claude:refactor-skill

To refactor a monolithic SKILL.md file, extract detailed content into references/ files and parameterize hardcoded values into a config.yaml. This modularization process creates a cleaner main file using progressive disclosure patterns for better maintainability.

You need to modularize your skill configuration when your SKILL.md file exceeds 500 lines, contains hardcoded project-specific values, or would benefit from progressive disclosure patterns. This refactoring addresses unmanageable, monolithic files that are difficult to maintain and update.

The best way to extract hardcoded values from a large skill file is to parameterize them into a config.yaml file. This separates project-specific settings from the core logic, allowing for easier customization and better code organization without modifying the main skill structure.

Yes, you can use shell commands to analyze, extract, and manipulate file content to split a complex skill file. This refactoring process moves detailed algorithms and explanations to references/ files, breaking down large monolithic structures into smaller, manageable components.

Progressive disclosure works effectively for managing complex skills over 1000 lines by moving detailed explanations to references/ files. This creates a concise main SKILL.md, a config.yaml for settings, and separate reference files, making the skill much easier to work with and update.