docs-audit_skill

This skill audits documentation against the current codebase, verifying accuracy, links, and docstring coverage to improve reliability.
  • TypeScript

25

GitHub Stars

1

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill outfitter-dev/agents --skill docs-audit

  • SKILL.md13.0 KB

Overview

This skill performs a comprehensive documentation audit against the current code state, producing a structured report with prioritized findings. It checks markdown accuracy, link validity, code example correctness, and docstring coverage while minimizing context usage. Reports are scaffolded and written to .pack/reports/ by default, with multi-file and custom-output options.

How this skill works

It discovers markdown files with git metadata, prioritizes a target set (stale, recent, or all), and runs deep analysis on the selected files. Deep checks include correctness of code examples and CLI snippets, internal link and anchor validation, completeness checks for required sections and exports, and docstring coverage scanning across detected languages. Findings are compiled into a YAML-frontmatter report (single or multi-file) and written to the generated path; a short summary and paths are returned for quick review.

When to use it

  • Before a release to ensure docs match code changes
  • When merging large refactors that may break examples or signatures
  • Periodic health checks on a documentation-heavy repo
  • Preparing onboarding or public-facing docs for customers
  • Investigating reports of broken links or outdated instructions

Best practices

  • Run discovery first to scope and prioritize rather than scanning everything
  • Use focus and limit flags to target highest-risk docs (stale or recent)
  • Prefer multi-file mode for detailed audits and single-file for quick runs
  • Keep .pack/reports/ gitignored for ephemeral runs, commit only curated reports
  • Create follow-up tasks for critical blocking issues to track fixes

Example use cases

  • Audit docs/ after API changes to update function signatures and examples
  • Run stale-focused audit on large monorepo to find docs lagging behind code
  • Generate a multi-file audit before a public launch to surface missing examples and docstrings
  • Run a quick link-only pass to fix internal anchors and image references
  • Measure docstring coverage per language to prioritize documentation workstreams

FAQ

Yes. Use the --output flag to write a single file or directory; the session ID remains in frontmatter for traceability.

How does it handle very large repositories?

It respects the --limit parameter and focus strategies (stale/recent/all) so you analyze a manageable subset and note the total file count in the report.

What languages are checked for docstrings?

Common languages (TypeScript/JavaScript, Python, Go, Rust) are detected by extension and scanned with lightweight grep-based summaries to compute coverage percentages.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational