Showing 28 verified skills. 284 preview entries are hidden until we confirm a real source. Show preview skills · Why?
Run 6 specialized review checks on every PR automatically
claude install anthropics/claude-code/pr-review-toolkitComprehensive PR review that dispatches specialized agents for different concerns, then synthesizes a unified review.
This is the actual SKILL.md file that powers this skill. Copy it to install.
---
name: pr-review-toolkit
description: |
Trigger on "review PR", "check my PR", "review this pull request", or when
the user shares a PR URL or number. Dispatches 6 specialized review agents.
allowed-tools:
- Bash(git *)
- Bash(gh *)
- Read
- Grep
- Glob
---
# PR Review Toolkit
6 specialized review agents that each focus on a different concern, then
synthesize into a unified review.
## Prerequisites
- Git repository with a PR branch
- Base branch available for comparison
## Steps
1. **Get the diff:**
```bash
git diff main...HEAD
```
2. **Run 6 specialized review agents:**
**Agent 1 - Comments & Documentation:**
Check for misleading comments, missing JSDoc/docstrings on public APIs,
TODO/FIXME items, and outdated documentation.
**Agent 2 - Test Coverage:**
Verify new code has tests. Check: are edge cases covered? Do tests
assert the right things? Are there flaky test patterns?
**Agent 3 - Error Handling:**
Find uncaught exceptions, missing try/catch, unhandled promise rejections,
silent failures, and unclear error messages.
**Agent 4 - Type Design:**
Review type definitions, check for `any` usage, verify generics are
constrained, ensure types are exported where needed.
**Agent 5 - Code Quality:**
Check for code smells: long functions, deep nesting, magic numbers,
duplicated logic, unclear variable names.
**Agent 6 - Simplification:**
Look for code that can be simplified: unnecessary abstractions,
over-engineering, complex conditionals that can be flattened.
3. **Synthesize results** into a single review with:
- Overall assessment (approve / request changes / comment)
- Grouped findings by severity
- Specific line references for each finding
4. **Post review** (if using GitHub):
```bash
gh pr review <number> --comment --body "<review>"
```
mkdir -p ~/.claude/skills/pr-review-toolkit~/.claude/skills/pr-review-toolkit/SKILL.mdResulting file structure:
~/.claude/
skills/
pr-review-toolkit/
SKILL.md <-- skill definitionSkills are loaded automatically by Claude Code when you start a new session. The skill name and description in the frontmatter determine when Claude triggers it.
Recommended from shared domain, career, and tool overlap with PR Review Toolkit
Query databases, inspect schemas, and explore data from your AI editor
Both used by Software Engineer
Inspect Redis keys, TTLs, and cache hit rates from your AI editor
Both used by Software Engineer
Query Sentry errors and analyze stack traces from your AI editor
Both used by Software Engineer
Manage Stripe payments and subscriptions from your AI assistant
Both used by Software Engineer
Manage Supabase databases, auth, and storage from your editor
Both used by Software Engineer
Control a browser from your AI editor - navigate, screenshot, interact
Both used by Software Engineer
PR Review Toolkit