caveman-compress

Compress markdown memory files into a shorter caveman format.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/junevm/avtoolz --skill caveman-compress-junevm
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: caveman-compress
Source: https://github.com/junevm/avtoolz/tree/main/.agents/skills/caveman-compress
Command: npx skills add https://github.com/junevm/avtoolz --skill caveman-compress-junevm

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

It reduces the token cost of loading large natural-language memory files (like CLAUDE.md and todo/preferences notes) by compressing them into a more concise “caveman” form.

Core Features & Use Cases

  • Token-saving memory compression: Compresses natural-language markdown into a shorter style while keeping meaning and structure.
  • Safe structure preservation: Preserves code blocks, inline code, URLs, file paths, headings, lists, tables, and YAML/frontmatter structure.
  • Validation with targeted repair: Validates the compressed output and, if validation fails, performs a limited fix rather than fully recompressing.
  • Backups for reversibility: Overwrites the target file with the compressed version and writes a human-editable backup as FILE.original.md.

Quick Start

Tell the skill to compress your memory file by issuing: /caveman-compress CLAUDE.md

Frequently Asked Questions about caveman-compress

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I reduce token usage when loading CLAUDE.md memory files in Claude?

To reduce token usage for CLAUDE.md memory files, compress the natural-language markdown into a shorter caveman format. This compression preserves technical content and structure while significantly shrinking the context size loaded on every session start.

Does markdown compression preserve code blocks, URLs, and file paths?

Markdown compression preserves code blocks, inline code, URLs, file paths, headings, lists, tables, and YAML frontmatter structure. It safely compresses natural-language text while keeping all technical references and formatting intact.

What is the best way to compress project memory markdown without losing technical content?

The best way to compress project memory markdown is using a caveman format compression that validates the output and performs targeted repairs if validation fails. This ensures technical content is retained while reducing overall token cost.

Can I restore original markdown files after context reduction compression?

You can restore original markdown files after context reduction because the compression process writes a human-editable backup as FILE.original.md before overwriting the target file with the compressed version.

What happens if validation fails during markdown compression?

If validation fails during markdown compression, the process performs a limited fix rather than fully recompressing the file. This targeted repair ensures the compressed output remains valid without losing the structure or content already processed.