requesting-code-review

Code review ensures quality by dispatching a dedicated reviewer subagent to analyze your changes. Provide the base and head Git SHAs, describe what you implemented, and reference your plan or requirements. The subagent categorizes issues by severity and offers recommendations before merge.

Include the base and head Git commit SHAs, a description of what was implemented, and your original plan or requirements. This gives the code-reviewer subagent the context needed to assess whether changes align with intent and catch issues early.

Request code review after completing each task in subagent-driven development, after implementing major features, before merging to main, when stuck, before refactoring, or after fixing complex bugs. It acts as a quality gate to prevent issues from cascading.

The code-reviewer subagent structures feedback by severity level—Critical, Important, and Minor—with specific recommendations for each. This prioritization lets you address the most impactful issues first and act on feedback systematically.

This skill requires Git SHAs to identify changes between commits. If your workflow doesn't use Git or version control that provides commit identifiers, you'll need to adapt the input format or use an alternative approach for capturing code deltas.

Code review focuses on adherence to requirements, design patterns, and potential logic issues through human reasoning via a subagent. Automated testing catches bugs through execution. Both are complementary—review catches design problems early; tests verify correctness at runtime.