agentleFS
Sign inStart free
ArangoGutierrezArangoGutierrez/claude-toolkit/.claude/skills/python-review/SKILL.md

python-review

Skill4 starsChanged 3 months agoMITView on GitHub

Python code review specialized for AI/MLOps/agent codebases — typing (pyright), async correctness, agent-loop invariants, ML reproducibility, model serialization safety. Triggered by "review Python", "review this agent code", "review the training pipeline", or /python-review

We split this skill into the rules it gives an agent

3 rules. 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 7 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    When invoked by the pr-review dispatcher, read references/python-review-checklist.md, review ONLY the changed lines in the PR diff, and …

    line 38
  2. Distinctive:

    Don't flag style that black/ruff already handle.

    line 57
  3. Distinctive:

    Don't demand type annotations on private helpers in an untyped codebase — type the public boundaries first.

    line 58

“Common” is measured: kept by at least 6 in 10 of the 7 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

Python Review

Systematic Python review for AI/MLOps/agent codebases. Flags correctness, typing, async, agent-loop, and reproducibility issues — not style that formatters own.

Process

  1. Static analysis (run only the tools the project already configures; never install anything into the project):
    • pyright on changed files when the project uses it (or mypy if that is the project's checker — check pyproject.toml / mypy.ini / setup.cfg).
    • ruff check on changed files when the project configures ruff.
  2. Walk checklist (see references/python-review-checklist.md):
    • Core correctness & typing, async correctness, agent-building invariants, ML reproducibility & serialization safety, packaging & environment.
  3. Report findings:
    • file:line for each issue
    • Category: correctness / security / reproducibility / performance
    • Severity: must-fix / should-fix / consider
    • Suggested fix (code snippet)

Dispatched mode (pr-review integration)

When invoked by the pr-review dispatcher, read references/python-review-checklist.md, review ONLY the changed lines in the PR diff, and return a findings list. Each finding has:

  • file — repo-relative path
  • line — NEW-file / RIGHT-side line number
  • description — 1–2 sentences
  • category — correctness / security / reproducibility / performance
  • severity — must-fix / should-fix / consider
  • reason — which checklist item flagged it

Take no external actions in dispatched mode — findings are returned as data.

Scope

Changed files only unless explicitly asked for a full-package review.

Gotchas

  • Don't flag style that black/ruff already handle.
  • Don't demand type annotations on private helpers in an untyped codebase — type the public boundaries first.
  • Respect the project's framework choices; review within them, don't rewrite.
  • Guard every static-analysis command: run it only if the repo configures it.

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.

Check your setup, free