Agentic Architecture for Code Review

GitHub Copilot code review now runs on an agentic tool-calling architecture, enabling it to gather broader repository context as needed when analyzing pull requests. The system can now reference relevant code, directory structures, and architectural references to provide feedback that reflects how changes fit into the larger codebase.

Key Improvements

With this architectural shift, developers can expect:

• Higher-quality findings that prioritize correctness and architectural integrity
• Lower noise with comments surfacing meaningful signal rather than volume
• More actionable guidance to help resolve issues quickly and confidently

This results in more targeted review comments that help teams merge with confidence and ship cleaner code faster.

Deployment and Setup

Copilot code review runs on GitHub Actions. Most users require no additional setup. However, if you've opted out of GitHub-hosted runners, you'll need to complete a one-time self-hosted runners setup to receive agentic code reviews on pull requests. Alternatively, code review can run on GitHub-hosted larger runners.

Availability

The feature is now generally available for all users with Copilot Pro, Copilot Pro+, Copilot Business, and Copilot Enterprise subscriptions.