caveman

Compress AI model responses into ultra-terse prose to minimize token usage.

Updated Jun 22, 2026
One-click install
npx skills add https://github.com/sqmasep/ecv-vinted --skill caveman-sqmasep
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: caveman
Source: https://github.com/sqmasep/ecv-vinted/tree/main/.claude/skills/caveman
Command: npx skills add https://github.com/sqmasep/ecv-vinted --skill caveman-sqmasep

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill addresses high token consumption by stripping away filler, pleasantries, and redundant prose while maintaining full technical accuracy and code integrity.

Core Features & Use Cases

  • Multi-Level Compression: Choose from six intensity levels ranging from professional-tight to extreme classical Chinese (wenyan) compression.
  • Technical Preservation: Ensures code blocks, error strings, API names, and technical logic remain untouched and verbatim.
  • Use Case: When working on long-running technical sessions, use this to reduce output length by 65-75% without losing the ability to debug complex code or system errors.

Quick Start

Activate the caveman skill by typing /caveman in the chat to immediately begin receiving ultra-compressed responses.

Frequently Asked Questions about caveman

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

FAQPage Schema
How do I reduce token usage in long-context technical debugging without losing code accuracy?

To reduce token usage during technical debugging, you can compress AI responses into ultra-terse prose. This strips away filler and redundant text while preserving code blocks, API names, and error strings verbatim, cutting output length by 65-75%.

What is the best way to compress AI model responses for token efficiency during code review?

The best way to compress AI model responses for token efficiency is applying intensity-based compression rules. This approach removes pleasantries and redundant prose across six intensity levels, ensuring technical logic and code integrity remain completely untouched.

Does ultra-compressed communication work for security-sensitive technical operations?

Ultra-compressed communication is not recommended for security-sensitive operations without precautions. The system features automatic clarity fallbacks specifically designed to prevent critical security details from being obscured during extreme text compression.

How do I activate ultra-terse prose compression for my documentation tasks?

To activate ultra-terse prose compression for documentation tasks, trigger the command interface in your chat. This immediately initiates the compression rules, allowing you to receive heavily truncated responses tailored for technical brevity.

Can I adjust the token compression intensity for different technical communication needs?

You can adjust token compression intensity across six distinct levels. These range from professional-tight compression for standard technical communication to extreme classical Chinese compression for maximum token reduction.

Why does my AI output still contain full code blocks when using text compression?

Your AI output retains full code blocks during text compression because technical preservation is strictly enforced. Code blocks, error strings, and API names remain untouched and verbatim to maintain debugging and system error resolution capabilities.