Reduce token usage with terse, high-signal responses across intensity levels.
npx skills add https://github.com/Jannael/Glinter --skill caveman-jannael
Please help me install this Agent Skill. Skill: caveman Source: https://github.com/Jannael/Glinter/tree/main/.agents/skills/caveman Command: npx skills add https://github.com/Jannael/Glinter --skill caveman-jannael
Ultra-compact communication mode that reduces token usage while preserving technical accuracy.
Ask for caveman mode by saying caveman mode or using /caveman lite, and receive concise, precise responses.
High-intent search queries and answers about installing and using this skill.
Ultra-compact communication mode reduces token usage by delivering terse, high-signal responses that strip non-essential fluff. This mechanism preserves technical accuracy while significantly lowering token consumption across your AI interactions.
Terse AI responses support intensity levels including lite, full, ultra, wenyan-lite, wenyan-full, and wenyan-ultra. These levels provide varied verbosity and style options to control the exact compactness of the generated output.
To activate caveman mode for concise prompt instructions, use specific trigger phrases like caveman mode or /caveman. The system also auto-triggers when token efficiency is requested, automatically applying your default intensity settings.
Caveman mode maintains persistence across interactions once activated. It enforces your selected intensity level and default settings throughout the session, ensuring consistent terse, precise responses until you change the preference.
You should avoid ultra-compact communication mode when your task requires a conversational tone, detailed explanations, or high-level context. The terse, no-fluff design prioritizes token efficiency over conversational nuance.
bbohhu-progEnforce pre-response skill discovery and tool invocation across conversations.
AveerCompress message ranges and prune verbose outputs to manage session context.
se02035Compress communication to reduce token usage while maintaining technical accuracy.
Parallel-7Generates README.md files with specified formatting and migrates legacy ones to match it.
