一键导入
reviewer
// Review code changes from git commit history for quality, security, and adherence to codebase patterns. Use when the user asks to review commits, review code changes, review git history, or perform a code review on recent changes.
// Review code changes from git commit history for quality, security, and adherence to codebase patterns. Use when the user asks to review commits, review code changes, review git history, or perform a code review on recent changes.
Orchestrate multiple skills through a phased workflow to solve problems, generate ideas, create plans, or build software. Use when the user asks to coordinate roles, run a full workflow, or when a task benefits from brainstorming, challenging, planning, and implementing in sequence.
Generate brainstorm ideas from multiple thinking frameworks and perspectives. Use when the user asks to brainstorm, ideate, explore options, think through a problem, generate alternatives, or needs creative input on any topic.
Aggressively challenge solutions, implementations, designs, and ideas by finding potential issues across performance, UX, security, bugs, scalability, maintainability, and edge cases. Use when the user asks to challenge, critique, stress-test, poke holes in, or review code, architecture, or ideas.
Analyze UI and UX for usability improvements across web, mobile, and CLI interfaces. Suggest and implement changes to make things easier to use and understand. Use when the user asks to review UI/UX, improve usability, simplify an interface, audit accessibility, or optimize user flows.
Create step-by-step plans for solving problems, debugging issues, building features, or architecting complex solutions. Always includes pros/cons analysis and risk consideration. Use when the user asks to plan, create a roadmap, break down a task, design an approach, outline steps, or needs a structured path forward.
Write unit tests for code changes across multiple stacks including JavaScript/TypeScript (Jest, Vitest), React (Testing Library), PHP (PHPUnit), Python (pytest), Django (TestCase), Hono.js, Express.js, and Flutter (Dart). Use when the user asks to write tests, add test coverage, create unit tests, or when making code changes that need test coverage. Automatically suggests tests for new or modified code.
| name | reviewer |
| description | Review code changes from git commit history for quality, security, and adherence to codebase patterns. Use when the user asks to review commits, review code changes, review git history, or perform a code review on recent changes. |
Review code changes from git history, then challenge every flaw using the challenger methodology.
The user may provide:
<start_hash> to <end_hash>HEAD~1..HEAD), stashed uncommited changes or unstashed changesCopy this checklist and track progress:
Review Progress:
- [ ] Step 1: Resolve commit range
- [ ] Step 2: Gather change context
- [ ] Step 3: Read changed files for pattern context
- [ ] Step 4: Analyze changes
- [ ] Step 5: Produce challenge report
Determine the commit range based on user input:
| Input | Git range |
|---|---|
Two hashes: <start> <end> | <start>..<end> |
One hash: <hash> | <hash>~1..<hash> |
| No hashes | HEAD~1..HEAD |
Run: git log --oneline <range> to confirm the commits exist and list them.
Run these commands to collect the full picture:
# List all changed files
git diff --name-status <range>
# Get the full diff with context
git diff <range>
# Get commit messages for intent
git log --format="%h %s%n%b" <range>
For each modified file (not deleted), read the full current version to understand:
For new files, find similar files in the same directory or module to establish expected patterns:
# Find sibling files for pattern reference
ls <directory_of_new_file>/
Read 1-2 sibling files to understand local conventions.
Review each change against these criteria:
Pattern Adherence (primary focus alongside challenger dimensions):
Apply the Challenger methodology from /Users/hbai/.cursor/skills/challenger/SKILL.md. Read that skill and follow its challenge dimensions and output format.
Add one extra dimension to the challenger report:
Structure the final output as:
## Code Review: <short description of changes>
### Commits Reviewed
- `<hash>` <message>
- ...
### Files Changed
- A: <new files>
- M: <modified files>
- D: <deleted files>
## Challenge Report
### Critical (must fix)
- **[DIMENSION]**: Description. Why it's dangerous. Attack/failure scenario.
### High Risk (strongly recommend fixing)
- **[DIMENSION]**: Description. Impact. Scenario.
### Medium Risk (should consider)
- **[DIMENSION]**: Description. Impact. Scenario.
### Low Risk (nitpick / hardening)
- **[DIMENSION]**: Description.
### Summary
X critical, Y high, Z medium, W low issues found.
Verdict: SAFE / RISKY / DANGEROUS