vibe-review-code

Perform a structured code review on a pull request by analyzing git diffs to identify correctness errors, dead code, and security violations, then generating a Markdown report with severity buckets and minimal fixes.

Dead code detection identifies unused functions and variables during code review. It fits into the process by flagging regression risks and API boundary violations alongside blocking or major severity issues in the final report.

Yes, you can use static analysis to validate fixes against reviewer feedback by checking source changes for regression risks and API boundary violations, ensuring corrections meet project policies and test coverage expectations.

Yes, vibe3 inspect and lint work with local branch diffs by evaluating symbol-level lookups and callers to determine change scope, validating against Python standards and LOC limits before opening a pull request.

The best way to categorize code review findings by severity is to bucket issues into Blocking, Major, Minor, and Nit classifications, mapping file and function level findings with minimal fixes and line references.

You should not use automated code review for source changes when the scope lacks clear git diffs or when changes require subjective architectural judgment beyond dead code, API boundary, and regression risk detection.