What problem does it solve? Skills often grow into bloated, poorly structured SKILL.md files that mix stable methodology with accumulating knowledge, trigger unreliably, and accumulate dead references. This Skill provides a proven method to design new skills or refactor existing ones with clean separation of concerns. ## Core Features & Use Cases - Skill Creation (Phases 1-4): Frame a single user intent, choose an architecture pattern (Template Method, Facade, Pipeline, Strategy), write the frontmatter description as a public interface, and review with a lint checklist. - Skill Refactoring (Phase 0): Diagnose mixed responsibilities, growth-inducing instructions, and dead references in an existing skill, then restructure content verbatim with zero information loss. - SOLID Transposed to Skills: Apply single-responsibility by change velocity, progressive disclosure (thin SKILL.md facade plus on-demand references), and append-only knowledge files with dated sections. - Use Case: Your SKILL.md has grown to 300 lines mixing workflow and accumulated pitfalls. Run the refactor path to split knowledge into dated reference files while keeping the core workflow under 70 lines. ## Quick Start Ask the AI to refactor this skill because the SKILL.md is too large and mixes workflow with accumulated knowledge.
