agentleFS
YougLin-devYougLin-dev/Aha-Loop/.agents/skills/doc-review/SKILL.md

doc-review

Skill181 starsChanged 9 months agoMITView on GitHub

Reviews and cleans up outdated documentation. Use periodically to maintain documentation health. Triggers on: review docs, clean up documentation, check for stale docs.

15 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

11 rules. 0 are common, 11 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 9 edited versions we read keep these.

  • 11

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Check for staleness indicators

    line 15
  2. Distinctive:

    Verify referenced code/files still exist

    line 16
  3. Distinctive:

    After each milestone completion

    line 27
  4. Distinctive:

    When documentation feels outdated

    line 33
  5. Distinctive:

    When onboarding reveals confusion

    line 35
  6. Distinctive:

    Not modified in 60+ days

    line 155
  7. Distinctive:

    No recent git commits mention the file

    line 158
  8. Distinctive:

    Update file paths (if clear mapping exists)

    line 172
  9. Distinctive:

    Remove broken external links

    line 173
  10. Distinctive:

    Update version numbers in dependency lists

    line 174
  11. Distinctive:

    Update docs with code: - When changing code, update related docs

    line 283

“Common” is measured: kept by at least 6 in 10 of the 9 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

Doc Review Skill

Systematically review documentation for staleness, accuracy, and relevance.


The Job

  1. Scan all documentation files
  2. Check for staleness indicators
  3. Verify referenced code/files still exist
  4. Identify outdated content
  5. Generate cleanup report
  6. Optionally apply fixes

When to Use

Scheduled

  • After each milestone completion
  • Monthly maintenance cycle
  • Before major releases

On Demand

  • When documentation feels outdated
  • After significant refactoring
  • When onboarding reveals confusion

Review Process

Step 1: Inventory Documentation

Scan for documentation files:

find . -name "*.md" -not -path "./.git/*" -not -path "./.vendor/*" -not -path "./.worktrees/*"

Common locations:

  • README.md (root and subdirs)
  • docs/ directory
  • AGENTS.md, CLAUDE.md
  • .claude/skills/*/SKILL.md
  • knowledge/ directory
  • scripts/aha-loop/templates/

Step 2: Check Staleness Indicators

For each document, check:

IndicatorCheck MethodThreshold
Last modifiedgit log -1> 30 days
Referenced filesgrep + verifyMissing = stale
Code examplesSyntax checkErrors = stale
Version numbersCompare to actualMismatch = stale
LinksHTTP checkBroken = stale

Step 3: Verify Code References

For each code reference in docs:

## Code Reference Check

File: [doc-path]
Reference: `src/main.rs:45-60`
Status: [Exists | Missing | Changed]

If Changed:
- Original: [what doc says]
- Current: [what code shows]
- Action: [Update doc | Flag for review]

Step 4: Check External Links

# Extract URLs
grep -oP 'https?://[^\s)]+' file.md

# Check each URL
curl -s -o /dev/null -w "%{http_code}" URL

Step 5: Generate Report

Create docs-review-report.md:

# Documentation Review Report

**Date:** [YYYY-MM-DD]
**Reviewed:** [N] files
**Issues Found:** [N]

## Summary

| Category | Count |
|----------|-------|
| Stale (>30 days) | [N] |
| Missing References | [N] |
| Broken Links | [N] |
| Outdated Examples | [N] |

## Issues by File

### [file1.md]
- [ ] Line 45: Reference to `src/old.rs` - file no longer exists
- [ ] Line 78: Code example uses deprecated API

### [file2.md]
- [ ] Line 12: Version "1.0.0" should be "2.0.0"
- [ ] Line 56: Broken link to external docs

## Recommended Actions

1. **Delete:** [files that should be removed]
2. **Update:** [files that need content updates]
3. **Review:** [files that need human review]

## Auto-fixable

The following can be auto-fixed:
- [ ] Update version numbers
- [ ] Remove dead links
- [ ] Update file paths

Run `./scripts/aha-loop/doc-cleaner.sh --fix` to apply.

Staleness Criteria

Definitely Stale

  • References files that don't exist
  • Contains code that doesn't compile
  • Links to 404 pages
  • Describes features that were removed

Probably Stale

  • Not modified in 60+ days
  • Uses old API patterns
  • References old version numbers
  • No recent git commits mention the file

Needs Human Review

  • Contains outdated but complex content
  • Describes deprecated but still-supported features
  • Historical documentation (changelogs, decisions)

Auto-Fix Rules

Safe to Auto-Fix

  • Update file paths (if clear mapping exists)
  • Remove broken external links
  • Update version numbers in dependency lists
  • Fix obvious typos in code references

Requires Human Review

  • Content changes beyond mechanical updates
  • Removing entire sections
  • Changing technical instructions
  • Modifying architecture decisions

Integration

With Doc Cleaner Script

# Generate report only
./scripts/aha-loop/doc-cleaner.sh --report

# Apply safe fixes
./scripts/aha-loop/doc-cleaner.sh --fix

# Interactive mode
./scripts/aha-loop/doc-cleaner.sh --interactive

With Observability

Log review activities:

## 2026-01-29 16:00:00 | Task: Maintenance | Phase: Doc Review

### Documentation Review
Scanned 45 files, found 7 issues.

### Key Findings
- AGENTS.md references outdated skill paths
- README example code uses deprecated API

### Actions Taken
- Auto-fixed 3 version number issues
- Flagged 4 items for human review

Checklist

Before completing review:

  • All markdown files scanned
  • Code references verified
  • External links checked
  • Report generated
  • Auto-fixes applied (if safe)
  • Human review items flagged

Example Review

Input

File: docs/api.md last modified 90 days ago

# API Documentation

## Authentication

See `src/auth/mod.rs` for implementation.

Use version 0.5.0 of the auth library:
```toml
auth-lib = "0.5.0"

More info: https://example.com/old-docs


### Review Findings

1. **Stale**: Last modified 90 days ago
2. **Missing Reference**: `src/auth/mod.rs` doesn't exist (moved to `src/api/auth.rs`)
3. **Outdated Version**: Current version is 1.2.0, not 0.5.0
4. **Broken Link**: https://example.com/old-docs returns 404

### Report Entry

```markdown
### docs/api.md

**Staleness:** 90 days since last update
**Issues:**
- [ ] Line 5: `src/auth/mod.rs` → `src/api/auth.rs`
- [ ] Line 8: Version 0.5.0 → 1.2.0
- [ ] Line 12: Broken link (404)

**Recommendation:** Update file paths and versions, remove broken link

Prevention

To reduce future staleness:

  1. Update docs with code - When changing code, update related docs
  2. Add doc checks to CI - Fail on broken references
  3. Schedule regular reviews - Monthly or per-milestone
  4. Use relative links - Easier to maintain than absolute
  5. Date-stamp examples - "As of v1.2.0"

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.