What problem does it solve? Structural verifiers confirm that tests exist and pass, but they cannot tell whether the code is actually correct or whether acceptance criteria are satisfied in spirit. This Skill performs a semantic, adversarial review of a single block's diff in a fresh-context subagent, catching logic bugs, missed edge cases, and superficially-passing tests before merge. ## Core Features & Use Cases - Three review lenses: Blind Hunter finds correctness bugs without trusting names or comments; Edge Case Hunter walks every branch, boundary, empty/error state, and concurrency path; Acceptance Auditor verifies each AC is genuinely satisfied and that discursive code rules hold. - Concurrency-claim auditing: dedicated checks that concurrency tests create real contention and that guarded operations are atomic rather than check-then-act (TOCTOU). - Finding triage and strict handoff: every finding is classified by severity (HIGH/MED/LOW) and disposition (Patch/Defer/Decision), producing a strict APPROVE/CHANGES/BLOCKED verdict consumed by the worker-composer. - Use Case: After the structural verifier passes on a block, the worker-composer dispatches this reviewer on the diff; it flags a sequential test masquerading as a concurrency test as HIGH, and the worker re-patches before merge. ## Quick Start Ask the worker-composer to run the mismagent code review on the current block's diff after the structural verifier passes.
