What problem does it solve?
Refactoring often silently breaks code when AI assistants rely on memory or assumptions instead of reading the original source, causing lost table columns, broken conditional branches, circular dependencies, and misaligned parallel lists. This Skill enforces a read-first, verify-after checklist workflow that catches these regressions before they ship.
Core Features & Use Cases
- Read-Before-Refactor Discipline: Requires reading complete original code (all if/else/switch branches) before restructuring, with comparison tables for tables, data structures, and field mappings.
- Ten Documented Pitfall Patterns: Covers circular dependencies from module extraction, state initialization conflicts in hooks, lost UI layouts in Modal/Dialog refactors, oversimplified render functions, field-name guessing, shotgun fixes, and distinct/filter misalignment in position-correlated lists.
- Multi-Language Examples: Provides Java, Go, and TypeScript implementations for paired-collection handling in the references directory.
- Use Case: When merging two tenant-specific table definitions into one shared component, the Skill forces a column-by-column comparison checklist so no column is dropped or renamed silently.
Quick Start
Ask the AI to refactor a component or merge duplicate logic, and it will apply the refactoring safety checklist to verify the result against the original code.