# Review code with AI swarm agents

Code reviews require manual effort and may miss security issues or performance problems. This skill coordinates multiple AI agents to review pull requests across security, performance, style, and architecture dimensions simultaneously.

## Install

```bash
npx skillstore add claude code flow/dnyoussef-github-code-review
```

## Metadata

- - Slug: dnyoussef-github-code-review
- - Version: 1.0.0
- - Author: Claude Code Flow
- - GitHub username: DNYoussef
- - License: MIT
- - Repository: https://github.com/DNYoussef/ai-chrome-extension/tree/main/.claude/skills/github-code-review
- - Ref: main
- - Supported tools: Claude, Codex, Claude Code
- - Risk level: safe
- - Risk factors: network, filesystem, external\_commands, env\_access
- - Quality score: 67
- - Quality tier: warning
- - Public page: https://skillstore.pages.dev/skills/dnyoussef-github-code-review
- - Manifest: https://skillstore.pages.dev/api/skills/dnyoussef-github-code-review/manifest

## Capabilities

- Deploys specialized AI agents for security, performance, style, and architecture reviews
- Analyzes pull requests using GitHub CLI integration
- Generates contextual review comments with severity levels
- Supports custom review agents and quality gate configuration
- Integrates with GitHub Actions for automated review workflows

## Use Cases

- Multi-agent security review: Deploy security-focused AI agents to analyze pull requests for vulnerabilities before merging
- Automated PR quality gates: Set up automated review workflows that check code quality, style, and documentation standards
- CI/CD review integration: Integrate AI-powered code review into GitHub Actions pipelines for continuous quality assurance

## Prompt Templates

### Basic PR review

```
Use the github-code-review skill to review PR 123 for security issues and code quality. Post the results as a comment.
```

### Security-focused review

```
Run a security-only review on PR 456 using the security review agent. Check for OWASP vulnerabilities and suggest fixes.
```

### Performance analysis

```
Analyze PR 789 with the performance review agent. Identify bottlenecks and suggest optimizations with benchmarks.
```

### Complete swarm review

```
Initiate a comprehensive review on PR 321 using security, performance, style, and architecture agents. Post all findings to the PR.
```

## Limitations

- Requires GitHub CLI \(gh\) and ruv-swarm tools to be installed
- Cannot execute code or run tests directly
- Review quality depends on AI model capabilities and context provided
- Labels and comments require appropriate GitHub permissions

## Best Practices

- Configure quality gates with clear thresholds for blocking merges on critical issues
- Use label-based agent assignment to route PRs to appropriate specialized reviewers
- Combine AI reviews with human oversight for final approval decisions

## Anti Patterns

- Relying solely on AI reviews without human verification of critical changes
- Ignoring severity classifications and treating all issues as equally important
- Running reviews on very large PRs without incremental or focused review modes

## Security Audit

- - Safe to publish: true
- - Audited at: 2026-01-17T03:28:29.889\+00:00
- - Summary: This skill contains ONLY markdown documentation. No executable code, no file system operations, no network operations, and no credential access. All external command references are within documentation examples only. The static scanner flagged pattern matches in bash examples and security documentation concepts, but these are illustrative examples for legitimate GitHub CLI and ruv-swarm tool usage. No malicious intent or actual security risk present.

## Stats

- - Views: 343
- - Downloads: 3
- - Favorites: 0
- - Popularity score: 0