Code Review
Comprehensive multi-agent code review system that examines code from multiple specialized perspectives to catch bugs, security issues, and quality problems before they reach production.
Focused on
Multi-perspective analysis - Six specialized agents examine code from different angles
Early bug detection - Catch bugs before commits and pull requests
Security auditing - Identify vulnerabilities and attack vectors
Quality enforcement - Maintain code standards and best practices
Overview
The Code Review plugin implements a multi-agent code review system where specialized AI agents examine code from different perspectives. Six agents work in parallel: Bug Hunter, Security Auditor, Test Coverage Reviewer, Code Quality Reviewer, Contracts Reviewer, and Historical Context Reviewer. This provides comprehensive, professional-grade code review before commits or pull requests.
Quick Start
# Install the plugin
/plugin install code-review@NeoLabHQ/context-engineering-kit
# Review uncommitted local changes
> /code-review:review-local-changes
# Review a pull request
> /code-review:review-pr #123Agent Architecture
Code Review Command
│
├──> Bug Hunter (parallel)
├──> Security Auditor (parallel)
├──> Test Coverage Reviewer (parallel)
├──> Code Quality Reviewer (parallel)
├──> Contracts Reviewer (parallel)
└──> Historical Context Reviewer (parallel)
│
▼
Aggregated ReportCommands Overview
/code-review:review-local-changes - Local Changes Review
Review uncommitted local changes using all specialized agents with code improvement suggestions.
Purpose - Comprehensive review before committing
Output - Structured report with findings by severity
/code-review:review-local-changes ["review-aspects"]Arguments
Optional review aspects to focus on (e.g., "security", "bugs", "tests")
How It Works
Change Detection: Identifies all uncommitted changes in the working directory
Staged changes
Unstaged modifications
New files
Parallel Agent Analysis: Spawns six specialized agents simultaneously
Bug Hunter - Identifies potential bugs and edge cases
Security Auditor - Finds security vulnerabilities
Test Coverage Reviewer - Evaluates test coverage
Code Quality Reviewer - Assesses code structure
Contracts Reviewer - Reviews API contracts
Historical Context Reviewer - Analyzes codebase patterns
Finding Aggregation: Combines all agent reports
Categorizes by severity (Critical, High, Medium, Low)
Removes duplicates
Adds file and line references
Report Generation: Produces actionable report with prioritized findings
Usage Examples
# Review all local changes
> /code-review:review-local-changes
# Focus on security aspects
> /code-review:review-local-changes security
# After implementing a feature
> claude "implement user authentication"
> /code-review:review-local-changesBest practices
Review before committing - Run review on local changes before
git commitAddress critical issues first - Fix Critical and High priority findings immediately
Iterate after fixes - Run again to verify issues are resolved
Combine with reflexion - Use
/reflexion:memorizeto save patterns for future reference
/code-review:review-pr - Pull Request Review
Comprehensive pull request review using all specialized agents.
Purpose - Review PR changes before merge
Output - Detailed findings with line-specific comments
/code-review:review-pr ["PR number or review-aspects"]Arguments
PR number (e.g., #123, 123) and/or review aspects to focus on
How It Works
PR Context Loading: Fetches PR details and diff
Changed files
Commit messages
PR description
Base branch context
Parallel Agent Analysis: Same six agents analyze the PR diff
Each agent examines changes from their specialty perspective
Considers PR context and commit messages
Finding Aggregation: Combines findings with PR-specific context
Links findings to specific lines in the diff
Considers breaking changes and backward compatibility
Report Generation: Produces PR-ready report
Structured for easy review
Action items for the PR author
Usage Examples
# Review PR by number
> /code-review:review-pr #123
# Review PR with focus on security
> /code-review:review-pr #123 security
# Review current branch's PR
> /code-review:review-prBest practices
Review before requesting human review - Address automated findings first
Share report with team - Include findings in PR comments
Track patterns - Use findings to improve coding guidelines
Don't ignore low priority - Create issues for future improvement
Review Agents
Bug Hunter
Focus: Identifies potential bugs and edge cases through root cause analysis
What it catches:
Null pointer exceptions
Off-by-one errors
Race conditions
Memory and resource leaks
Unhandled error cases
Logic errors
Security Auditor
Focus: Security vulnerabilities and attack vectors
What it catches:
SQL injection risks
XSS vulnerabilities
CSRF missing protection
Authentication/authorization bypasses
Exposed secrets or credentials
Insecure cryptography usage
Test Coverage Reviewer
Focus: Test quality and coverage
What it evaluates:
Test coverage gaps
Missing edge case tests
Integration test needs
Test quality and meaningfulness
Code Quality Reviewer
Focus: Code structure and maintainability
What it evaluates:
Code complexity
Naming conventions
Code duplication
Design patterns usage
Code smells
Contracts Reviewer
Focus: API contracts and interfaces
What it reviews:
API endpoint definitions
Request/response schemas
Breaking changes
Backward compatibility
Type safety
Historical Context Reviewer
Focus: Changes relative to codebase history
What it analyzes:
Consistency with existing patterns
Previous bug patterns
Architectural drift
Technical debt indicators
Report Structure
Reviews produce structured output organized by severity:
# Code Review Report
## Executive Summary
[Overview of changes and quality assessment]
## Critical Issues (Must Fix)
- [Issue with location and suggested fix]
## High Priority (Should Fix)
- [Issue with location and suggested fix]
## Medium Priority (Consider Fixing)
- [Issue with location]
## Low Priority (Nice to Have)
- [Issue with location]
## Action Items
- [ ] Critical action 1
- [ ] High priority action 1Last updated