requesting-code-review

To run a pre-commit code review, trigger the skill on your staged git diff. It scans added lines for security issues, runs baseline-aware tests and lint checks, and delegates logic validation to an independent fail-closed reviewer.

Baseline-aware testing compares your current test and lint results against a pre-existing failure baseline. This blocks only newly introduced issues from your staged diff, ignoring known historical test failures.

Yes, you can automatically fix security issues using the optional auto-fix loop. It attempts up to two targeted fix-and-reverify cycles for reported shell injection, secrets, or unsafe deserialization problems.

The static scan detects shell injection, eval, unsafe deserialization, secrets, and common SQL-injection patterns. It targets only the added lines in your staged git diff before passing results to an independent reviewer.

A fail-closed independent reviewer is a separate validation step that receives your diff and static scan output. It blocks the commit by returning a JSON-only verdict if any security or logic failures are detected.

Use automated pre-commit verification for git-based feature work and bug fixes involving multi-file edits. It prevents broken or insecure changes from reaching your repository by validating logic and security before you commit or push.