skills.rest
svssdeva
svssdevaCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

caveman-compress

Compress natural-language memory markdown files while preserving code, URLs, and structure.

1|1|Updated Apr 20, 2026
Search Tags:#claude code#token optimization#guardrails#memory files#markdown compression#prose shortening
One-click install
npx skills add https://github.com/svssdeva/agentic-skills --skill caveman-compress-svssdeva
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: caveman-compress
Source: https://github.com/svssdeva/agentic-skills/tree/main/utils/caveman-compress
Command: npx skills add https://github.com/svssdeva/agentic-skills --skill caveman-compress-svssdeva

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Caveman Compress reduces input token usage by shrinking natural-language memory files while keeping all technical substance, code, URLs, and structure intact.

Core Features & Use Cases

  • Token-saving compression: Converts prose in memory files (like CLAUDE.md, todos, preferences) into a shorter “caveman” style.
  • Safety-preserving preservation: Preserves code blocks, inline code, URLs, links, file paths, commands, and technical terms exactly.
  • Operational guardrails: Writes a human-readable backup as FILE.original.md and avoids compressing the backup itself to prevent compounding changes.

Quick Start

Run caveman-compress on a memory file by asking an AI to compress the file at the path you provide.

Frequently Asked Questions about caveman-compress

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

FAQPage Schema
How do I reduce Claude input tokens in my markdown memory files?▼

To reduce Claude input tokens in markdown memory files, you compress natural-language prose into a shorter style. This process shrinks CLAUDE.md-style instructions and preferences while preserving technical content and exact markdown structure.

What is the best way to shorten markdown prose without losing code blocks and URLs?▼

The best way to shorten markdown prose without losing code blocks and URLs is to apply compression guardrails. These guardrails enforce exact preservation of inline code, file paths, commands, and links while converting surrounding text to a condensed format.

Does compressing CLAUDE.md files permanently overwrite my original instructions?▼

Compressing CLAUDE.md files does not permanently overwrite original instructions without safety measures. The process writes a human-readable backup as FILE.original.md and only overwrites the original file after validated compression with bounded retries.

Can I automate token optimization for markdown todos and preferences?▼

You can automate token optimization for markdown todos and preferences by running a script that applies prose shortening. This targets natural-language memory files, converting verbose text into a compressed format to cut down Claude input token usage.

Why does markdown compression sometimes break file paths and commands?▼

Markdown compression breaks file paths and commands when applying unguarded prose shortening. Using a compression method with strict preservation guardrails ensures technical syntax, commands, and file paths remain exactly intact during token reduction.

Related Skills

Explore More Content & Communication Skills→
RDCrystalLabRDCrystalLab

compress

Compress markdown memory files to reduce token usage while preserving technical content.

Official
Advanced
valorisavalorisa
4

token-optimization

Implement cache management and context optimization rules for Claude Code sessions.

Community
Advanced
selfagencyselfagency

caveman-compress

Compress markdown memory files into shorter phrasing while preserving code blocks and URLs.

Community
Advanced
SaltFishGCSaltFishGC

caveman-compress

Compress prose-heavy memory files into compact caveman format with backups.

Community
Advanced