agentleFS
mkolewmkolew/ai/skills/reviewer/reviewer-docs/SKILL.md

reviewer-docs

Skill0 starsChanged 56 days agoMITView on GitHub

Review documentation: Markdown (README, specs, guides, attribution files) and licence texts. Use when reviewing changes to .md files or licence and attribution files, or when the user types /reviewer-docs. Catches commands, paths, flags, script names and versions that no longer match the repo, docs that miss what the change added or removed, broken links and anchors, commands that fail when pasted, missing or wrong third-party attribution, invalid tables, and leaked secrets or private URLs. Part of the reviewer bundle.

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 2 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Look at the code change the docs belong to, so you know what should have been documented.

    line 14
  2. Very specific:

    Every command, script, flag, path and env var exists now and does what the doc says.

    line 20
  3. Distinctive:

    No secrets, tokens, internal hostnames or personal data.

    line 52
  4. Very specific:

    Use the reviewer format: [severity] path:line — problem, then failure: (the concrete input, state or platform → wrong result) and fix: (the …

    line 67

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “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

Docs reviewer

Scope: *.md, LICENSE*, attribution and font licence texts. HTML pages go to reviewer-web.

Before reviewing

  • List every concrete claim the changed docs make: commands, script names, flags, paths, env vars, versions, URLs, counts.
  • Look at the code change the docs belong to, so you know what should have been documented.

Checklist

Truth (the most common defect)

  • Every command, script, flag, path and env var exists now and does what the doc says. Grep for it, run --help, ls the path. Renamed scripts and flags are the usual miss.
  • Versions, counts and defaults match the code and config.
  • Behaviour described as current is current. No history ("we used to", "now uses", "Phase 2") outside a changelog.

Completeness

  • A new script, command, env var, setup step or user-visible behaviour in the change is documented where users look. Removed things are removed from the docs.
  • Setup steps work on a fresh clone in the order given.

Commands

  • Copy-pasteable: one command per fenced block where the reader runs them, no $ prompts, no output mixed into the block, and the working directory stated or obvious.
  • Fenced blocks have a language tag (bash, json, text).

Links

  • Relative links and #anchors resolve. External links point at the intended page.
  • No localhost, private hostnames or links that need a login, unless they are clearly local-only.

Licences and attribution

  • Third-party content (data, fonts, art, music, code) is credited where the licence requires it: licence name, link, author, and a note on changes for CC BY and CC BY-SA.
  • Font licences (OFL) ship with the fonts. The LICENSE file matches the licence declared in the package metadata.
  • Attribution required in the shipped product is actually bundled into the build, not only kept in the repo.

Form

  • Tables have the same number of columns on every row. Headings go in order without skipped levels.
  • Answer-first sections, short sentences, one term per concept throughout the document.

Safety

  • No secrets, tokens, internal hostnames or personal data.

Readability

Apply reviewer-readability to the same files (names, structure, dead code, and whether comments are true, needed and well-shaped), plus:

  • Docs readability is covered by the checklist above (truth, answer-first, one term per concept). Code samples inside the docs follow the readability rules of their language.

Not findings

  • Wording preferences when the text is correct and clear.
  • Line wrapping and other formatting that Markdown renders the same way.

Report

Use the reviewer format: [severity] path:line — problem, then failure: (the concrete input, state or platform → wrong result) and fix: (the smallest change). Most severe first. Prove each finding before reporting it, and say No findings with what you checked when there are none.

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.