apply-progressive-disclosure

Reduce startup token costs by fragmenting large CLAUDE.md and SKILL.md files into lazy-loaded targeted references and scripts using progressive disclosure. This minimizes the tokens loaded on startup to optimize Claude Code configurations.

Progressive disclosure for Claude Code skill management is an optimization technique that fragments large configuration files into lazy-loaded components. It analyzes token budgets across project roots and skill directories to minimize startup load and deduplicate content.

Analyze and optimize token budgets by generating before/after context reports across project roots and skill directories. This workflow requires a CLI analysis tool and fragmentation-rules references to guide manual migration and deduplication of configuration artifacts.

Yes, you can fragment SKILL.md files to minimize tokens loaded on startup. The progressive disclosure process operates on skill directories containing SKILL.md, extracting code and applying fragmentation rules to create targeted references and scripts.

Yes, applying progressive disclosure requires a CLI analysis tool to evaluate token budgets and generate before/after reports. Optional scripting is also utilized to fragment documents and extract code during the optimization workflow.

A limitation of fragmenting CLAUDE.md is that it requires manual migration guided by generated before/after context reports. The fragmentation process relies on specific rules and optional scripting, meaning deduplication and optimization are not fully automated.