requesting-code-review

To run a pre-commit security scan on staged git changes, you can trigger an agent to verify your cached diff, executing targeted static checks on added lines to block secrets and injection vectors before committing.

Baseline-aware testing prevents regressions in your git diff by capturing existing project test and lint failures first, then blocking only the new test failures or lint errors introduced by your staged changes.

The best way to auto-fail code review on unparseable output is using a fail-closed independent reviewer subagent that requires strict JSON output, automatically failing the review if the response cannot be parsed.

To handle large diffs during a pre-commit code review, the verification pipeline splits the staged changes by file, ensuring the independent reviewer and security scans process manageable chunks without missing added lines.

Yes, you can detect project test and lint tools automatically before pushing, as the pre-commit workflow identifies your project's existing test stack and lint configurations to run baseline-aware checks against your staged changes.

The independent reviewer subagent sees just the diff and scan results, receiving only your newly staged changes to provide an isolated security and logic check without full repository context.