002Fmain\u002Ftests\u002Fcontrol-plane\u002Ffixtures\u002Fmalformed\u002Fno-rules-dir\u002F.agents\u002Fskills\u002Fsample\u002FSKILL.md",[],[],{"skillNo":84,"name":57,"nameZh":6,"description":64,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":18},"01M759695BRBKRSK4KDY4VCFR4","https:\u002F\u002Fgithub.com\u002FJackSmack1971\u002Ffreeforge-chat\u002Fblob\u002Fmain\u002Ftests\u002Fcontrol-plane\u002Ffixtures\u002Fmalformed\u002Fmissing-skill-md\u002F.agents\u002Fskills\u002Fsample\u002FSKILL.md",[],[]]