requesting-code-review

Automate code reviews by dispatching a code-reviewer subagent to inspect implementation against requirements, catching issues before they cascade. This enforces mandatory reviews after each task, before merging to main, and at critical checkpoints without manual delays.

Request code reviews after completing major features, after each subagent-driven task, before merging to main, and optionally when stuck, before refactoring, or after fixing complex bugs. Structured review points ensure consistent quality and early issue detection.

The skill automatically identifies BASE_SHA and HEAD_SHA to generate precise diffs for review, then dispatches the code-reviewer subagent with a structured payload containing the comparison. This enables targeted feedback on exact changes without manual SHA lookup.

Code review feedback is categorized by severity—critical, important, and minor issues. Address each according to its level: resolve critical issues immediately, schedule important ones in your workflow, and consider minor suggestions for future refactoring or debt reduction.

Yes. Mandatory reviews at key development stages catch issues before they compound, reducing bugs that reach production and preventing technical debt accumulation. Early detection through consistent automation maintains code quality throughout the workflow.

Basic Git familiarity helps but isn't mandatory—the skill handles BASE_SHA and HEAD_SHA extraction automatically. You need to understand when to trigger reviews in your workflow and how to respond to feedback, not Git internals.