Copilot code review settings are now manageable via REST and GraphQL APIs
As of Nov 6, enterprise, org, and repo admins can read and change Copilot code review enablement, default effort, runner config, and whether reviews use custom instructions and MCP tools through the API, and review comments now expose attribution, effort, and the skills or MCP servers cited as sources. Enterprise and org endpoints don't accept GitHub Apps or fine-grained tokens; Balanced reviews also now run three reviewer models plus a combiner.
