What problem does it solve? After code is built against a plan, it is easy to trust a green test suite and miss live failures, broken consumers, or plan deviations. This Skill runs a full post-build review in a single chat without dispatching subagents, catching defects before commit. ## Core Features & Use Cases - Build and test re-verification: Reruns the build and test suite in the current session and quotes the output rather than trusting prior results. - Plan conformance and stack-trap audit: Gates the diff against the plan's tasks, decisions, and log points, then checks it against the stack's known data-access, lifecycle, and concurrency traps. - Live probing and contract tracing: Boots the app through its production composition root, probes failing inputs, and traces changed wire contracts to their consumers. - Use Case: After an implementer finishes a streamed export endpoint, run this review to discover that a malformed query parameter returns a live 500 instead of 400, and that a sibling web client still expects the old response shape. ## Quick Start Ask the AI to review the build and check the code against its plan before committing.
