菜单
Autonomous iterative optimization loop — run N iterations against a mechanical metric, learn from git history, auto-keep/discard changes. Use for improving measurable metrics (coverage, performance, bundle size, etc.) through repeated experimentation.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/phuc-nt/my-translator --skill ck-autoresearch命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
AI-optimized browser automation CLI with context-efficient snapshots. Use for long autonomous sessions, self-verifying workflows, video recording, and cloud browser testing (Browserbase).
Generate images via Nano Banana with 129 curated prompts. Mandatory validation interview refines style/mood/colors (use --skip to bypass). 3 modes: search, creative, wild. Styles: Ukiyo-e, Bento grid, cyberpunk, cinematic, vintage patent.
Analyze images/audio/video with Gemini API (better vision than Claude). Generate images (Imagen 4, Nano Banana 2, MiniMax), videos (Veo 3, Hailuo), speech (MiniMax TTS), music (MiniMax). Use for vision analysis, transcription, OCR, design extraction, multimodal AI.
基于 SOC 职业分类
| name | ck:autoresearch |
| description | Autonomous iterative optimization loop — run N iterations against a mechanical metric, learn from git history, auto-keep/discard changes. Use for improving measurable metrics (coverage, performance, bundle size, etc.) through repeated experimentation. |
| argument-hint | [Goal/Metric description] or inline config block |
| metadata | {"author":"claudekit","attribution":"Core patterns adapted from autoresearch by Udit Goenka (MIT)","license":"MIT","version":"1.0.0"} |
Constraint + Mechanical Metric + Fast Verification = Autonomous Improvement
| Situation | Better Tool |
|---|---|
| Subjective goals ("make it cleaner") | ck:cook |
| Bug fixing with known root cause | ck:fix or ck:debug |
| One-shot tasks, no repetition needed |
ck:cook| No mechanical metric to measure progress | ck:cook --interactive |
| Files outside a defined scope | manual approach |
Parsed from user message. Missing required fields trigger a batched AskUserQuestion.
| Field | Description | Example |
|---|---|---|
Goal | Human description of what to improve | "Increase test coverage in src/utils" |
Scope | Glob pattern(s) for editable files | "src/utils/**/*.ts" |
Verify | Shell command that outputs a single number | "npx jest --coverage --json | jq '.coverageMap | .. | .s? | to_entries | map(.value) | (map(select(.>0)) | length) / length * 100' | tail -1" |
| Field | Default | Description |
|---|---|---|
Guard | none | Regression check command (exit 0 = pass) |
Iterations | 10 | Maximum iterations to run |
Noise | medium | Tolerance for metric variance: low / medium / high |
Min-Delta | 0 | Minimum improvement to count as progress |
Direction | higher | Whether higher or lower metric value is better |
When required fields are missing, ask all at once:
AskUserQuestion({
questions: [
{ question: "What metric do you want to improve? (e.g. 'test coverage in src/utils')", field: "Goal" },
{ question: "Which files may be edited? (glob, e.g. 'src/utils/**/*.ts')", field: "Scope" },
{ question: "Verify command — must print a single number to stdout", field: "Verify" },
{ question: "Guard command for regression check? (optional, press Enter to skip)", field: "Guard" }
]
})
See references/autonomous-loop-protocol.md for the full 8-phase specification.
Key invariants:
git revert over git reset — preserve historyEach iteration appends a TSV line to loop-results.tsv in the working directory:
iteration commit metric delta status description
0 a1b2c3d 80.0 - baseline initial measurement
1 e4f5a6b 82.4 +2.4 keep add null checks to parser.ts
2 - 81.9 -0.5 discard extract helper function
See references/autonomous-loop-protocol.md — Phase 7 for full schema.
| Condition | Action |
|---|---|
| 5 consecutive discards | Analyze patterns → shift strategy (different files, different approach) |
| 10 consecutive discards | STOP — report findings, surface to user |
/ck:autoresearch
Goal: Increase test coverage in src/utils from ~60% to 80%
Scope: src/utils/**/*.ts, tests/utils/**/*.test.ts
Verify: npx jest tests/utils --coverage --coverageReporters=json-summary 2>/dev/null | node -e "const d=require('./coverage-summary.json');console.log(d.total.lines.pct)"
Guard: npx tsc --noEmit && npx jest --passWithNoTests
Iterations: 15
Direction: higher
/ck:autoresearch
Goal: Reduce main bundle size below 200KB
Scope: src/**/*.ts, src/**/*.tsx
Verify: npx vite build 2>/dev/null | grep "dist/index" | awk '{print $2}' | sed 's/kB//'
Guard: npx tsc --noEmit
Direction: lower
Min-Delta: 0.5
/ck:autoresearch
Goal: Drive ESLint error count to zero in src/api
Scope: src/api/**/*.ts
Verify: npx eslint src/api --format=json 2>/dev/null | node -e "const r=require('/dev/stdin');console.log(r.reduce((a,f)=>a+f.errorCount,0))" || echo 999
Direction: lower
Iterations: 20
ScopeGuard commandVerify command must complete in < 30 seconds (otherwise loop is impractical)references/autonomous-loop-protocol.md — Full 8-phase loop spec, decision matrix, anti-patternsreferences/git-memory-pattern.md — Git as cross-iteration memory, revert vs reset, commit conventionsreferences/guard-and-noise.md — Guard pattern for regression prevention, noise-aware verificationreferences/results-logging.md — TSV format spec, progressive summariesreferences/metric-library.md — Common verify commands by domain