code_review

A strategic code review for architectural planning is performed by spawning an architect subagent that analyzes your codebase to provide design pattern recommendations and systematic guidance without directly writing code.

Analyzing technical debt and root causes is achieved by passing a specific task and relevant files to an architect subagent, which drives structured analysis to identify systemic issues and evaluate design flaws.

You cannot use this for direct code edits during major feature work. It provides evaluation, planning, and debugging guidance focused on architecture, requiring a separate delegation skill for actual code modifications.

To start system design analysis, you invoke the architect subagent with a clear task definition and a list of relevant files, adding optional parameters like responsibility and git diff details to drive the structured analysis.

You should not use an architect subagent for codebase debugging when your task requires direct code modifications, as this approach is strictly reserved for deep architectural analysis, planning, and design evaluation.