refactoring

Refactoring a large file starts with read-only architecture analysis to inventory symbols and map dependency impacts. The process outputs a YAML plan with target files, budgets, and migration steps for a worker-coder to execute.

Architecture analysis for code decomposition is a read-only scan that inventories lines, symbols, and responsibilities within a chosen scope. It detects boundary violations and flags refactoring smells like mixed concerns and risky coupling.

Yes, you can use a YAML plan to decompose a module without changing behavior. The plan designs new module boundaries and delegates code changes to worker-coder tasks with verification gates to prevent unexpected behavior shifts.

Yes, refactoring analysis detects boundary violations and risky coupling. It flags common refactoring smells like overly large files, mixed concerns, duplication, and dependency impacts to map the blast radius of changes.

The best way to split a large file and extract new boundaries is generating a structured refactoring plan. It applies cohesion heuristics, maps dependency impacts, and outputs a YAML contract with migration steps for safe execution.

You should not use an automated refactoring plan when the target scope lacks clear symbols or dependencies to map. The analysis requires symbol discovery and dependency impact mapping to enforce file and class cohesion heuristics accurately.