Code review has always been a context problem. The reviewer staring at a diff often needs to know things that aren't in the diff: what the linked Jira ticket says, what the internal API contract requires, what the incident post-mortem recommended. Until now, Copilot's code review was blind to all of that. That just changed.

GitHub has shipped two public previews for Copilot code review: support for custom agent skills and MCP server connections, plus a new "medium" analysis tier that routes complex pull requests to a higher-reasoning model. Both are available today for Copilot Pro, Pro+, Business, and Enterprise users.

The context gap Copilot couldn't cross

The agentic shift in development has led to disjointed workflows, more context switching, and too much time spent reviewing agent-generated code. The core problem was always the same: a lot of what reviewers need to know lives in other tools, not in the diff itself.

Agent skills and MCP bring that context into Copilot's reviews, ensuring that reviews don't stall on questions already answered elsewhere. This means senior engineers stop being the bottleneck for consistency across repositories.

Two new levers: skills and MCP

These two features are distinct but complementary. Here's what each one does:

• Custom agent skills: Custom agent skills invoke your team's internal tools and standards during a review, extending Copilot beyond its built-in analysis. You define them as markdown files in your repo.
• MCP server connections: Copilot code review can use MCP servers to pull context directly into the review from the third-party platforms and internal systems your team uses, including issue tracking, documentation, service catalogs, and incident tooling.