Summary
GitHub changed Copilot code review from a file-reading reviewer into an agent that can run builds, tests, targeted scripts, tools and APIs behind the Copilot agent firewall. Lite reviews now use an ensemble of agents and merge their findings into one review. GitHub reports 47% more addressed high-severity comments, 31% more medium-severity comments and about 8% lower review cost in its experiments. Copilot also auto-resolves addressed threads and writes commit messages for applied fixes.
Why it matters
For teams already using Copilot review, validation is moving from static inspection toward execution-backed review without requiring the highest effort tier. Trial it on repositories with reliable tests and compare high-severity precision, runtime and false positives against the previous Lite baseline.
Technical details
| Review Runtime | Full Copilot SDK shell-tool set behind the Copilot agent firewall; can run builds, tests, targeted scripts, and query available tools/APIs |
|---|---|
| Lite Ensemble | Multiple review agents contribute perspectives and GitHub merges their findings into one Lite review |
| Vendor Experiment | +47% addressed high-severity comments; +31% medium; +11% low; review cost about -8% |
| Workflow | Auto-resolves comments after a later commit addresses them; generates context-specific commit messages for applied suggestions |
Tags
copilotcode-reviewmulti-agenttool-useevaluationobservability