Shape Copilot code review around your team
GitHub has introduced two public previews for Copilot code review: Agent skills and MCP (Model Context Protocol). These features enable teams to shape the review process around their specific tools and coding standards, allowing the review depth to scale with the complexity of each change. The Agent skills capability lets developers define custom review behaviors, while MCP provides a standardized way to integrate external context into reviews. This move aims to make Copilot code review more adaptable and relevant to diverse team workflows, moving beyond a one-size-fits-all approach. The previews are available now, and GitHub encourages teams to experiment with them to tailor the review experience.
Teams can now customize Copilot code review to match their specific tools and standards.