requesting-code-review

To catch code defects before merging to main, you can use an automated code review process that dispatches a dedicated subagent with your git base and head SHAs, requirements document, and work summary to generate structured, severity-categorized feedback.

A structured code review categorizes feedback by severity, separating Critical, Important, and Minor issues to provide a clear merge readiness assessment. This approach prevents unvetted code with bugs or requirement mismatches from cascading into production.

Yes, automated code review works with subagent-driven development, plan execution, and ad-hoc feature workflows. It serves as a mandatory review checkpoint after task completion, before refactoring, and prior to merging to your main branch.

To review code changes against a feature plan, you need to provide the git base and head SHAs for the change range, a summary of the implemented work, and the associated plan or requirements document to generate accurate, structured feedback.

Dispatching a separate subagent for peer code review provides focused context of your changes and requirements while avoiding bias from your session history. This isolated approach enforces a standard review output with clear merge readiness assessment.

You should run a code review check after task completion, before refactoring, and prior to merging to main. These mandatory checkpoints catch requirement deviations and architectural flaws early, eliminating costly rework after deployment.