Skip to content
Get packs

Ultracave Terse Responses

Makes the agent reply in ultra-compressed caveman-style fragments until you turn it off

Source Download skill
Repository
JuliusBrussee/caveman
Stars
116,835
Licence
Apache-2.0
Files
1 file

What it does

Once installed and switched on, the agent answers in very short, clipped fragments, dropping filler words and repeated facts while keeping technical content, code, commands, numbers and error messages exactly as written. It stays in this mode for the whole session until you say "stop caveman" or "normal mode". It is a style setting for the agent's replies, not a new tool that does work on its own.

Great for

  • Developers who want quick, dense answers and find long explanations slow to read
  • People who already know the subject and only need the key facts
  • Users who want to cut token use and reading time during long coding sessions

Skip it if

  • Beginners or readers who need full explanations, since terse fragments can hide reasoning and be easy to misread
  • Formal or customer-facing writing, because the output is deliberately informal and clipped
  • Anyone expecting it to run code or change files by itself; it only shapes how the agent writes

Try asking

  • “/ultracave then ask: why does my React component re-render on every click?”
  • “/caveman ultra, then review this function for bugs”
  • “/ultracave, and keep it on while we debug the database migration”
  • “Turn it off by saying "normal mode" after the session is finished”
Step-by-step guide: how to use Ultracave Terse ResponsesInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the Apache-2.0 licence and an attribution note.

    ultracave.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip ultracave.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip ultracave.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Ask for a full sentence when a fragment is unclear; the skill tells the agent to do this for risky or ambiguous points
  • Use /caveman status to check the current mode, but note it only reports accurately if the host tracks it
  • Remember that security warnings and irreversible-action confirmations are still written in plain prose

What's inside

  • SKILL.md

Redistributed unmodified from JuliusBrussee/caveman under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.