Review Pull Request

Arguments: $ARGUMENTS — a PR number or URL, and/or a mode. BUGS, SECURITY, or PERFORMANCE narrows the review to that axis; otherwise apply all.

Input

A diff pasted inline → review that text, nothing to fetch. Otherwise the PR number or URL, else the current branch's PR → gh pr view + gh pr diff. No PR found → ask for one.

Approval standard

Approve when the change definitely improves overall code health, even if it isn't perfect. Don't block a change because it isn't how you would have written it. If it improves the codebase and follows its conventions, approve.

If the change is too large to review well (~1000+ lines), asking the author to split it is a valid outcome. Suggest a strategy: stack (small change, next one based on it), by file group, horizontal (shared code first, then consumers), or vertical (one end-to-end slice per PR).

Review criteria

Review every changed file, not just the latest commit. Review the tests first: do they test behavior, cover the edge cases, and would they catch a regression? With the branch checked out locally, answer the last one by experiment: invert one condition the change adds, run the suite, restore the file. Still green → a missing test; name it. Reviewing alongside other agents → skip the experiment: test-auditor runs mutants in a scratch copy, and an in-place edit would race it.

• Correctness: logic bugs, off-by-ones, race conditions, unhandled states, missing error paths
• Readability: functions > 50 lines, nesting > 2 levels, dead code, unused imports
• Security: exposed secrets, unvalidated external data
• Type safety & immutability: any or unjustified assertions; in-place mutation of shared data
• Patterns: consistent with codebase conventions, no reinvented wheels
• Performance: unnecessary re-renders, O(n²) where O(n) works
• Code smells: match the diff against the baseline in smells.md — always judgement calls; the repo's documented style overrides the baseline
• Dependency upgrades (manifest/lockfile in the diff): one dependency per change — a bulk bump that breaks hides which package did it; verify against the changelog, not the version number; review the lockfile diff (one direct bump pulls dozens of transitive changes); flag hand-edited lockfiles

Output

Group by severity, file:line on every finding, a suggested fix for every Critical:

• Critical — must fix before merge (bugs, vulnerabilities)
• Suggestions — improvements worth considering
• Nit — minor and optional; label true nitpicks as Nit rather than dropping or inflating them
• FYI — informational
• Positives — good patterns to call out
• Verified — what you ran (tests, build, experiments) and what you only read

One structural problem and ten nits: the structural problem is the review — lead with it. Close with the verdict against the approval standard.