Skip to content

Skills / review

review

Review a diff for correctness, clarity and risk, and report findings by severity. Use when asked to review changes, a pull request or a commit. Do not use to write the fix; report first, change code only when asked.

OfficialAuditA
Latest
v1.1.0
Aisle
code, review
Works with
Claude Code, Codex, Cursor
Stars
2

Install

Install review

npx @agentsaisle/cli add review

Installs into your project's skills folder. Add --global to install it for every project.

Audit

A
Critical
0
High
0
Medium
0
Low
0

3 files · audited 2026-10-29 · rules 1.0.0

No findings.

What this skill can use

Tools

  • Read
  • Grep
  • Glob
  • Bash(git diff *)
  • Bash(git log *)
  • Bash(git show *)
  • Bash(git status *)

Scripts

None

Domains mentioned

None

Automated checks found no known risky patterns. This is not a guarantee of safety. How audits work · Read the rules

Show this grade in a README

The badge updates when the registry is rebuilt. Paste this Markdown into the README:

[![aisle audit](https://img.shields.io/endpoint?url=https%3A%2F%2Fagentsaisle-vercel.vercel.app%2Fbadges%2Freview.json)](https://agentsaisle-vercel.vercel.app/skills/review)

SKILL.md

Rendered from commit 4036c95

Review

Use when

  • You are asked to review a diff, a branch, a pull request or a commit.
  • The user wants a second pair of eyes before merging.

Don't use when

  • The user wants the problems fixed. Report first; edit only when asked.
  • There is no change to look at.

Steps

  1. Get the diff (git diff, a branch comparison or the pull request) and read the description or commit messages for intent.
  2. Check correctness first: wrong conditions, missing error paths, off-by-one errors, empty or null cases, concurrency, resource leaks.
  3. Check the tests: do they cover the new behavior and the failure cases, and would they fail if the code were wrong?
  4. Check risk: security, data loss, breaking changes to public behavior, performance on large inputs.
  5. Only then note clarity: naming, structure, duplication, comments that no longer match.
  6. For each finding give where (file:line), what is wrong, a concrete input that triggers it and a suggested fix. Say plainly when a point is a preference.

Budget

  • Search first (grep, glob, symbol search). Read line ranges, not whole files.
  • Read at most 8 files per step. If you need more, write down what you learned and narrow the question.
  • Do not re-read a file you just wrote or edited.
  • Prefer the smallest command output that answers the question: head, tail, quiet flags, filtered lists.
  • Read the diff first and open surrounding code only for the lines you question.

Output

  • Findings ordered by severity, each with location, problem, trigger and fix.
  • If nothing serious turned up, say so. Do not invent findings to look thorough.

Done when

  • Every changed file was considered and each finding names a line.
  • No file was edited unless the user asked.

Version history

VersionCommitInstall
1.1.0latest4036c95add review

Report this skillAgent Aisle on GitHub