Caveman Compress

If you're working with large CLAUDE.md files or project memory that's eating your token budget, this compresses them into caveman speak while keeping all the technical bits intact. It strips articles, filler words, and pleasantries but leaves code blocks, URLs, commands, and file paths completely untouched. The original gets backed up as .original.md before being overwritten. It's a pretty aggressive approach to token savings, and the compressed output reads like telegram shorthand, but if you're repeatedly hitting context limits on documentation heavy projects, shaving 30 to 40 percent off your memory files without losing substance is worth the tradeoff.

GitHub stars: 504
last updated: Jun 11, 2026

View on GitHub

Install

$ npx -y skills add hoangnguyen0403/agent-skills-standard --skill caveman-compress --agent claude-code

Installs into .claude/skills of the current project.

Files

Caveman Compress

Purpose

Compress natural language files (CLAUDE.md, todos, preferences) into caveman-speak to reduce input tokens. Compressed version overwrites original. Human-readable backup saved as <filename>.original.md.

Trigger

/caveman:compress <filepath> or when user asks to compress a memory file.

Process

The compression scripts live in caveman-compress/scripts/ (adjacent to this SKILL.md). If the path is not immediately available, search for caveman-compress/scripts/__main__.py.
Run:

cd caveman-compress && python3 -m scripts <absolute_filepath>

The CLI will:

detect file type (no tokens)
call Claude to compress
validate output (no tokens)
if errors: cherry-pick fix with Claude (targeted fixes only, no recompression)
retry up to 2 times
if still failing after 2 retries: report error to user, leave original file untouched