What changed
GitHub's Copilot code review now resolves its own addressed comments during rereview, generates commit messages for accepted fixes, uses the Copilot SDK's shell tools behind the agent firewall and runs an ensemble of agents at the Lite effort level.
Why it matters
Review automation gains more execution reach and can reduce thread cleanup, while the ensemble changes both review behavior and cost. The broader shell surface and automatic closure make missed findings and auditability more consequential.
What to do
Pilot on a non-sensitive repository, inspect the commands and network calls available during review, replay known defects, verify that only truly addressed comments close, and compare severity recall, false positives, latency and cost with the prior setup.
Caveat
GitHub reports 47%, 31% and 11% increases in addressed high-, medium- and low-severity comments and roughly 8% lower cost from its own experiments. These are vendor results without task mix, confidence intervals or independent reproduction.