Documentation Audit FAQs

Question 1

How does the skill verify that documentation claims are actually true?

Accepted Answer

It uses a multi-tier verification hierarchy. It cross-references text claims against canonical sources like OpenAPI specs and Protobuf files, uses symbolic analysis to check code structure, and performs deep investigations into behavioral claims to ensure your docs aren't lying to users.

Question 2

When is the best time to use the Documentation Audit skill?

Accepted Answer

It is most effective when onboarding to a new codebase, after significant refactoring, or when preparing for a new release. It ensures that 'doc rot' doesn't set in and that your READMEs reflect the current state of the project.

Question 3

What does the Documentation Audit skill do for my repository?

Accepted Answer

This skill performs a multi-pass analysis of your repository to ensure your Markdown documentation matches your actual source code. It extracts technical claims, verifies them against symbols and API contracts, identifies undocumented features, and repairs broken links or diagrams.

Question 4

Can it validate diagrams like Mermaid or PlantUML?

Accepted Answer

Yes. The skill parses text-based diagrams and compares them against your actual code structure. It distinguishes between 'normative' diagrams (which must match the code exactly) and 'illustrative' diagrams (which represent conceptual flows) to ensure visual aids remain technically relevant.

Question 5

Will this skill automatically overwrite my documentation?

Accepted Answer

No, it uses a risk-tiered repair system. While it can auto-fix low-risk items like typos and broken links, it requires explicit user approval for technical corrections, deletions, or structural changes. You can also run it in 'dry_run' mode to see a report before any changes are made.

Documentation Audit

Documentation Audit

主な機能

ユースケース

主な機能

ユースケース