implement-refactor

A safe code refactor requires an orchestrated workflow using characterization tests, isolated worktrees, and transformation waves. By validating tests and types after each wave, you preserve behavior while migrating code end-to-end.

A transformation wave is a dependency-aware scheduled phase of an architected refactor that applies code changes sequentially. Each wave is followed by test and type validation to ensure behavior is preserved during the migration.

You set up an isolated environment for code migration by creating a worktree. This worktree-based workflow ensures refactors are executed safely away from the main branch while maintaining progress tracking and decision logging.

Yes, you can automate test generation and documentation during a structured refactor by spawning specialized agents. A test-writer agent handles characterization tests while a documenter agent ensures key docs exist and produces a final summary.

You need dependency-aware scheduling for a code transformation when executing an architected refactor with multiple transformation waves. It coordinates the order of changes so behavior is preserved and validation runs successfully after each wave.

Orchestrating refactors with multiple agents requires an existing plan with characterization, transformation waves, and reviews. It is limited to structured migrations where key docs exist and cannot safely handle unstructured code changes without dependency-aware scheduling.