refactor

To refactor messy code and reduce coupling, you extract modules, invert dependencies, and migrate to a healthier architectural structure. This process uses a diagnose, target, plan, and execute workflow to prevent behavior regressions while improving testability.

The best way to restructure code architecture without breaking behavior is using a test-driven migration plan. You execute atomic steps with TDD requirements, running tests and builds after each step while updating specifications to prevent spec drift and regressions.

You diagnose structural anti-patterns by generating a diagnose report that names anti-patterns, identifies hotspots, and determines a clear structural root cause. This diagnosis happen before proposing concrete refactoring targets and architectural boundaries.

Yes, dependency inversion and module extraction are core techniques for architectural redesign. They allow you to invert dependency directions and extract modules to untangle outdated codebases, differentiating structural refactoring from simple dead-code removal.

Yes, using an architecture decision record and a confirmation gate prevents spec drift by requiring explicit user agreement on proposed targets before execution. This ensures all atomic migration steps align with the confirmed architectural boundaries.

You should not use an architectural refactoring workflow for simple dead-code removal or minor cleanup. This process focuses on structural redesign, dependency inversion, and module extraction rather than routine maintenance tasks.