Agent Code Review Swarm

This spins up multiple specialized AI agents to review your pull requests, each with a different focus: security vulnerabilities, performance bottlenecks, architecture patterns, and style conventions. It hooks into GitHub CLI to pull PR diffs, runs parallel agent reviews checking things like OWASP issues and N+1 queries, then posts findings back as comments or review requests. The interesting bit is the configurable thresholds where security issues can block merges while style suggestions stay advisory. Best for teams that want deeper analysis than linters provide but don't want to manually coordinate different review perspectives. The learning from past reviews feature could reduce noise over time, though you'll need to tune the agents to your codebase's quirks.

Install to Claude Code

npx -y skills add ruvnet/ruflo --skill agent-code-review-swarm --agent claude-code