How does the parallel review work?

The skill uses a Task tool to launch multiple specialized sub-agents—such as security-reviewer and performance-reviewer—simultaneously. Each agent focuses on its specific domain and their findings are aggregated into a single cohesive report.

Can I review specific pull requests?

Yes, by using the --pr flag followed by the PR number, the skill uses the GitHub CLI to fetch the diff and provide a detailed analysis of the changes in that specific pull request.

What focus areas can I specify?

You can narrow the review to specific domains using the --focus flag. Supported areas include architecture, security, testing, performance, and style.

Where are the code review reports saved?

Reports are automatically saved as Markdown files in the docs/reviews/ directory of your project, named with the date and branch/PR identifier for easy reference.

Does it support local changes?

Absolutely. You can use the --staged flag to review only changes in your git index, or run it without arguments to compare your current branch changes against the main branch.

Multi-Agent Code Review

Name: Multi-Agent Code Review
Author: TomazWang

byTomazWang

0•

협업 도구

Orchestrates specialized AI agents to perform comprehensive, parallel code reviews focusing on security, architecture, performance, and testing.

This skill enhances your development workflow by transforming Claude into a sophisticated code review coordinator. It leverages specialized sub-agents to analyze code changes from multiple perspectives simultaneously, ensuring high-quality standards are met before merging. Whether reviewing a specific GitHub pull request, a historical commit, or local staged changes, the skill provides deep insights into security vulnerabilities, architectural patterns, and performance bottlenecks, aggregating findings into a structured report with actionable recommendations.

주요 기능

01Automated report generation saved directly to your repository documentation for audit trails.

02Integrated task management that offers to create development tasks for identified critical and high-priority issues.

03Parallel multi-agent execution for comprehensive analysis of architecture, security, testing, performance, and style.

040 GitHub stars

05Flexible scoping to analyze specific files, staged changes, individual commits, or entire branch diffs.

06Native integration with GitHub CLI to review pull requests and post inline comments automatically.

사용 사례

01Performing a final security and architecture audit on a complex pull request before it hits production.

02Running a quick style and performance check on local staged changes to maintain code quality during active development.

03Generating a comprehensive quality report for a specific commit hash to debug regression or compliance issues.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tomazwang/agent-toolkit code-review

For use in Claude.ai and ChatGPT

Download Skill