review-pr

star 9

Comprehensive PR review against team standards. Spins up multiple agents to review code design, structure, and quality.

babylonlabs-io By babylonlabs-io schedule Updated 2/23/2026

name: review-pr description: Comprehensive PR review against team standards. Spins up multiple agents to review code design, structure, and quality. disable-model-invocation: true argument-hint: [pr-number-or-url]

Spin up multiple agents in parallel to carefully review PR $ARGUMENTS. Each agent should focus on a different aspect of the review.

First, fetch the PR diff and details:

!gh pr diff $0 !gh pr view $0

Review Criteria

Agents should cover the following areas. Provide specific file references with line numbers and suggest concrete improvements.

1. Magic Numbers & Constants

  • Flag any hardcoded numbers or strings without explanation
  • Suggest extracting to named constants, enums, or configuration
  • Check for suspicious values (timeouts, limits, indices, hex values)

2. File Structure & Organization

  • Are new/modified files in the right location following project conventions?
  • Are files getting too large (>500 lines source, >1000 lines tests) and should be split?
  • Is related functionality co-located?
  • Are index files and exports properly maintained?

3. Method Placement & Cohesion

  • Are methods in the right file/module?
  • Flag methods with too many responsibilities (SRP violation)
  • Identify duplicate or near-duplicate logic across files
  • Check if utility functions belong in a shared module instead

4. Extraction Opportunities

  • Common patterns that should become reusable functions
  • Large methods (>50 lines) or components that should be broken down
  • Repeated logic that could be abstracted
  • Configuration that should be externalized

5. Code Quality

  • Type safety and null/undefined handling
  • Error handling completeness (missing catch, unhandled edge cases)
  • Naming clarity (variables, functions, files)
  • No over-engineering or unnecessary abstractions
  • No dead code or unused imports introduced

6. Architecture & Patterns

  • Adherence to existing project patterns and conventions
  • Unnecessary complexity or premature optimization
  • Performance concerns (unnecessary re-renders, expensive operations in hot paths)
  • Security concerns (injection, XSS, secrets exposure)
  • Proper separation of concerns

7. Testing

  • Are critical paths covered by tests?
  • Are edge cases and error conditions tested?
  • Are test descriptions clear and accurate?
  • Any flaky test patterns (timing, ordering dependencies)?
Install via CLI
npx skills add https://github.com/babylonlabs-io/babylon-toolkit --skill review-pr
Repository Details
star Stars 9
call_split Forks 6
navigation Branch main
article Path SKILL.md
More from Creator
babylonlabs-io
babylonlabs-io Explore all skills →