Summary

A refactoring guidance skill that instructs agents to restructure code while preserving behavior. It establishes clear boundaries for safe refactoring: no feature changes mixed with structural edits, one ownership move at a time, and verification before and after changes. The skill prioritizes intermediate buildable/testable states and explicit scoping of exceptions.

Detected Capabilities

Trigger Keywords

Phrases that agents use to match this skill to user intent.

Use Cases

• Extract functions or modules from monolithic code
• Consolidate duplicated logic across files
• Move code ownership between modules or teams
• Reorganize project structure without changing behavior
• Clean up dependencies and imports while maintaining compatibility
• Rename and restructure public interfaces with safety verification

Quality Notes

• Strong emphasis on behavior preservation through verification boundaries
• Clear constraints (no feature changes, one ownership move at a time) that reduce risk
• Practical guidance on intermediate states (buildable/testable)
• Concise and actionable principles that an agent can follow
• Limitations section missing — does not explicitly address what types of refactors are out of scope (e.g., algorithm changes, major dependency rewrites)
• No concrete examples or walkthrough of how to establish/run proofs
• Missing guidance on when refactoring should stop or fail (e.g., if tests fail, if performance degrades)