agentleFS
Cogni-AI-OUCogni-AI-OU/cogni-ai-agent-skills/docs-review/SKILL.md

docs-review

Skill4 starsChanged 6 months agoMITView on GitHub

Enforce documentation quality, completeness, and mutual consistency across architecture files, ADRs, runbooks, READMEs, and code-level documentation. You MUST load this skill when asked to review or check consistency of documentation (such as *.md/*.mmd files).

25 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

10 rules. 0 are common, 7 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 24 edited versions we read keep these.

  • 7

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    After any significant architectural decision to verify documentation exists and is accurate.

    line 20
  2. Distinctive:

    When encountering complex code to ensure it's properly documented.

    line 21
  3. Distinctive:

    When verifying an operational procedure or tradeoff.

    line 22
  4. Very specific:

    When the user explicitly requests purely formatting/linting fixes (use markdown or yaml skills).

    line 27
  5. Distinctive:

    Look for contradictions, outdated references, missing updates, or structural mismatches between files.

    line 41
  6. Distinctive:

    Check if any of the documentation lags behind the current state of the repository.

    line 43
  7. Distinctive:

    CRITICAL INSTRUCTION: ONLY apply changes to these files if actual inconsistencies, contradictions, or outdated information are found.

    line 45
  8. Very specific:

    Verify: Every significant decision has an ADR (typically stored in docs/decisions/).

    line 51
  9. Distinctive:

    Verify: Every non-obvious code block has a "why" comment.

    line 55
  10. Very specific:

    Verify: Every on-call alert has a linked runbook (typically stored in docs/runbooks/).

    line 60

“Common” is measured: kept by at least 6 in 10 of the 24 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
<!-- markdownlint-disable MD013 MD023 MD031 MD032 -->

Docs Review

Guidance for reviewing core architecture, documentation, and constraint files for mutual consistency and accuracy. Use this skill when checking repository documentation.

When to Use

  • Checking core architecture files (*.mmd, *.mzn) for consistency.
  • Reviewing documentation (**/AGENTS.md, **/README.md) for outdated references.
  • After any significant architectural decision to verify documentation exists and is accurate.
  • When encountering complex code to ensure it's properly documented.
  • When verifying an operational procedure or tradeoff.

When Not to Use

  • For performing deep code execution or logic validation (use tester or code-review instead).
  • When the user explicitly requests purely formatting/linting fixes (use markdown or yaml skills).
  • For writing completely new features from scratch without reviewing existing architecture.

Common Pitfalls

  • Overzealous Editing: Modifying architectural constraint files (CONSTRAINTS.mzn) arbitrarily without understanding the global impact, leading to runtime failures.
  • Surface-Level Checks: Verifying that a markdown file exists without actually reading its content to ensure it accurately reflects the current codebase.
  • Ignoring Dependency Context: Updating a README file while neglecting the corresponding AGENTS.md or ADR, creating localized inconsistency.

Core Process

  1. Analyze Relationships: Cross-reference the relationships, definitions, rules, and configurations defined across these files.
  2. Identify Inconsistencies: Look for contradictions, outdated references, missing updates, or structural mismatches between files.
  3. Verify Up-to-Date Status: Check if any of the documentation lags behind the current state of the repository.
  4. Conditional Edits: CRITICAL INSTRUCTION: ONLY apply changes to these files if actual inconsistencies, contradictions, or outdated information are found. If all files are mutually consistent and up-to-date, do not make any edits—simply report back that the files are in order.

Architectural Decision Records (ADRs) Verification

Verify: Every significant decision has an ADR (typically stored in docs/decisions/).

Code-Level Documentation Verification

Verify: Every non-obvious code block has a "why" comment. Verify: Code-level comments don't state the obvious.

Runbooks Verification

Verify: Every on-call alert has a linked runbook (typically stored in docs/runbooks/).

README Currency Verification

Verify: README reflects current state (not v1 state). Verify: Setup instructions work on a fresh machine. Verify: Architecture diagrams are up to date.

Subdirectory Agents (AGENTS.md) Verification

Verify: Subdirectories with specific functionality have local AGENTS.md context files. Verify: Local AGENTS.md aligns with the global repository invariants.

Verification Checklist

  • ADRs written for significant architectural decisions
  • Non-obvious code blocks have "why" comments
  • Every production alert has a linked runbook
  • README is current and setup instructions work
  • Subdirectories with specific functionality have local AGENTS.md context files

Related Skills

  • critical-thinking: You MUST load this skill when deconstructing complex documentation and identifying structural contradictions.
  • docs-writer: Guidance for writing and maintaining documentation. You MUST load this skill when asked to write, document, or generate new documentation.

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.