audit-documentation

A skill for checking Markdown documentation against the files and configuration in a project, then fixing safe factual errors. Markdown is a plain-text format commonly used for README files and other project docs.

In plain words
What is it for?
It is for reviewing project documentation for accuracy and consistency across documents. By default it checks tracked Markdown files while excluding several named planning and memory files.
Why use it?
It helps find claims that no longer match the code, commands, paths, or configuration. It limits edits to evidence-backed documentation fixes and leaves uncertain points clearly identified.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/conn-castle/agent-layer/audit-documentation
Any agent
npx skills add conn-castle/agent-layer --skill audit-documentation
Clone the repo
git clone --depth 1 https://github.com/conn-castle/agent-layer

Made for: Claude Code, Codex.

Per session 30 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 462 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00030 $0.00462
Opus 5 $0.00015 $0.00231
Sonnet 5 $0.00006 $0.00092
Haiku 4.5 $0.00003 $0.00046

Measured 3d ago against content hash 0468e77bae8f, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

audit-documentation scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

internal/templates/skills/audit-documentation/SKILL.md · 57 lines

What it actually says

audit-documentation

Audit Markdown against repository evidence; fix safe inaccuracies.

Scope

  • Use supplied scope; otherwise audit all tracked *.md files.
  • Exclude ISSUES.md, BACKLOG.md, DECISIONS.md, COMMANDS.md, and CONTEXT.md. Use /audit-memory when the user requests memory files.
  • A finding limit caps reporting, not coverage. Empty scope returns no-findings.

Contract

  • Validate actionable claims against files, configuration, symbols, scripts, and contracts. An unexecuted command is not verified behavior.
  • Fix the smallest evidence-backed error that could materially mislead. If docs express better intended behavior than code, report a code gap. Preserve ambiguous content and name the missing decision.
  • Do not edit source, tests, memory files, or product policy.

Workflow

Record scope and inspect enough context to establish its claims. For large scopes, parallelize non-overlapping read-only investigations; the owning agent validates candidates, resolves conflicts, and edits.

Check:

  • referenced commands, paths, configuration keys, and interface names
  • architecture and workflow claims against current code and configuration
  • contradictions, renamed concepts, and drift across in-scope documents
  • templates or examples that disagree with their canonical source

Mark statically indeterminate claims unverified. Ignore style, fix supported inaccuracies, and leave code gaps or unresolved intent unchanged. Do not log the audit to memory.

Write a concise summary with:

  1. # Documentation Audit Summary — scope and verdict
  2. ## Fixes Applied
  3. ## Code Gaps
  4. ## Decisions Needed

Findings include severity, location, evidence, impact, outcome (fixed, code-gap, or needs-user-decision), and type (command, path, config, interface, architecture, or cross-doc). Use None for empty sections. Finish when every document is checked and every material finding has an outcome.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 57 lines · 30 tokens per session scan A 0468e77bae8f

Subscribe to this mod's changes

audit-documentation is a skill published in the GitHub repository conn-castle/agent-layer (10 stars, last pushed 4d ago), licensed MIT. It adds 30 tokens to every session and 462 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories