Optimize CLAUDE.md files using progressive disclosure. Goal: Maximize information efficiency, readability, and maintainability. Use when: User wants to optimize CLAUDE.md, information is duplicated across files, or LLM repeatedly fails to follow rules.
Two-level CLAUDE.md architecture
Backup created
0%
0%
Two-level structure present
100%
100%
Reference naming convention
25%
37%
Information recording principles
10%
100%
Reference index at document start
12%
100%
Reference index at document end
37%
100%
Before-modifying-code table
0%
100%
Trigger conditions on references
80%
100%
Code patterns retained in Level 1
100%
100%
SOP content moved to Level 2
100%
100%
Historical decisions moved to Level 2
100%
100%
Verification report present
100%
100%
No line count metrics
100%
100%
Content migration integrity verification
Connection pool parameters identified
100%
100%
Error diagnostic compression detected
100%
100%
Missing reference file identified
100%
100%
Historical decisions missing
100%
100%
Deployment SOP content loss detected
100%
100%
Level 1 code restored verbatim
100%
100%
Deployment SOP restored verbatim
100%
100%
Missing reference file created
100%
100%
Historical decisions file created
100%
100%
No line count metrics used
100%
100%
Quality-based optimization metrics
No line count language in strategy
100%
100%
No line count language in CLAUDE.md
100%
100%
Quality framework articulated
100%
100%
Success signals defined by behavior
100%
90%
Code patterns kept in Level 1
100%
100%
SOP moved to Level 2
100%
100%
Historical context moved to Level 2
100%
100%
Section-by-section decision justification
100%
100%
Future information guidance present
100%
100%
Iron rules retained in Level 1
100%
100%
392d34c
Table of Contents
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.