github-code-review

Coordinate multi-agent code reviews for GitHub pull requests.

1|Updated Nov 13, 2025
One-click install
npx skills add https://github.com/Cornjebus/amair --skill github-code-review-cornjebus
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: github-code-review
Source: https://github.com/Cornjebus/amair/tree/main/.claude/skills/github-code-review
Command: npx skills add https://github.com/Cornjebus/amair --skill github-code-review-cornjebus

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Manual code reviews are often slow, inconsistent, and prone to missing critical issues like security vulnerabilities, performance bottlenecks, or architectural flaws, especially in large codebases or fast-paced development environments.

Core Features & Use Cases

  • Multi-Agent Review System: Deploy specialized AI agents (security, performance, architecture, style, accessibility) for comprehensive, parallel code analysis.
  • PR-Based Swarm Management: Dynamically create and manage review swarms based on Pull Request context, labels, and complexity.
  • Intelligent Comment Generation: Generate contextual, actionable review comments with code examples and suggested fixes directly on GitHub PRs.
  • Automated Workflows: Trigger auto-reviews on PR creation/synchronization and enforce quality gates with configurable thresholds.
  • Context-Aware Reviews: Analyze PRs with full project context, learning from historical reviews and performing cross-PR analysis.
  • Use Case: Automatically initiate a comprehensive review for a critical PR, deploying security, performance, and architecture agents to identify issues, post intelligent comments, and enforce quality gates before merging.

Quick Start

Initialize a multi-agent review swarm for a Pull Request (PR 123): gh pr view 123 --json files,diff | npx ruv-swarm github review-init --pr 123 Post an initial review status comment to the PR: gh pr comment 123 --body "🔍 Multi-agent code review initiated"

Frequently Asked Questions about github-code-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate code reviews for GitHub pull requests?

Automate code reviews by deploying multi-agent swarm coordination on GitHub PRs. Initialize review agents across security, performance, architecture, and style domains to analyze diffs in parallel, then post contextual comments and enforce quality gates automatically on PR creation or synchronization.

Can I use AI agents to catch security vulnerabilities and performance issues in pull requests?

Yes. Specialized AI agents review code for security vulnerabilities, performance bottlenecks, architectural flaws, and accessibility gaps across the entire PR diff. Each agent analyzes its domain in parallel, aggregates findings, and generates actionable review comments with suggested fixes.

What's the best way to enforce code quality gates on GitHub PRs?

Enforce quality gates by configuring automated thresholds that trigger on multi-agent review completion. The swarm orchestrates agents to analyze PR complexity and domain-specific risks, then blocks or approves merges based on configurable quality criteria before manual review.

Does this work with large codebases and fast-paced development workflows?

Yes. Multi-agent parallel analysis scales to large codebases by deploying specialized review agents simultaneously rather than sequential manual review. Context-aware analysis learns from historical reviews and performs cross-PR analysis to maintain consistency across fast release cycles.

How do I integrate automated code review comments directly into GitHub?

Integration uses GitHub CLI to fetch PR metadata and diffs, then posts intelligent review comments with code examples and fixes directly to the PR. The swarm coordinates comment generation across agents and timestamps reviews tied to PR synchronization events.

What if I only need to review specific domains like security or performance?

Configure the review swarm to deploy only the agents you need—security, performance, architecture, style, or accessibility—rather than all domains. Each agent focuses on its specialty, so you can prioritize reviews based on PR risk profile or organizational standards.