Skip to content
Agentmart
Submit a tool

docs-reviewer

Review documentation diffs for docstring accuracy, README consistency, and stale examples (.md/.rst and docstring changes only).

Skill by caiobellizzi

IndexedMIT
Write a review

Chalk Labs

Need help deploying docs-reviewer?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/caiobellizzi-docs-reviewer

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Review documentation diffs for docstring accuracy, README consistency, and stale examples (.md/.rst and docstring changes only).

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/docs-reviewer

What's inside

Skill name
docs-reviewer
Description
Review documentation diffs for docstring accuracy, README consistency, and stale examples (.md/.rst and docstring changes only).
License
MIT
Instructions
83 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

## Docs Reviewer

Validates documentation quality in diffs that contain `.md`, `.rst` files, or changed docstrings. Dispatched as a shared-context SkillTool pass (D-04). Gated on the presence of documentation paths in the diff per D-03.

## Discipline

Act as a documentation accuracy gate. Ensure that documentation changes stay in sync with implementation changes in the same diff. Do not evaluate code correctness — focus on documentation fidelity.

Decision rules:
- Docstring accuracy: if a function/class signature changes in the diff, any docstring in the same diff must reflect the new signature. Stale parameter names or return type descriptions are Important.
- README consistency: if `README.md` is changed, verify that code examples in the diff match the actual public API in `diff.changes`. Stale examples that would fail if copy-pasted are Important.
- CHANGELOG entries: if the diff touches a `CHANGELOG.md` or `HISTORY.md`, verify the entry references the correct version identifier present elsewhere in the diff.
- Broken links: new Markdown links (`[text](url)`) that reference a file path not present in the repo or changed files are Important.
- Stale `TODO`/`FIXME` comments: new `TODO` comments introduced by the diff without a tracking reference (issue number, task ID) are Minor.

Never do:
- Never flag pre-existing stale docs outside the diff.
- Never require perfect prose or stylistic changes — focus on factual accuracy.
- Never modify the diff — read only (D-06).

Exhaustion behavior (D-09): skip and log an audit entry when budget is exhausted.

## Behavior

1. Parse `--query` JSON: `{task_id, plan, diff, context}`. Skip immediately if no `.md`, `.rst`, or docstring changes in `diff.changes`.
2. For each changed documentation file, inspect `+` lines only.
3. Cross-reference changed function signatures in the same diff against docstring parameter lists.
4. Validate README code examples against public API changes in the diff.
5. Check new Markdown links for file-path references that do not exist in the diff's known file set.
6. Check for new `TODO`/`FIXME` without tracking references.
7. Aggregate findings with severity:
   - `Critical`: documentation that contradicts the implementation in a way that would cause user errors (e.g., wrong function name, wrong parameter order).
   - `Important`: stale parameter name in docstring, stale README example, broken path link.
   - `Minor`: new TODO without reference, inconsistent capitalization in headings.
8. Set `passed: true` if no Critical findings exist.
9. Emit the Output Schema JSON to stdout.

## Output Schema

Emit exactly this JSON object to stdout (no surrounding text):

```json
{
  "role": "docs-reviewer",
  "passed": true,
  "findings": [
    {
      "file": "path/or/*",
      "line": 42,
      "problem": "string — what is wrong",
      "required_fix": "string — concrete fix required",
      "severity": "Important"
    }
  ],
  "tokens_used": 0,
  "cost_cents": 0
}
```

Fields:
- `role`: always `"docs-reviewer"`.
- `passed`: `true` if no Critical findings; `false` otherwise.
- `findings`: list of finding objects; empty list if `passed: true`.
- `severity`: one of `"Critical"`, `"Important"`, `"Minor"`.
- `tokens_used`: integer — total tokens consumed by this invocation.
- `cost_cents`: integer — cost in USD-equivalent cents.

## Using Pre-Injected Brain Context

When the `context` field contains a `## Brain: Repo Digest` block:
- Extract the relevant sections and apply them as constraints on your output
- DO NOT make a separate brain-query call — the digest is already pre-injected
- If a section is absent from the context, proceed without it (fail-open)

## Dry-run Behavior

If `--dry-run` appears in the trigger, emit the following hardcoded example and stop:

```json
{"role": "docs-reviewer", "passed": true, "findings": [], "tokens_used": 0, "cost_cents": 0}
```

From caiobellizzi/ultra-workshop, MIT.

More from caiobellizzi/ultra-workshop

Signals

Stars
0
Forks
0
Open issues
0
Last commit (this folder)
May 31, 2026 (5 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from caiobellizzi

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get docs-reviewer running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying docs-reviewer?

CallGet help
Cite this page

docs-reviewer skill. Agent Mart. https://agentmart.dev/skills/caiobellizzi-docs-reviewer (last updated 2026-10-24)