agentleFS
fritzprixfritzprix/libr-agent/.agents/skills/audit-docs/SKILL.md

audit-docs

Skill11 starsChanged 18 days agoMITView on GitHub

Audit documentation gaps by comparing code changes between the last docs update (or release version) and current HEAD. Detects undocumented features, modified tools/skills, configuration changes, and KO-EN translation parity drift. Use when checking for missing documentation, auditing doc gaps, verifying translation parity, or preparing documentation before releases. Triggers on: "audit docs", "audit-docs", "문서 누락 검토", "문서 누락 감사", "docs audit", "문서 드리프트 감사", "마지막 docs 업데이트 분석".

17 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use diagnose.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • MicrosoftSkill

    diagnose

    Use to systematically diagnose why scenarios fail — identifies root causes by analyzing agent traces, codebase, and behavioral patterns

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 2 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 16 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Run the helper script to detect baseline commits, code changes, and doc parity:

    line 39
  2. Very specific:

    Check the KO ↔ EN Documentation Parity Audit table from the script:

    line 82
  3. Distinctive:

    When writing or reviewing English documentation, enforce ASD-STE100 principles:

    line 93
  4. Very specific:

    Avoid ambiguous modals (should, might).

    line 101

“Common” is measured: kept by at least 6 in 10 of the 16 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Audit Docs Skill

Analyze the gap between the codebase changes (features, tools, skills, settings) made since the last documentation update (or release tag) and current HEAD. Verify whether user documentation (website/ and docs/user/) accurately reflects these changes and maintains Korean-English translation parity.


🧭 Workflow Decision Tree

1. Run Automated Audit Script
   └─ python3 .agents/skills/audit-docs/scripts/audit_docs.py [--since <REF> | --release]
2. Inspect Code Changes vs Document Mapping
   ├─ Built-in Tools → docs/user/guides/builtin-tools.md
   ├─ Bundled Skills → docs/user/guides/skills.md
   ├─ Settings/Config → docs/user/guides/troubleshooting.md & getting-started/
   └─ Scenarios/Workflows → docs/user/scenarios/
3. Audit KO ↔ EN Translation Parity & Truncation
   └─ Check if EN ratio < 70% or headings mismatch
4. Verify ASD-STE100 Style Compliance
   └─ Check active voice, imperative verbs, sentence length <= 20-25 words
5. Output Structured Audit Report & Action Plan

Step 1: Run the Audit Script

Run the helper script to detect baseline commits, code changes, and doc parity:

# 1. Automatic mode: Compares against the last commit that touched docs/user/
python3 .agents/skills/audit-docs/scripts/audit_docs.py

# 2. Release mode: Compares against the latest Git release tag
python3 .agents/skills/audit-docs/scripts/audit_docs.py --release

# 3. Custom baseline: Compares against a specific branch or commit ref
python3 .agents/skills/audit-docs/scripts/audit_docs.py --since dev/0.9.x~10

# 4. JSON format for machine parsing
python3 .agents/skills/audit-docs/scripts/audit_docs.py --json --output audit_report.json

Step 2: Analyze Code Diffs & Map to Docs

Inspect code diffs between baseline and HEAD across five key architectural domains:

Code DomainSource PathsTarget User Documentation
Bundled Skillssrc-tauri/bundled_skills/docs/user/guides/skills.md
Built-in Toolssrc-tauri/src/mcp/, src-tauri/src/agent/tools/docs/user/guides/builtin-tools.md
Settings & Configsrc/components/settings/, src-tauri/src/config/docs/user/guides/troubleshooting.md, getting-started/connecting-models.md
UI Views / Routessrc/views/, src/routes/docs/user/guides/navigation-guide.md, docs/user/index.md
Automation / Cronsrc-tauri/src/scheduled/docs/user/guides/scheduled-tasks.md, docs/user/guides/automation.md

Verification Checklist:

  • Were new skills added to bundled_skills without being listed in the Skills Catalog?
  • Were new MCP tools or tool parameters added without reference entries?
  • Were UI setting keys renamed or added without updating troubleshooting and setup guides?
  • Were new execution modes (e.g., YOLO, Unsafe, Normal) added without documentation?

Step 3: Audit KO ↔ EN Translation Parity

A common pitfall is updating Korean documentation while leaving English documents as outdated skeletons.

Check the KO ↔ EN Documentation Parity Audit table from the script:

  1. MISSING_EN: Korean document exists, but English counterpart does not exist. (Critical)
  2. SEVERE_TRUNCATION (Ratio < 40%): English document is severely abbreviated or a placeholder stub. (High)
  3. MODERATE_TRUNCATION (Ratio 40% ~ 70%): English document misses key tables, tips, or sections. (Medium)
  4. OK (Ratio >= 70%): Content volume and structural depth match.

Step 4: Verify ASD-STE100 Style Guidelines

When writing or reviewing English documentation, enforce ASD-STE100 principles:

  1. Sentence Length:
    • Procedural instructions: <= 20 words.
    • Descriptive statements: <= 25 words.
  2. Imperative Verbs:
    • Begin instruction steps with active verbs (Specify, Configure, Inspect, Run).
  3. Clarity and Precision:
    • Avoid ambiguous modals (should, might). Use must for requirements, can for capabilities.
    • Restrict consecutive noun clusters to 3 words or fewer.

Step 5: Generate the Final Audit Report

Structure the final audit report using this markdown format:

# 📋 Documentation Gap Audit Report

## 1. Audit Scope & Baseline

- Baseline: `<COMMIT_HASH>` (<SUMMARY>)
- Target Head: `<HEAD_HASH>`
- Audit Mode: `<AUTOMATIC | RELEASE | CUSTOM>`

## 2. Code Changes vs. Document Mapping Gaps

| Change Type | Component / File | Detected Code Change    | Affected Documentation    | Status      |
| ----------- | ---------------- | ----------------------- | ------------------------- | ----------- |
| New Skill   | `skill-name`     | Added to bundled_skills | `guides/skills.md`        | 🔴 Missing  |
| Tool Flag   | `browser`        | Added `--sandbox` arg   | `guides/builtin-tools.md` | ⚠️ Outdated |

## 3. Translation Parity Findings (KO ↔ EN)

- Total Files Audited: XX
- Truncated / Outdated EN Files: X
- Parity Gaps List:
  - `path/to/doc.md`: Ratio XX%, missing sections: [...]

## 4. Actionable Remediation Plan

1. [P0] Update `guides/skills.md` (KO & EN) to include `new-skill`.
2. [P1] Re-translate `path/to/doc.md` to ASD-STE100 standards.
3. [P2] Run `pnpm --filter libragent-docs build` to verify clean build.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.