Reduces documentation size while maintaining perfect execution equivalence for AI agents.
The shrink-doc skill optimizes Claude-facing documentation by condensing verbose explanations and reorganizing content into a more efficient hub-and-spoke model. It employs a validation-driven workflow that compares compressed versions against original baselines to ensure instructions remain functionally identical for AI agents while significantly reducing token usage. By distinguishing between Claude-facing configurations and human-facing docs, it safely streamlines your project's internal guidance system without compromising readability for developers.
Características Principales
01Validation-driven compression with execution equivalence scoring
02Intelligent filtering to prevent modification of human-facing documentation
03Smart reorganization of CLAUDE.md into efficient hub-and-spoke patterns
042 GitHub stars
05Section-header protection for style detection and rule-based files
06Automated baseline preservation and persistent versioning for safe rollbacks
Casos de Uso
01Optimizing slash command prompts and agent definitions for faster execution
02Consolidating redundant project instructions across multiple documentation files
03Shrinking oversized CLAUDE.md files that consume too much context
