compress

Compress natural language memory files into caveman format with backup.

171|10|Updated Feb 20, 2026
One-click install
npx skills add https://github.com/fmflurry/settings-opencode --skill compress-fmflurry
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: compress
Source: https://github.com/fmflurry/settings-opencode/tree/main/skills/compress
Command: npx skills add https://github.com/fmflurry/settings-opencode --skill compress-fmflurry

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Compress natural language memory files into caveman format to drastically reduce per-session tokens, while preserving content semantics. The compressed file overwrites the original and a human-readable backup is saved as FILE.original.md.

Core Features & Use Cases

  • Overwrites the original with a compact version while retaining essential content and structure.
  • Creates a human-readable backup named FILE.original.md for reference and restoration.
  • Applies across memory files like CLAUDE.md, todos, and preferences to reduce token usage during Claude sessions.

Quick Start

cd into this skill's directory and run python3 -m scripts <absolute_filepath> to compress a memory file.

Frequently Asked Questions about compress

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

FAQPage Schema
How do I reduce token usage in my Claude memory files?

You can reduce token usage by compressing natural language memory files into a compact caveman format. This overwrites the original with a token-optimized version while preserving content semantics.

What is the caveman format for markdown compression?

Caveman format is a compressed representation of natural language that drastically reduces input tokens during Claude sessions. It retains essential content and structure while stripping verbose syntax from markdown files.

Can I compress a CLAUDE.md file without losing the original content?

Yes, compressing a CLAUDE.md file saves a human-readable backup named FILE.original.md before overwriting the original. This ensures you can reference or restore the uncompressed version at any time.

How do I automate markdown compression for project todos and preferences?

Automate markdown compression by running the provided Python script with an absolute filepath argument. The script applies compression, performs validation, and retries corrections up to two cycles automatically.

Does the compression script validate the compressed memory files?

Yes, the compression script performs validation on the compressed memory files and runs up to two retry cycles for corrections. This ensures the compressed output maintains semantic accuracy before finalizing.

What types of memory files can I compress for Claude sessions?

You can compress natural language memory files like CLAUDE.md, project todos, and user preferences. The tool applies across projects and sessions to consistently reduce per-session token load.