caveman

Compresses AI response output into terse technical prose across six intensity levels.

Updated Sep 10, 2026
One-click install
npx skills add https://github.com/longqt321/Rendez --skill caveman-longqt321
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: caveman
Source: https://github.com/longqt321/Rendez/tree/main/.agent/skills/caveman
Command: npx skills add https://github.com/longqt321/Rendez --skill caveman-longqt321

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Long AI responses waste tokens and reading time when you only need the technical answer. This Skill strips filler, hedging, and pleasantries from every reply while preserving technical accuracy, code, and error strings. ## Core Features & Use Cases - Six compression levels: lite, full, ultra, plus wenyan-lite, wenyan-full, and wenyan-ultra for classical Chinese terseness. - Session persistence: the chosen style applies to every response until you say "stop caveman" or "normal mode". - Auto-clarity fallback: security warnings, irreversible actions, and ambiguous multi-step sequences revert to clear prose automatically. - Use Case: During a long debugging session, enable full mode so every diagnosis arrives as short fragments like "Bug in auth middleware. Token expiry check use < not <=." with code blocks unchanged. ## Quick Start Ask the assistant to enable caveman mode at full level so all further answers in this session are compressed.

Frequently Asked Questions about caveman

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

FAQPage Schema
How do I make AI responses shorter and more concise?

Enable caveman mode with /caveman or by asking for brief answers. The full level drops articles and filler while keeping technical terms, code, and error messages exact, cutting output tokens without losing substance.

What compression levels does caveman mode support?

Six levels: lite, full, ultra, and three classical Chinese variants wenyan-lite, wenyan-full, wenyan-ultra. Lite keeps full sentences, ultra strips conjunctions, and wenyan modes compress into classical Chinese with up to 90% character reduction.

Does caveman mode work in languages other than English?

Yes, it preserves the user's dominant language exactly and only compresses style. For languages with grammatical particles or case markers, those are kept while politeness and filler are removed.

When does caveman mode stop compressing responses?

It automatically reverts to clear prose for security warnings, irreversible action confirmations, and multi-step sequences where fragments could cause misreading. It also stops when you say "stop caveman" or "normal mode".

Does caveman mode change code blocks or error messages?

No. Code blocks, API names, CLI commands, commit keywords, and exact error strings stay verbatim at every level. Only surrounding prose is compressed, and persisted artifacts like commits and docs are written in normal prose.