bmad-os-review-pr

To automate adversarial code review, this Skill performs a dual-layer analysis combining adversarial skepticism with automated edge-case detection on GitHub pull requests, generating professional, actionable engineering feedback.

Standard pull request reviews often miss logic gaps and security risks due to a lack of rigorous depth. An adversarial review approach actively hunts for boundary conditions, null dereferences, and unhandled exceptions to resolve this inconsistency.

Yes, performing automated pull request analysis requires the GitHub CLI and local repository access. These dependencies are necessary to execute diff analysis and post the resulting review comments back to the specific pull request.

The best way to find edge cases and null dereferences is using a dual-layer review mechanism. It automatically scans code for boundary conditions and unhandled exceptions while an adversarial persona evaluates logic gaps and security risks.

To post professional engineering feedback, the review Skill operates on specific pull request identifiers. It merges adversarial findings and edge-case detections into a neutral report, then uses the GitHub CLI to post comments.