code-review

To automate code review for a git diff, you can run the skill in default self-review mode. It processes your diff inputs to generate a shortlist of structured findings with severity levels and file:line references.

The best way to structure code review findings is to use a mode-driven framework that outputs traceable results. This approach generates findings with clear severity levels and specific file:line references to guide remediation.

Yes, you can use this code review process for cross-team audits across diverse codebases and languages. It features a mode router that delegates to specific audit paths to generate consistent, traceable findings.

A mode-driven code review framework handles feedback processing by routing inputs to a specific feedback handler. This ensures that review comments and optional PR data are processed systematically to produce actionable findings and status updates.

Yes, the code review skill works in both local and PR contexts. It supports standard development workflows by satisfying inputs such as args, git diff, and optional PR data to produce findings and status.