Copilot code review adds adaptive effort and secret detection previews
GitHub Copilot code review can choose Lite, Balanced or Max effort from pull-request complexity, use related public repositories as context and flag both patterned and model-detected secrets. Azure DevOps code review moved to general availability with audit and Azure billing controls.
Why it matters: Review depth, cross-repository reads and secret finding now influence cost and exposure automatically. Enterprise maximums constrain adaptive effort, but public context and probabilistic secret detection still need explicit false-positive and leakage tests.