Requesting Code Review

This formalizes the practice of dispatching a code-reviewer subagent to catch issues before they compound. You grab git SHAs, fill out a template with what you built and what it should do, then dispatch the reviewer subagent to analyze the diff. The skill pushes review after every task in subagent-driven development, which sounds aggressive but makes sense if you're trying to prevent cascading problems. It includes severity levels (Critical, Important, Minor) and clear guidance on when to fix versus when to push back. The example shows real workflow integration with specific commit SHAs and actionable feedback. Honest take: the mandatory review after each task will slow you down, but probably less than debugging compounded issues three tasks later.