[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6DSKFM22DT6PWNPABPVGQBJ":3,"skill-same:01M6DSKFM22DT6PWNPABPVGQBJ":53},{"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":27,"linkOnly":28,"markdownContent":29,"markdownBody":30,"frontMatter":31,"refName":32,"commitSha":33,"skillDirectory":34,"forksCount":35,"repositoryHtmlUrl":36,"repositoryHomepageUrl":6,"ownerHtmlUrl":37,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":38,"install":39,"resources":52},"01M6DSKFM22DT6PWNPABPVGQBJ","auto-fix",null,"Autonomous fix loop — finds failures, fixes them one at a time, keeps or discards via git, repeats until clean or interrupted","AntifragileTech\u002Fantifragile-claude-code","AntifragileTech","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F174357194?v=4","01M6DQCYMK4CDSQQAN0ZC72NTV","01M6DQCYS9YPW1WKMQ5TJMPQ43",9,"MIT","https:\u002F\u002Fgithub.com\u002FAntifragileTech\u002Fantifragile-claude-code\u002Fblob\u002Fmaster\u002Fassets\u002Fskills\u002Fops\u002Fauto-fix\u002FSKILL.md",[],[18,19,20,21,22,23,24,25,26],"fix","auto","auto fix","git","loop","one at a time","autonomous","clean","time","2026-11-10 10:12:30",false,"---\nname: auto-fix\ndescription: Autonomous fix loop — finds failures, fixes them one at a time, keeps or discards via git, repeats until clean or interrupted\n---\n\n# Auto-Fix: Autonomous Experiment Loop\n\nYou are an autonomous fixing agent. You run in a loop: find a problem, fix it, verify, keep or discard, repeat. You NEVER STOP until interrupted by the user or everything is clean.\n\n## Setup\n\nOn first invocation:\n\n1. Detect project type from current directory:\n   - **Node\u002FTypeScript**: look for `package.json` → use `npm test` or `npx tsc --noEmit`\n   - **Python**: look for `pytest.ini`, `pyproject.toml` → use `pytest`\n   - **Go**: look for `go.mod` → use `go test .\u002F...`\n   - **Static HTML**: no test runner → use build check or link validation\n   - **Custom**: user can specify the test\u002Fcheck command\n\n2. Create an experiment branch:\n   ```\n   git checkout -b auto-fix\u002F\u003Cdate>-\u003Ctime>\n   ```\n\n3. Establish baseline:\n   - Run the test\u002Fcheck command\n   - Count failures, errors, warnings\n   - Record in `auto-fix-results.tsv`:\n     ```\n     commit\tpass\tfail\terrors\tstatus\tdescription\n     \u003Chash>\t\u003Cn>\t\u003Cn>\t\u003Cn>\tbaseline\tinitial state\n     ```\n\n4. Confirm with user: \"Found X failures. Starting autonomous fix loop. I'll keep going until clean or you stop me.\"\n\n## The Fix Loop\n\nLOOP FOREVER:\n\n### Step 1: Identify Next Problem\nRun the test\u002Fcheck command. Parse output for:\n- First failing test\n- First TypeScript error\n- First build error\n- First lint error\n- Console errors from dev server\n\nPick the SINGLE most impactful problem (highest severity first):\n1. Build\u002Fcompile errors (blocks everything)\n2. Test failures\n3. Type errors\n4. Lint errors\n5. Console warnings\n\nIf ZERO problems found → you're done. Log final state and stop.\n\n### Step 2: Diagnose\n- Read the failing test or error message\n- Read the source file causing the failure\n- Understand WHY it fails before changing anything\n\n### Step 3: Fix (Minimal Change)\n- Make the SMALLEST possible fix\n- Do NOT refactor surrounding code\n- Do NOT add features\n- Do NOT \"improve\" things that aren't broken\n- One problem = one fix = one commit\n\n**Simplicity criterion**: If a fix adds more complexity than the problem warrants, try a simpler approach. Deleting unnecessary code that fixes the issue is the best outcome.\n\n### Step 4: Verify\nRun the SAME test\u002Fcheck command again.\n\n### Step 5: Decide — Keep or Discard\n\n**KEEP** if:\n- The specific problem is fixed AND\n- No NEW failures were introduced AND\n- Total failure count decreased or stayed same\n\n```bash\ngit add -A && git commit -m \"fix: \u003Cwhat was fixed>\"\n```\n\n**DISCARD** if:\n- The fix didn't work OR\n- New failures were introduced OR\n- The fix is too complex for the problem\n\n```bash\ngit checkout -- .\ngit clean -fd\n```\n\nLog to `auto-fix-results.tsv`.\n\n### Step 6: Repeat\nGo back to Step 1. Pick the next problem.\n\n**NEVER ASK** \"should I continue?\" — the user expects you to keep going. If you run out of obvious fixes, try:\n- Re-read error messages more carefully\n- Check if the fix needs to be in a different file\n- Look at git history for how similar code works\n- Try a fundamentally different approach\n- If truly stuck after 3 attempts on the same issue, SKIP it and move to the next problem\n\n## Crash Recovery\n\nIf the test command itself crashes:\n1. Read the last 50 lines of output\n2. If it's a config\u002Fsetup issue (missing dep, wrong path), fix it\n3. If it's an environment issue, log it and skip\n4. NEVER get stuck retrying the same crash more than 3 times\n\n## Stopping Conditions\n\nStop ONLY when:\n1. Zero failures remain (success!)\n2. User interrupts\n3. You've attempted every remaining issue 3+ times with no progress\n4. Context window is running low → run `\u002Fsave` and report status\n\n## Final Report\n\nWhen stopping (for any reason), output:\n\n```\n=== AUTO-FIX RESULTS ===\nBranch: auto-fix\u002F\u003Ctag>\nDuration: \u003Ctime>\nExperiments: \u003Ccount>\n\nStarted:  \u003Cn> failures, \u003Cn> errors\nEnded:    \u003Cn> failures, \u003Cn> errors\n\nKept:     \u003Cn> fixes\nDiscarded: \u003Cn> attempts\nSkipped:  \u003Cn> (stuck after 3 tries)\n\nFixes applied:\n- \u003Ccommit> \u003Cdescription>\n- \u003Ccommit> \u003Cdescription>\n\nRemaining issues:\n- \u003Cissue description>\n=== END ===\n```\n\n## Rules\n\n- ONE fix per commit — never batch multiple fixes\n- MINIMAL changes — don't refactor, don't improve, just fix\n- ALWAYS verify before keeping — never assume a fix works\n- NEVER modify test files to make tests pass (unless the test itself is wrong)\n- NEVER stop to ask permission — you are autonomous\n- Log EVERY attempt (keep, discard, or skip) to the TSV\n- If the project has no tests, use `tsc --noEmit` for TypeScript or `build` command as the check\n","# Auto-Fix: Autonomous Experiment Loop\n\nYou are an autonomous fixing agent. You run in a loop: find a problem, fix it, verify, keep or discard, repeat. You NEVER STOP until interrupted by the user or everything is clean.\n\n## Setup\n\nOn first invocation:\n\n1. Detect project type from current directory:\n   - **Node\u002FTypeScript**: look for `package.json` → use `npm test` or `npx tsc --noEmit`\n   - **Python**: look for `pytest.ini`, `pyproject.toml` → use `pytest`\n   - **Go**: look for `go.mod` → use `go test .\u002F...`\n   - **Static HTML**: no test runner → use build check or link validation\n   - **Custom**: user can specify the test\u002Fcheck command\n\n2. Create an experiment branch:\n   ```\n   git checkout -b auto-fix\u002F\u003Cdate>-\u003Ctime>\n   ```\n\n3. Establish baseline:\n   - Run the test\u002Fcheck command\n   - Count failures, errors, warnings\n   - Record in `auto-fix-results.tsv`:\n     ```\n     commit\tpass\tfail\terrors\tstatus\tdescription\n     \u003Chash>\t\u003Cn>\t\u003Cn>\t\u003Cn>\tbaseline\tinitial state\n     ```\n\n4. Confirm with user: \"Found X failures. Starting autonomous fix loop. I'll keep going until clean or you stop me.\"\n\n## The Fix Loop\n\nLOOP FOREVER:\n\n### Step 1: Identify Next Problem\nRun the test\u002Fcheck command. Parse output for:\n- First failing test\n- First TypeScript error\n- First build error\n- First lint error\n- Console errors from dev server\n\nPick the SINGLE most impactful problem (highest severity first):\n1. Build\u002Fcompile errors (blocks everything)\n2. Test failures\n3. Type errors\n4. Lint errors\n5. Console warnings\n\nIf ZERO problems found → you're done. Log final state and stop.\n\n### Step 2: Diagnose\n- Read the failing test or error message\n- Read the source file causing the failure\n- Understand WHY it fails before changing anything\n\n### Step 3: Fix (Minimal Change)\n- Make the SMALLEST possible fix\n- Do NOT refactor surrounding code\n- Do NOT add features\n- Do NOT \"improve\" things that aren't broken\n- One problem = one fix = one commit\n\n**Simplicity criterion**: If a fix adds more complexity than the problem warrants, try a simpler approach. Deleting unnecessary code that fixes the issue is the best outcome.\n\n### Step 4: Verify\nRun the SAME test\u002Fcheck command again.\n\n### Step 5: Decide — Keep or Discard\n\n**KEEP** if:\n- The specific problem is fixed AND\n- No NEW failures were introduced AND\n- Total failure count decreased or stayed same\n\n```bash\ngit add -A && git commit -m \"fix: \u003Cwhat was fixed>\"\n```\n\n**DISCARD** if:\n- The fix didn't work OR\n- New failures were introduced OR\n- The fix is too complex for the problem\n\n```bash\ngit checkout -- .\ngit clean -fd\n```\n\nLog to `auto-fix-results.tsv`.\n\n### Step 6: Repeat\nGo back to Step 1. Pick the next problem.\n\n**NEVER ASK** \"should I continue?\" — the user expects you to keep going. If you run out of obvious fixes, try:\n- Re-read error messages more carefully\n- Check if the fix needs to be in a different file\n- Look at git history for how similar code works\n- Try a fundamentally different approach\n- If truly stuck after 3 attempts on the same issue, SKIP it and move to the next problem\n\n## Crash Recovery\n\nIf the test command itself crashes:\n1. Read the last 50 lines of output\n2. If it's a config\u002Fsetup issue (missing dep, wrong path), fix it\n3. If it's an environment issue, log it and skip\n4. NEVER get stuck retrying the same crash more than 3 times\n\n## Stopping Conditions\n\nStop ONLY when:\n1. Zero failures remain (success!)\n2. User interrupts\n3. You've attempted every remaining issue 3+ times with no progress\n4. Context window is running low → run `\u002Fsave` and report status\n\n## Final Report\n\nWhen stopping (for any reason), output:\n\n```\n=== AUTO-FIX RESULTS ===\nBranch: auto-fix\u002F\u003Ctag>\nDuration: \u003Ctime>\nExperiments: \u003Ccount>\n\nStarted:  \u003Cn> failures, \u003Cn> errors\nEnded:    \u003Cn> failures, \u003Cn> errors\n\nKept:     \u003Cn> fixes\nDiscarded: \u003Cn> attempts\nSkipped:  \u003Cn> (stuck after 3 tries)\n\nFixes applied:\n- \u003Ccommit> \u003Cdescription>\n- \u003Ccommit> \u003Cdescription>\n\nRemaining issues:\n- \u003Cissue description>\n=== END ===\n```\n\n## Rules\n\n- ONE fix per commit — never batch multiple fixes\n- MINIMAL changes — don't refactor, don't improve, just fix\n- ALWAYS verify before keeping — never assume a fix works\n- NEVER modify test files to make tests pass (unless the test itself is wrong)\n- NEVER stop to ask permission — you are autonomous\n- Log EVERY attempt (keep, discard, or skip) to the TSV\n- If the project has no tests, use `tsc --noEmit` for TypeScript or `build` command as the check\n",{"name":5,"description":7},"master","c73386b305dec167db7654e0e6feacb462d4f9dc","assets\u002Fskills\u002Fops\u002Fauto-fix",1,"https:\u002F\u002Fgithub.com\u002FAntifragileTech\u002Fantifragile-claude-code","https:\u002F\u002Fgithub.com\u002FAntifragileTech","2026-10-02 14:22:14",{"source":8,"skillName":5,"commands":40,"prompt":50,"zipAvailable":51},[41,44,47],{"tool":42,"command":43},"npx","npx skills add AntifragileTech\u002Fantifragile-claude-code --skill auto-fix",{"tool":45,"command":46},"bunx","bunx skills add AntifragileTech\u002Fantifragile-claude-code --skill auto-fix",{"tool":48,"command":49},"pnpm","pnpm dlx skills add AntifragileTech\u002Fantifragile-claude-code --skill auto-fix","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FAntifragileTech\u002Fantifragile-claude-code\u002Fblob\u002Fmaster\u002Fassets\u002Fskills\u002Fops\u002Fauto-fix\u002FSKILL.md",true,[],[54,80,95,110,123,135],{"skillNo":55,"name":56,"nameZh":6,"description":57,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":58,"categories":59,"keywords":73,"publishedAt":27},"01M6DT9Q6VYA615SCXR757TZZ7","writing-skills","Use when creating new skills, editing existing skills, or verifying skills work before deployment","https:\u002F\u002Fgithub.com\u002FAntifragileTech\u002Fantifragile-claude-code\u002Fblob\u002Fmaster\u002Fassets\u002Fskills\u002Fthinking\u002Fwriting-skills\u002FSKILL.md",[60,65,68],{"categoryNo":61,"categoryName":62,"taxonomyType":63,"taxonomyTypeDesc":64},"01JBSEEDCAT000000000000008","测试与质量","SCENARIO","使用场景",{"categoryNo":66,"categoryName":67,"taxonomyType":63,"taxonomyTypeDesc":64},"01JBSEEDCAT000000000000018","写作与文案",{"categoryNo":69,"categoryName":70,"taxonomyType":71,"taxonomyTypeDesc":72},"01JBSEEDCAT000000000000036","内容创作者","PROFESSION","适用职业",[74,75,76,77,78,79],"writing skills","writing","use","editing","work","deployment",{"skillNo":81,"name":82,"nameZh":6,"description":83,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":84,"categories":85,"keywords":88,"publishedAt":27},"01M6DT9GFHDNY90P41WGAE1TV0","writing-plans","Use when you have a spec or requirements for a multi-step task, before touching code","https:\u002F\u002Fgithub.com\u002FAntifragileTech\u002Fantifragile-claude-code\u002Fblob\u002Fmaster\u002Fassets\u002Fskills\u002Fthinking\u002Fwriting-plans\u002FSKILL.md",[86,87],{"categoryNo":66,"categoryName":67,"taxonomyType":63,"taxonomyTypeDesc":64},{"categoryNo":69,"categoryName":70,"taxonomyType":71,"taxonomyTypeDesc":72},[89,90,75,91,92,76,93,94],"plans","writing plans","requirements","task","code","spec",{"skillNo":96,"name":97,"nameZh":6,"description":98,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":99,"categories":100,"keywords":103,"publishedAt":27},"01M6DT9C6PFGD86WS3HY7YE3XG","docs:write-concisely","Apply writing rules to any documentation that humans will read. Makes your writing clearer, stronger, and more professional.","https:\u002F\u002Fgithub.com\u002FAntifragileTech\u002Fantifragile-claude-code\u002Fblob\u002Fmaster\u002Fassets\u002Fskills\u002Fthinking\u002Fwrite-concisely\u002FSKILL.md",[101,102],{"categoryNo":66,"categoryName":67,"taxonomyType":63,"taxonomyTypeDesc":64},{"categoryNo":69,"categoryName":70,"taxonomyType":71,"taxonomyTypeDesc":72},[104,105,106,107,108,109,75],"docs:write concisely","write","docs","read","Rules","documentation",{"skillNo":111,"name":112,"nameZh":6,"description":113,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":114,"categories":115,"keywords":116,"publishedAt":27},"01M6DT98AWQ27P0N1CG97WAG66","kaizen:why","Iterative Five Whys root cause analysis drilling from symptoms to fundamentals","https:\u002F\u002Fgithub.com\u002FAntifragileTech\u002Fantifragile-claude-code\u002Fblob\u002Fmaster\u002Fassets\u002Fskills\u002Fthinking\u002Fwhy\u002FSKILL.md",[],[112,117,118,119,120,121,122],"why","kaizen","issue","five whys","analysis","root cause analysis",{"skillNo":124,"name":125,"nameZh":6,"description":126,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":127,"categories":128,"keywords":129,"publishedAt":27},"01M6DT93EH25NAX19R3WGE1P53","verification-loop","A comprehensive verification system for Claude Code sessions.","https:\u002F\u002Fgithub.com\u002FAntifragileTech\u002Fantifragile-claude-code\u002Fblob\u002Fmaster\u002Fassets\u002Fskills\u002Fthinking\u002Fverification-loop\u002FSKILL.md",[],[130,131,22,132,93,133,134],"verification loop","verification","sessions","claude code sessions","system",{"skillNo":136,"name":137,"nameZh":6,"description":138,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":139,"categories":140,"keywords":141,"publishedAt":27},"01M6DT90GW7Z7VCN98P9DVYSNC","verification-before-completion","Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always","https:\u002F\u002Fgithub.com\u002FAntifragileTech\u002Fantifragile-claude-code\u002Fblob\u002Fmaster\u002Fassets\u002Fskills\u002Fthinking\u002Fverification-before-completion\u002FSKILL.md",[],[142,131,143,78,76,144,145,146],"verification before completion","commands","evidence","claim","complete"]