Documentation Audit

VerifiedSafe

Audits all project documentation by cross-referencing it against the actual codebase. Checks environment variables, file paths, Docker configs, and feature descriptions, then identifies gaps and applies fixes.

Sby Skills Guide Bot
DocumentationIntermediate
906/2/2026
Claude Code
#documentation-audit#code-cross-reference#accuracy-check#gap-analysis

Recommended for

Our review

Audits project documentation for accuracy and completeness by cross-referencing against the actual codebase.

Strengths

  • Catches stale documentation
  • Identifies missing documentation
  • Ensures environment variables match schema
  • Verifies file paths exist

Limitations

  • Requires documentation to exist in a recognized format
  • May miss implicit documentation outside files
  • Cannot verify logical correctness of descriptions
When to use it

When you need to ensure project documentation is up-to-date and consistent with the codebase before a release or handoff.

When not to use it

When documentation is intentionally minimal or your project is in early development without stable code.

Security analysis

Safe
Quality score90/100

The skill only reads documentation and code to audit and fix documentation using the Edit tool. It does not execute external commands, exfiltrate data, or perform destructive actions. The operations are confined to improving documentation accuracy.

No concerns found

Examples

Full documentation audit
Run a documentation audit on this project: check README.md, CLAUDE.md, .env.example, and all SKILL.md files against the codebase. Report stale items, missing items, and apply fixes.
Check environment variable docs
Verify that all environment variables mentioned in the documentation match the keys in scripts/config_schema.py's ENV_MAP. Fix any discrepancies.
Find undocumented features
Scan the codebase for features that are not covered in any documentation file and list them as missing items.

Documentation Audit

Audit all project documentation for accuracy and completeness, cross-referencing against the actual codebase.

Usage

/documentation

Instructions

Step 1: Inventory

List all documentation files in the project:

  • README.md
  • CLAUDE.md
  • .env.example
  • All SKILL.md files under .claude/skills/
  • Significant inline documentation (Python docstrings, shell script comments)

Use Glob and Read tools to locate and read each file.

Step 2: Cross-Reference Code

For each documentation file, verify the following against the actual codebase:

  • Environment variables — All environment variable names mentioned must match keys in scripts/config_schema.py's ENV_MAP and ALIASES. Use Grep and Read to check.
  • File paths and directories — All file paths, directory references, and command examples must actually exist and work. Use Glob to verify paths.
  • Build arguments and Docker config — All build arguments, Docker image names, and port numbers must match Dockerfile and docker-compose.yml. Use Read to compare.
  • Feature descriptions — Feature descriptions must match actual code behavior. Read the relevant source code to verify.

Step 3: Identify Gaps

Check for:

  • Features present in code but missing from documentation.
  • Deprecated or removed features still mentioned in documentation.
  • Inconsistencies between different documentation files (e.g., README.md says one thing, .env.example says another).

Step 4: Report and Fix

Produce a summary of findings organized as:

  • Accurate items — Documentation that correctly reflects the codebase.
  • Stale items — Documentation that references removed or changed features.
  • Missing items — Features or configuration present in code but not documented.

Apply fixes directly to the documentation files using the Edit tool. For each fix, explain what was changed and why.

Related skills