What problem does it solve? Code review feedback often gets addressed superficially — comments are accepted without verifying the referenced code, or dismissed without technical justification. This Skill enforces a disciplined protocol for handling PR comments, reviewer suggestions, and automated tool findings so every item is read, verified against the actual code, and either fixed or pushed back with evidence. ## Core Features & Use Cases - Six-Step Protocol: Read, Understand, Verify, Evaluate, Respond, and Implement each review item individually, with fixes applied sequentially and tests run after each change. - Source-Based Trust Calibration: Applies different verification rigor depending on whether feedback comes from the user, external reviewers, or automated tools like linters and AI reviewers. - Mandatory Adversarial Review: Runs a scoped adversarial review over changed files after implementation, handling CRITICAL, WARNING, and INFO findings by severity. - Use Case: A reviewer leaves five PR comments claiming a function mishandles null values and suggesting a refactor. The Skill verifies each claim against the actual code, fixes the two valid bugs, pushes back on one incorrect claim with file:line evidence, and runs the test suite after every fix. ## Quick Start Ask the assistant to process the latest PR review comments on your branch using the receive-review protocol and address each item one at a time.