Refactor Code

Workflow

1. Establish current observable behavior, compatibility constraints, and the specific structural problem to solve.
2. Inspect repository patterns and define a narrow refactor boundary. Do not expand scope merely because nearby code is imperfect.
3. Identify available verification before editing: focused tests, types, lint, builds, snapshots, or stable runtime checks.
4. Plan reversible, incremental changes that keep review and failure diagnosis straightforward.
5. Refactor consistently with the repository, favoring simplification over new architectural ceremony.
6. Run meaningful behavior-preservation checks after each risky step and at the end.
7. Inspect the final diff for accidental behavior changes, compatibility breaks, unrelated formatting, and unnecessary abstractions.

Constraints

• A refactor needs a concrete reason: duplicated logic, unclear boundaries, excessive complexity, poor testability, or a repeated error-prone pattern.
• Do not mix feature work or unrelated defect fixes into the refactor.
• Preserve APIs, data formats, ordering, side effects, and error behavior unless the user explicitly requests a compatibility change.
• Do not add dependencies or abstraction layers solely for stylistic preference.
• Report validation limits and any behavior that could not be proven equivalent.