[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6CQY5Z92PRY5V1TQCWRWY5B":3,"skill-same:01M6CQY5Z92PRY5V1TQCWRWY5B":50},{"skillNo":4,"name":5,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":21,"frontMatter":22,"refName":23,"commitSha":24,"skillDirectory":25,"forksCount":13,"repositoryHtmlUrl":26,"repositoryHomepageUrl":6,"ownerHtmlUrl":27,"ownerBlogUrl":28,"ownerDescription":6,"updatedAt":29,"install":30,"resources":43},"01M6CQY5Z92PRY5V1TQCWRWY5B","self-improvement",null,"dundas\u002Fagentbootup-public","dundas","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F627426?v=4","01M6CQVP2KHXEP7JH9WKKF59S7","David K. Dundas","01M6CQVP76D48NCMMXYDK9CBA8",0,"MIT","https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public\u002Fblob\u002Fmain\u002Ftemplates\u002F.claude\u002Fskills\u002Fself-improvement\u002FSKILL.md",[],[18],"self improvement","2026-11-10 10:12:30",false,"# Self-Improvement Skill\n\n## Goal\n\nEnable any agent to continuously learn from its own coding sessions, build permanent knowledge, and share learnings across a network of brains.\n\n## When to Use\n\n- At the **start of each session**: Load memory and apply past learnings\n- At the **end of significant sessions**: Trigger analysis to extract insights\n- **Periodically**: Run analysis on recent sessions to catch anything missed\n- **When deploying new brains**: Propagate accumulated knowledge to new instances\n\n## Prerequisites\n\n- `agentbootup` installed in the project (`npx agentbootup`)\n- `MECH_APP_ID` and `MECH_API_KEY` environment variables set\n- `memory\u002F` directory exists in project root\n\n## Workflow\n\n### Phase 1: Analyze Sessions\n\nRun transcript analysis to extract insights from recent coding sessions.\n\n```bash\n# Analyze last 24 hours\nanalyze-transcripts\n\n# Analyze all unprocessed sessions\nanalyze-transcripts --all\n\n# Preview what would be analyzed\nanalyze-transcripts --dry-run --verbose\n```\n\n**What happens:**\n1. Finds Claude Code session transcripts for this project\n2. Filters to recent\u002Funprocessed sessions\n3. Parses each transcript (messages, files modified, errors)\n4. Sends session context to LLM for insight extraction\n5. Writes insights to `memory\u002Fdaily\u002FYYYY-MM-DD.md`\n6. Updates `memory\u002FMEMORY.md` with significant learnings (deduplicated)\n\n### Phase 2: Review and Curate\n\nAfter analysis, review what was extracted:\n\n```bash\n# Check today's daily log\ncat memory\u002Fdaily\u002F$(date +%Y-%m-%d).md\n\n# Check what was added to long-term memory\ngit diff memory\u002FMEMORY.md\n\n# View analysis statistics\nanalyze-transcripts --stats\n```\n\n**Curation actions:**\n- **Promote**: Move auto-extracted insights to hand-written sections in MEMORY.md\n- **Refine**: Edit auto-extracted text for clarity and precision\n- **Remove**: Delete insights that are incorrect or not useful\n- **Extract skill**: If a learning is a reusable capability, create a skill\n\n### Phase 3: Apply in Next Session\n\nAt session start, memory is automatically loaded via CLAUDE.md. The agent should:\n\n1. Check `memory\u002FMEMORY.md` for relevant patterns and anti-patterns\n2. Check recent daily logs for context on current work\n3. Apply \"never do X\" rules to avoid past mistakes\n4. Use learned patterns to make better decisions\n\n### Phase 4: Share Across Brains\n\nFor multi-brain setups:\n\n1. **Memory sync daemon** pushes\u002Fpulls memory changes via Mech Storage\n2. **agentbootup templates** carry curated knowledge to new projects\n3. **Skills** are shared when synced to Mech Storage\n\n## Commands\n\n| Command | Description |\n|---------|-------------|\n| `analyze-transcripts` | Analyze recent sessions (last 24h) |\n| `analyze-transcripts --all` | Analyze all unprocessed sessions |\n| `analyze-transcripts --hours 168` | Analyze last week |\n| `analyze-transcripts --session abc123` | Analyze specific session |\n| `analyze-transcripts --dry-run` | Preview without writing |\n| `analyze-transcripts --verbose` | Show extracted insights |\n| `analyze-transcripts --reset --all` | Re-analyze everything |\n| `analyze-transcripts --stats` | Show analysis statistics |\n| `memory-sync-daemon start` | Start continuous sync |\n\n## Memory Architecture\n\n```\nmemory\u002F\n├── MEMORY.md              ← Long-term (loaded every session, \u003C200 lines)\n├── README.md              ← System documentation\n└── daily\u002F\n    ├── 2026-02-05.md      ← Today's session logs\n    ├── 2026-02-04.md      ← Yesterday's logs\n    └── ...\n```\n\n### MEMORY.md Structure\n\n```markdown\n# Autonomous Memory System\n\n## Core Identity\n(Agent name, role, purpose)\n\n## Critical Learnings\n### Auto-extracted (2026-02-05)\n- Learning from transcript analysis\n- Another learning\n\n### Hand-written\n- Curated permanent knowledge\n\n## Skills Acquired (N)\n- List of capabilities\n\n## Standing Orders\n- Always-active behavioral rules\n```\n\n### Daily Log Structure\n\n```markdown\n# Daily Log: 2026-02-05\n\n## Session abc12345 (14:30)\n**Summary:** Built fuzzy search for transcript queries\n**Duration:** 2h 15m\n**Activity:** 45 messages, 8 files modified\n\n### Technical Learnings\n- Levenshtein distance catches typos within threshold of 2\n\n### Mistakes & Corrections\n- **Mistake:** Wrong stemming suffix order\n  - **Correction:** Apply longer suffixes first\n  - **Lesson:** Order matters in sequential string operations\n```\n\n## Significance Criteria\n\nNot all learnings make it to MEMORY.md. To qualify, a learning must:\n\n1. **Contain a signal keyword**: never, always, critical, important, security, must, bug, fix, pattern, breaking, gotcha, workaround\n2. **OR be substantial**: More than 50 characters of specific, actionable insight\n\nLearnings that are too generic (\"learned about APIs\") or too short are kept in daily logs only.\n\n## Deploying to Company Brains\n\n### Step 1: Curate Your Knowledge\n\nReview your brain's `memory\u002FMEMORY.md` and identify universally valuable learnings.\n\n### Step 2: Promote to Templates\n\nAdd curated knowledge to agentbootup templates:\n- `templates\u002Fmemory\u002FMEMORY.md` - Default memory template\n- `templates\u002F.ai\u002Fprotocols\u002FSELF_IMPROVEMENT.md` - This protocol\n- `templates\u002F.claude\u002Fskills\u002Fself-improvement\u002F` - This skill\n\n### Step 3: Publish\n\n```bash\ncd ~\u002Fdev_env\u002Fagentbootup\n# Bump version in package.json\nnpm publish\n```\n\n### Step 4: Deploy to Each Brain\n\n```bash\ncd ~\u002Fdev_env\u002Fproject-brain\nnpx agentbootup\n```\n\n### Step 5: Enable Continuous Learning\n\nEach brain needs:\n1. `MECH_APP_ID` and `MECH_API_KEY` in environment\n2. `analyze-transcripts` run periodically (cron or daemon)\n3. `memory-sync-daemon` for cross-session persistence\n\n## Error Handling\n\n| Error | Cause | Fix |\n|-------|-------|-----|\n| \"MECH_APP_ID required\" | Missing env vars | Set MECH_APP_ID and MECH_API_KEY |\n| \"No transcripts found\" | Wrong project path | Use `--project` to specify correct path |\n| \"Session not significant\" | Too few messages\u002Fchanges | Normal - only substantial sessions get analyzed |\n| \"All learnings already in MEMORY.md\" | Dedup working correctly | Normal - prevents duplicates |\n| LLM API error | Mech service issue | Check credentials, retry later |\n\n## References\n\n- `templates\u002F.ai\u002Fprotocols\u002FSELF_IMPROVEMENT.md` - Full protocol documentation\n- `lib\u002Fanalysis\u002FREADME.md` - Technical architecture\n- `templates\u002F.ai\u002Fprotocols\u002FAUTONOMOUS_OPERATION.md` - Autonomous behavior rules\n- `templates\u002F.ai\u002Fskills\u002Fmemory-manager\u002FSKILL.md` - Memory management skill\n",{},"main","47b89745588c3057f0ccd0517a2eb8967df27054","templates\u002F.claude\u002Fskills\u002Fself-improvement","https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public","https:\u002F\u002Fgithub.com\u002Fdundas","derivative.io","2026-09-02 20:12:15",{"source":7,"skillName":5,"commands":31,"prompt":41,"zipAvailable":42},[32,35,38],{"tool":33,"command":34},"npx","npx skills add dundas\u002Fagentbootup-public --skill self-improvement",{"tool":36,"command":37},"bunx","bunx skills add dundas\u002Fagentbootup-public --skill self-improvement",{"tool":39,"command":40},"pnpm","pnpm dlx skills add dundas\u002Fagentbootup-public --skill self-improvement","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public\u002Fblob\u002Fmain\u002Ftemplates\u002F.claude\u002Fskills\u002Fself-improvement\u002FSKILL.md",true,[44],{"path":45,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":48,"sourceHtmlUrl":49},"skill-bundle-manifest.json","REFERENCE","参考文档",804,"https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public\u002Fblob\u002Fmain\u002Ftemplates\u002F.claude\u002Fskills\u002Fself-improvement\u002Fskill-bundle-manifest.json",[51,65,85,105,114,136],{"skillNo":52,"name":53,"nameZh":6,"description":54,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":55,"categories":56,"keywords":62,"publishedAt":19},"01M6CQZ7DFJB965BJJ5ZZ54CXY","fixture-skill","Secret-free recovery fixture skill.","https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public\u002Fblob\u002Fmain\u002Ftests\u002Fruntime-adapters\u002Ffixtures\u002Fhermes\u002F0.18.2-darwin-arm64-real\u002Froot\u002Fskills\u002Ffixture-skill\u002FSKILL.md",[57],{"categoryNo":58,"categoryName":59,"taxonomyType":60,"taxonomyTypeDesc":61},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[63,64],"fixture skill","recovery",{"skillNo":66,"name":67,"nameZh":6,"description":68,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":69,"categories":70,"keywords":76,"publishedAt":19},"01M6CQZ4NN0KVH1Z4E1Y1GJ3X9","user-story-generator","Generate standalone user stories for features without creating a full PRD. Perfect for backlog grooming and story refinement.","https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public\u002Fblob\u002Fmain\u002Ftemplates\u002F.claude\u002Fskills\u002Fuser-story-generator\u002FSKILL.md",[71],{"categoryNo":72,"categoryName":73,"taxonomyType":74,"taxonomyTypeDesc":75},"01JBSEEDCAT000000000000031","产品经理","PROFESSION","适用职业",[77,78,79,80,81,82,83,84],"user story generator","Story","user story","user","prd","documentation","generate","backlog",{"skillNo":86,"name":87,"nameZh":6,"description":88,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":89,"categories":90,"keywords":97,"publishedAt":19},"01M6CQZ16VC27HMWM31CCKK6BC","user-journey-mapper","Map user flows and journeys through your application for UX design and workflow visualization, independent of testing.","https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public\u002Fblob\u002Fmain\u002Ftemplates\u002F.claude\u002Fskills\u002Fuser-journey-mapper\u002FSKILL.md",[91,94],{"categoryNo":92,"categoryName":93,"taxonomyType":60,"taxonomyTypeDesc":61},"01JBSEEDCAT000000000000008","测试与质量",{"categoryNo":95,"categoryName":96,"taxonomyType":74,"taxonomyTypeDesc":75},"01JBSEEDCAT000000000000032","设计师",[98,99,80,82,100,101,102,103,104],"user journey mapper","journey","ux","visualization","workflow","design","testing",{"skillNo":106,"name":107,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":108,"categories":109,"keywords":110,"publishedAt":19},"01M6CQYXSQQJ4Z5CRAX2SSMHW1","transcript-query","https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public\u002Fblob\u002Fmain\u002Ftemplates\u002F.claude\u002Fskills\u002Ftranscript-query\u002FSKILL.md",[],[111,112,113],"transcript","transcript query","query",{"skillNo":115,"name":116,"nameZh":6,"description":117,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":118,"categories":119,"keywords":124,"publishedAt":19},"01M6CQYT7A1G4CPQTZKXR5TCK9","test-plan-generator","Generate comprehensive E2E test plans from PRDs, task lists, or descriptions with environment-aware prerequisites and issue tracking.","https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public\u002Fblob\u002Fmain\u002Ftemplates\u002F.claude\u002Fskills\u002Ftest-plan-generator\u002FSKILL.md",[120,121],{"categoryNo":92,"categoryName":93,"taxonomyType":60,"taxonomyTypeDesc":61},{"categoryNo":122,"categoryName":123,"taxonomyType":60,"taxonomyTypeDesc":61},"01JBSEEDCAT000000000000010","后端与接口",[125,126,127,128,129,130,131,132,133,134,83,135],"plan","test plan generator","test plan","test","task","issue","e2e test","environment","tracking","quality","e2e",{"skillNo":137,"name":138,"nameZh":6,"description":139,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":140,"categories":141,"keywords":143,"publishedAt":19},"01M6CQYQAWFSPTYE2VTJJVW72G","tasklist-generator","Generate high-level tasks and gated sub-tasks from a PRD, with agent assignments, commit\u002FPR strategy, and relevant files.","https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public\u002Fblob\u002Fmain\u002Ftemplates\u002F.claude\u002Fskills\u002Ftasklist-generator\u002FSKILL.md",[142],{"categoryNo":58,"categoryName":59,"taxonomyType":60,"taxonomyTypeDesc":61},[144,145,146,147,81,83,148,149,150,102],"tasklist generator","gated","files","strategy","tasks","pr","commit"]