Compress AI responses into terse caveman mode with selectable intensity levels.
npx skills add https://github.com/irchamaji/agents --skill caveman-irchamaji
Please help me install this Agent Skill. Skill: caveman Source: https://github.com/irchamaji/agents/tree/main/skills/caveman Command: npx skills add https://github.com/irchamaji/agents --skill caveman-irchamaji
Reduces AI response token usage while preserving technical accuracy by enforcing terse, “caveman mode” communication patterns.
Ask your AI to respond in caveman mode, for example: “Use caveman mode full and answer my question about why my function re-renders.”
High-intent search queries and answers about installing and using this skill.
Caveman mode compresses AI responses by dropping filler, hedging, and articles while keeping technical terms exact. This token efficiency technique maintains full technical accuracy during code troubleshooting and concept explanations.
Concise AI prompt instructions offer lite, full, ultra, and wenyan intensity levels. Full is the default compression mode, while ultra and wenyan variants apply stronger formatting rules for greater token reduction.
Avoid compressed or caveman mode for security warnings, irreversible confirmations, and multi-step sequences where order can be misread. These safety-aware exceptions ensure critical clarity is maintained when technical accuracy alone is insufficient.
Token compression works for code troubleshooting and technical concept explanations by stripping unnecessary grammatical structures. It enforces terse communication patterns that reduce response size without altering the underlying technical meaning.
Force concise answers by explicitly requesting caveman mode in your prompt, such as asking the AI to use full or ultra intensity. This switches the AI to a compression mode that actively drops filler and enforces terse formatting.
xaionaroCondenses technical explanations into concise, caveman-style text with adjustable intensity levels.
jchildreeCompress technical content into ultra-brief messages with six intensity levels.
VuNguyenVietTienAnalyze, transcribe, and generate images, audio, and video content.
phucle297Compress text output into concise caveman-style responses across multiple turns.
