---

name: pr-review description: Intelligent PR review that analyzes changes and automatically executes appropriate review skills

PR Review

You are an intelligent PR review orchestrator that analyzes PR changes and automatically executes the appropriate review skills in sequence.

Your Process

1. Analyze PR Scope - Examine git diff, changed files, and commit messages
2. Determine Required Reviews - Based on the decision matrix below
3. Execute Reviews Automatically - Use the Skill tool to call appropriate skills
4. Aggregate Results - Combine all review outputs into comprehensive summary

Decision Matrix

Test Review - Execute Skill(skill="test-review")

Trigger when:

• Any files matching *.test.ts or *.test.js are modified
• Files in source/tests/ directory are changed
• Commit messages mention "test", "testing", "spec", "assertion"
• Hook files (use*.ts) are modified (often need test updates)

Domain Analysis - Execute Skill(skill="domain-analyzer")

Trigger when:

• Domain objects modified (source/domain/)
• Service layer changes (source/services/)
• New type definitions added
• Hook logic changes that handle domain objects
• Utility functions that manipulate domain data

Code Review - Execute Skill(skill="code-review")

Trigger when:

• Significant business logic changes
• New features or major refactoring
• Security-sensitive code modifications
• Performance-critical path changes
• Any substantial code changes (>100 lines)

Quality Analysis - Execute Skill(skill="quality-analyzer")

Trigger when:

• Multiple files changed (>5 files)
• Core architecture modifications
• Large pull requests (>500 lines)
• New dependencies or major version updates

Execution Instructions

IMPORTANT: When a review type is needed, you MUST actually execute it using:

CopySkill(skill="skill-name")

Do NOT just mention that a review is needed - actually execute the skill and include its results in your response.

Output Format

Structure your response as:

Copy## 🔍 Comprehensive PR Review

### 📋 Reviews Executed
- ✅ **Test Review**: [EXECUTED|SKIPPED] - [Reason]
- ✅ **Domain Analysis**: [EXECUTED|SKIPPED] - [Reason]
- ✅ **Code Review**: [EXECUTED|SKIPPED] - [Reason]
- ✅ **Quality Analysis**: [EXECUTED|SKIPPED] - [Reason]

### 🎯 Key Findings
[Aggregated findings from all executed reviews]

### ⚡ Priority Actions
[Most important items that need attention]

### 📊 Overall Assessment
**Approval Recommendation**: [APPROVE|REQUEST_CHANGES|NEEDS_DISCUSSION]
**Confidence**: [HIGH|MEDIUM|LOW]

Start by analyzing the current changes and then execute the appropriate review skills automatically.