github-code-review

The github-code-review skill transforms standard code review workflows into a robust, multi-agent automated process. Designed for software engineering teams using GitHub, this skill leverages a swarm of specialized AI agents to inspect code changes beyond simple linting. By integrating with the GitHub CLI (gh), it allows developers to initiate deep analysis directly from the terminal, providing immediate feedback on pull requests based on specific domains like security vulnerabilities, architectural design patterns, performance bottlenecks, and adherence to team style conventions.

This skill is ideal for lead developers, DevOps engineers, and security teams who need to enforce quality gates without manual overhead. It supports complex workflows such as OWASP security checks, Big O complexity analysis, SOLID principle evaluation, and accessibility compliance. By automating the triage of code quality, the swarm coordination system helps reduce reviewer fatigue and ensures consistent standards are applied across every commit, significantly shortening feedback loops in high-velocity development environments.

• Multi-agent swarm orchestration for parallelized code analysis and domain-specific oversight.

• Integration with GitHub CLI to fetch PR diffs, metadata, and post automated status updates or change requests.

• Specialized agents including security, performance, architecture, style, and accessibility reviewers.

• Automated quality gates that can block or flag pull requests based on severity levels (e.g., critical security flaws).

• Context-aware analysis that learns from history and utilizes repository-level insights for more accurate suggestions.

• Requires GitHub CLI (gh) installed and configured with proper authentication to access repositories.

• Works optimally when provided with full PR context (files, diffs, titles, and body) via pipe commands.

• Supports various levels of depth, from basic syntax checks to comprehensive architectural deep-dives.

• Output can be directed to PR comments, labels, or status checks to integrate seamlessly into existing CI/CD pipelines.

• Users can trigger specific agents selectively to save compute resources or run the full suite for enterprise-grade audits.