Will I lose information when optimizing my CLAUDE.md?

No. The skill follows a 'move, don't compress' rule, ensuring that 100% of the original content is preserved in either the main file or the reference folder.

How does Claude know when to look at the Level 2 reference files?

Level 1 contains a 'Reference Index' and 'Trigger Conditions' (such as specific error codes or task types) that explicitly tell Claude when a deeper dive into a specific document is necessary.

Why does this skill forbid using line count as a metric?

Shorter files aren't always better. The goal is to eliminate duplicate information and cognitive noise, not to arbitrarily cut useful instructions that Claude needs for specific tasks.

What is progressive disclosure in a CLAUDE.md file?

It is a strategy to keep the main CLAUDE.md file high-signal by moving detailed, low-frequency information into specialized reference files that are only accessed when specific triggers occur.

Claude.md Progressive Disclosure Optimizer

Name: Claude.md Progressive Disclosure Optimizer
Author: LuodashiV5

byLuodashiV5

•

开发者工具

Optimizes CLAUDE.md files using a multi-layered architecture to maximize information efficiency and AI performance.

This skill implements a 'Progressive Disclosure' framework for managing project-specific AI instructions within CLAUDE.md files. By organizing content into 'Level 1' high-frequency rules and 'Level 2' on-demand reference files, it ensures Claude receives high-signal tokens without context bloat. The system uses multi-entry indexing and specific trigger conditions to help Claude retrieve detailed SOPs, error diagnostics, and code patterns only when relevant to the current task, preventing the common issue of rule-following failure in long-running sessions.

主要功能

01Strict 'No Information Loss' migration protocols

02Multi-entry indexing based on errors, tasks, and file locations

03Multi-layered hierarchy (Level 1 Core vs. Level 2 References)

04Context-aware triggers for on-demand loading of detailed SOPs

05Automated 'Information Recording Principles' to prevent future file bloat

061 GitHub stars

使用场景

01Preventing context degradation in long development conversations by isolating low-frequency info

02Refactoring oversized CLAUDE.md files that have become difficult for the LLM to follow

03Standardizing project documentation for better AI agent navigation and consistency

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add luodashiv5/please_answer_this_era_5.0 claude-md-progressive-disclosurer

For use in Claude.ai and ChatGPT

主要功能

01Strict 'No Information Loss' migration protocols

02Multi-entry indexing based on errors, tasks, and file locations

03Multi-layered hierarchy (Level 1 Core vs. Level 2 References)

04Context-aware triggers for on-demand loading of detailed SOPs

05Automated 'Information Recording Principles' to prevent future file bloat

061 GitHub stars

使用场景

01Preventing context degradation in long development conversations by isolating low-frequency info

02Refactoring oversized CLAUDE.md files that have become difficult for the LLM to follow

03Standardizing project documentation for better AI agent navigation and consistency

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add luodashiv5/please_answer_this_era_5.0 claude-md-progressive-disclosurer

For use in Claude.ai and ChatGPT