[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6KJCTEPAAPV3YE6GXFC1RQ4":3,"skill-same:01M6KJCTEPAAPV3YE6GXFC1RQ4":55},{"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":22,"publishedAt":29,"linkOnly":30,"markdownContent":31,"markdownBody":32,"frontMatter":33,"refName":34,"commitSha":35,"skillDirectory":36,"forksCount":13,"repositoryHtmlUrl":37,"repositoryHomepageUrl":6,"ownerHtmlUrl":38,"ownerBlogUrl":6,"ownerDescription":39,"updatedAt":40,"install":41,"resources":54},"01M6KJCTEPAAPV3YE6GXFC1RQ4","token-budget",null,"Manages token budget estimation and tracking to prevent context overflow","nposadaa\u002Fwa_trm_notifier","nposadaa","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F57275210?v=4","01M6KJ95HX8AHGQ83PJKSD5HTE","01M6KJ95NTKE2PHAY9Z88HFXYS",0,"MIT","https:\u002F\u002Fgithub.com\u002Fnposadaa\u002Fwa_trm_notifier\u002Fblob\u002Fmaster\u002F.agents\u002Fskills\u002Ftoken-budget\u002FSKILL.md",[17],{"categoryNo":18,"categoryName":19,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000040","财务","PROFESSION","适用职业",[23,24,25,26,27,28],"Budget","token","token budget","context","tracking","estimation","2026-11-10 10:12:30",false,"---\nname: token-budget\ndescription: Manages token budget estimation and tracking to prevent context overflow\n---\n\n# Token Budget Skill\n\n\u003Crole>\nYou are a token-efficient agent. Your job is to maximize output quality while minimizing token consumption.\n\n**Core principle:** Every token counts. Load only what you need, when you need it.\n\u003C\u002Frole>\n\n---\n\n## Token Estimation\n\n### Quick Estimates\n\n| Content Type | Tokens\u002FLine | Notes |\n|--------------|-------------|-------|\n| Code | ~4-6 | Depends on verbosity |\n| Markdown | ~3-4 | Less dense than code |\n| JSON\u002FYAML | ~5-7 | Structured, repetitive |\n| Comments | ~3-4 | Natural language |\n\n**Rule of thumb:** `tokens ≈ lines × 4`\n\n### File Size Categories\n\n| Category | Lines | Est. Tokens | Action |\n|----------|-------|-------------|--------|\n| Small | \u003C50 | \u003C200 | Load freely |\n| Medium | 50-200 | 200-800 | Consider outline first |\n| Large | 200-500 | 800-2000 | Use search + snippets |\n| Huge | 500+ | 2000+ | Never load fully |\n\n---\n\n## Budget Thresholds\n\nBased on PROJECT_RULES.md context quality thresholds:\n\n| Usage | Quality | Budget Status |\n|-------|---------|---------------|\n| 0-30% | PEAK | ✅ Proceed freely |\n| 30-50% | GOOD | ⚠️ Be selective |\n| 50-70% | DEGRADING | ? Compress & summarize |\n| 70%+ | POOR | ? State dump required |\n\n---\n\n## Budget Tracking Protocol\n\n### Before Each Task\n\n1. **Estimate current usage:**\n   - Count files in context\n   - Estimate tokens per file\n   - Calculate approximate %\n\n2. **Check budget status:**\n   ```\n   Current: ~X,000 tokens (~Y%)\n   Budget: [PEAK|GOOD|DEGRADING|POOR]\n   ```\n\n3. **Adjust strategy:**\n   - PEAK: Proceed normally\n   - GOOD: Prefer search-first\n   - DEGRADING: Use outlines only\n   - POOR: Trigger state dump\n\n### During Execution\n\nTrack cumulative context:\n\n```markdown\n## Token Tracker\n\n| Phase | Files Loaded | Est. Tokens | Cumulative |\n|-------|--------------|-------------|------------|\n| Start | 0 | 0 | 0 |\n| Task 1 | 2 | ~400 | ~400 |\n| Task 2 | 3 | ~600 | ~1000 |\n```\n\n---\n\n## Optimization Strategies\n\n### 1. Progressive Loading\n\n```\nLevel 1: Outline only (function signatures)\nLevel 2: + Key functions (based on task)\nLevel 3: + Related code (if needed)\nLevel 4: Full file (only if essential)\n```\n\n### 2. Just-In-Time Loading\n\n- Load file only when task requires it\n- Unload mentally after task complete\n- Don't preload \"just in case\"\n\n### 3. Search Before Load\n\nAlways use context-fetch skill first:\n1. Search for relevant terms\n2. Identify candidate files\n3. Load only needed sections\n\n### 4. Summarize & Compress\n\nAfter understanding a file:\n- Document key insights in STATE.md\n- Reference summary instead of re-reading\n- Use \"I've analyzed X, it does Y\" pattern\n\n---\n\n## Budget Alerts\n\n### At 50% Budget\n\n```\n⚠️ TOKEN BUDGET: 50%\nSwitching to efficiency mode:\n- Outlines only for new files\n- Summarizing instead of loading\n- Recommending compression\n```\n\n### At 70% Budget\n\n```\n? TOKEN BUDGET: 70%\nQuality degradation likely. Recommend:\n1. Create state snapshot\n2. Run \u002Fpause\n3. Continue in fresh session\n```\n\n---\n\n## Integration\n\nThis skill integrates with:\n- `context-fetch` — Search before loading\n- `context-health-monitor` — Quality tracking\n- `context-compressor` — Compression strategies\n- `\u002Fpause` and `\u002Fresume` — Session handoff\n\n---\n\n## Anti-Patterns\n\n❌ **Loading files \"for context\"** — Search first\n❌ **Re-reading same file** — Summarize once\n❌ **Full file when snippet suffices** — Target load\n❌ **Ignoring budget warnings** — Quality will degrade\n\n---\n\n*Part of GSD v1.6 Token Optimization. See PROJECT_RULES.md for efficiency rules.*\n","# Token Budget Skill\n\n\u003Crole>\nYou are a token-efficient agent. Your job is to maximize output quality while minimizing token consumption.\n\n**Core principle:** Every token counts. Load only what you need, when you need it.\n\u003C\u002Frole>\n\n---\n\n## Token Estimation\n\n### Quick Estimates\n\n| Content Type | Tokens\u002FLine | Notes |\n|--------------|-------------|-------|\n| Code | ~4-6 | Depends on verbosity |\n| Markdown | ~3-4 | Less dense than code |\n| JSON\u002FYAML | ~5-7 | Structured, repetitive |\n| Comments | ~3-4 | Natural language |\n\n**Rule of thumb:** `tokens ≈ lines × 4`\n\n### File Size Categories\n\n| Category | Lines | Est. Tokens | Action |\n|----------|-------|-------------|--------|\n| Small | \u003C50 | \u003C200 | Load freely |\n| Medium | 50-200 | 200-800 | Consider outline first |\n| Large | 200-500 | 800-2000 | Use search + snippets |\n| Huge | 500+ | 2000+ | Never load fully |\n\n---\n\n## Budget Thresholds\n\nBased on PROJECT_RULES.md context quality thresholds:\n\n| Usage | Quality | Budget Status |\n|-------|---------|---------------|\n| 0-30% | PEAK | ✅ Proceed freely |\n| 30-50% | GOOD | ⚠️ Be selective |\n| 50-70% | DEGRADING | ? Compress & summarize |\n| 70%+ | POOR | ? State dump required |\n\n---\n\n## Budget Tracking Protocol\n\n### Before Each Task\n\n1. **Estimate current usage:**\n   - Count files in context\n   - Estimate tokens per file\n   - Calculate approximate %\n\n2. **Check budget status:**\n   ```\n   Current: ~X,000 tokens (~Y%)\n   Budget: [PEAK|GOOD|DEGRADING|POOR]\n   ```\n\n3. **Adjust strategy:**\n   - PEAK: Proceed normally\n   - GOOD: Prefer search-first\n   - DEGRADING: Use outlines only\n   - POOR: Trigger state dump\n\n### During Execution\n\nTrack cumulative context:\n\n```markdown\n## Token Tracker\n\n| Phase | Files Loaded | Est. Tokens | Cumulative |\n|-------|--------------|-------------|------------|\n| Start | 0 | 0 | 0 |\n| Task 1 | 2 | ~400 | ~400 |\n| Task 2 | 3 | ~600 | ~1000 |\n```\n\n---\n\n## Optimization Strategies\n\n### 1. Progressive Loading\n\n```\nLevel 1: Outline only (function signatures)\nLevel 2: + Key functions (based on task)\nLevel 3: + Related code (if needed)\nLevel 4: Full file (only if essential)\n```\n\n### 2. Just-In-Time Loading\n\n- Load file only when task requires it\n- Unload mentally after task complete\n- Don't preload \"just in case\"\n\n### 3. Search Before Load\n\nAlways use context-fetch skill first:\n1. Search for relevant terms\n2. Identify candidate files\n3. Load only needed sections\n\n### 4. Summarize & Compress\n\nAfter understanding a file:\n- Document key insights in STATE.md\n- Reference summary instead of re-reading\n- Use \"I've analyzed X, it does Y\" pattern\n\n---\n\n## Budget Alerts\n\n### At 50% Budget\n\n```\n⚠️ TOKEN BUDGET: 50%\nSwitching to efficiency mode:\n- Outlines only for new files\n- Summarizing instead of loading\n- Recommending compression\n```\n\n### At 70% Budget\n\n```\n? TOKEN BUDGET: 70%\nQuality degradation likely. Recommend:\n1. Create state snapshot\n2. Run \u002Fpause\n3. Continue in fresh session\n```\n\n---\n\n## Integration\n\nThis skill integrates with:\n- `context-fetch` — Search before loading\n- `context-health-monitor` — Quality tracking\n- `context-compressor` — Compression strategies\n- `\u002Fpause` and `\u002Fresume` — Session handoff\n\n---\n\n## Anti-Patterns\n\n❌ **Loading files \"for context\"** — Search first\n❌ **Re-reading same file** — Summarize once\n❌ **Full file when snippet suffices** — Target load\n❌ **Ignoring budget warnings** — Quality will degrade\n\n---\n\n*Part of GSD v1.6 Token Optimization. See PROJECT_RULES.md for efficiency rules.*\n",{"name":5,"description":7},"master","afd8a0f0dcb9b77a6380674baf34aac507e69d70",".agents\u002Fskills\u002Ftoken-budget","https:\u002F\u002Fgithub.com\u002Fnposadaa\u002Fwa_trm_notifier","https:\u002F\u002Fgithub.com\u002Fnposadaa","Fun = Nerdism + Technology","2026-11-05 00:20:35",{"source":8,"skillName":5,"commands":42,"prompt":52,"zipAvailable":53},[43,46,49],{"tool":44,"command":45},"npx","npx skills add nposadaa\u002Fwa_trm_notifier --skill token-budget",{"tool":47,"command":48},"bunx","bunx skills add nposadaa\u002Fwa_trm_notifier --skill token-budget",{"tool":50,"command":51},"pnpm","pnpm dlx skills add nposadaa\u002Fwa_trm_notifier --skill token-budget","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fnposadaa\u002Fwa_trm_notifier\u002Fblob\u002Fmaster\u002F.agents\u002Fskills\u002Ftoken-budget\u002FSKILL.md",true,[],[56,69,83,95,109,122],{"skillNo":57,"name":58,"nameZh":6,"description":59,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":60,"categories":61,"keywords":62,"publishedAt":29},"01M6KJD0BT88C4KVQ7RJJTPT7Y","verifier","Validates implemented work against spec requirements with empirical evidence","https:\u002F\u002Fgithub.com\u002Fnposadaa\u002Fwa_trm_notifier\u002Fblob\u002Fmaster\u002F.agents\u002Fskills\u002Fverifier\u002FSKILL.md",[],[58,63,64,65,66,67,68],"spec","work","evidence","requirements","spec requirements","with",{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":29},"01M6KJCN162BN1C6P66K52TZVP","planner","Creates executable phase plans with task breakdown, dependency analysis, and goal-backward verification","https:\u002F\u002Fgithub.com\u002Fnposadaa\u002Fwa_trm_notifier\u002Fblob\u002Fmaster\u002F.agents\u002Fskills\u002Fplanner\u002FSKILL.md",[],[71,76,77,78,79,80,81,68,82],"verification","analysis","dependency","goal","task","plans","dependency analysis",{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":29},"01M6KJCHHQ3HNHDH8DAEJN3JDC","plan-checker","Validates plans before execution to catch issues early","https:\u002F\u002Fgithub.com\u002Fnposadaa\u002Fwa_trm_notifier\u002Fblob\u002Fmaster\u002F.agents\u002Fskills\u002Fplan-checker\u002FSKILL.md",[],[90,91,92,93,94,81],"plan","plan checker","checker","issues","execution",{"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":101,"publishedAt":29},"01M6KJCDQS4C0DSVZ52CW053MJ","status","Show MemPalace status — room counts, storage usage, and palace health.","https:\u002F\u002Fgithub.com\u002Fnposadaa\u002Fwa_trm_notifier\u002Fblob\u002Fmaster\u002F.agents\u002Fskills\u002Fmempalace\u002F.codex-plugin\u002Fskills\u002Fstatus\u002FSKILL.md",[],[97,102,103,104,105,106,107,108],"mempalace","Bash","usage","show","storage","health","read",{"skillNo":110,"name":111,"nameZh":6,"description":112,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":113,"categories":114,"keywords":115,"publishedAt":29},"01M6KJCA2PFKKNEFZWMMTE3XHM","search","Search your MemPalace — semantic search across all mined memories, projects, and conversations.","https:\u002F\u002Fgithub.com\u002Fnposadaa\u002Fwa_trm_notifier\u002Fblob\u002Fmaster\u002F.agents\u002Fskills\u002Fmempalace\u002F.codex-plugin\u002Fskills\u002Fsearch\u002FSKILL.md",[],[111,103,116,117,108,118,119,102,120,121],"semantic search","projects","across","memories","semantic","conversations",{"skillNo":123,"name":124,"nameZh":6,"description":125,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":126,"categories":127,"keywords":128,"publishedAt":29},"01M6KJC5MKM8KWT703Q4GC8CP9","mine","Mine a project or conversation into your MemPalace — extract and store memories for later retrieval.","https:\u002F\u002Fgithub.com\u002Fnposadaa\u002Fwa_trm_notifier\u002Fblob\u002Fmaster\u002F.agents\u002Fskills\u002Fmempalace\u002F.codex-plugin\u002Fskills\u002Fmine\u002FSKILL.md",[],[124,129,108,130,131,132,119,102,133,134,135,136,103],"grep","extract","later","conversation","project","store","retrieval","for"]