Compress AI communication by removing filler while preserving technical terms.
npx skills add https://github.com/koseki2580/skills --skill caveman-koseki2580
Please help me install this Agent Skill. Skill: caveman Source: https://github.com/koseki2580/skills/tree/main/skills/caveman Command: npx skills add https://github.com/koseki2580/skills --skill caveman-koseki2580
This Skill reduces token usage and verbosity when you want brief, technically accurate answers without conversational fluff.
Ask for caveman mode, for example: "Explain this bug in caveman mode and include only the likely cause and exact fix."
High-intent search queries and answers about installing and using this skill.
You can reduce token usage by applying a compression mode that removes filler and pleasantries while preserving technical terms and code blocks. This yields deterministic, brief answers without conversational fluff.
Caveman mode is a prompting style that compresses AI communication to minimize token usage. It applies rule-based removal of articles and hedging, using terse phrasing like [thing] [action] [reason] to maintain technical accuracy.
To get brief answers without losing accuracy, request a compression mode that drops conversational fluff but strictly preserves technical terms and code blocks. This applies to troubleshooting and implementation guidance scenarios.
Compressed AI communication temporarily disables brevity for security warnings, irreversible confirmations, and multi-step sequences. This exception restores full clarity to prevent misunderstandings during critical operations.
You should not use ultra-compressed AI communication for multi-step sequences where ordering could be misunderstood, or during security warnings and irreversible actions. The system restores clarity for these critical scenarios.
ahostbrCompress technical responses by removing filler and hedging while preserving code and warnings.
tyevansExpand sparse findings into detailed descriptions with cited evidence.
heyramziPlan feature UX and UI through a phased discovery interview workflow.
babuallyProduce terse, high-density responses that preserve technical meaning.
