What problem does it solve? Adding features or fixing bugs in large working codebases often leads to full-file rewrites that break dependencies, waste tokens, and introduce invisible regressions. This Skill enforces a patch-over-rewrite methodology so existing functionality stays intact. ## Core Features & Use Cases - 6-Phase Process: Audit the current file state, map priorities by dependency, apply targeted str_replace patches, validate syntax and feature presence before integration, verify dependency wiring, and run regression checks on all prior features. - Token Economy: Uses grep and ranged file views instead of re-reading entire files, keeping context usage low on files with thousands of lines. - Language-Specific Patterns: Covers TypeScript typing, React component decomposition thresholds, CSS/Tailwind safe changes, Python pitfalls, dead code elimination, dependency injection, strangler fig migration, and SQL query optimization. - Use Case: You need to add a PDF export feature to a 2000-line single-file React/HTML app. The Skill audits insertion points, patches only the necessary spots, validates JSX via Babel before integrating, and confirms every old feature still works. ## Quick Start Ask the AI to add a feature to your large existing file surgically without rewriting it, validating the change before integrating and regression-checking afterwards.
