ultracave
An agent skill by JuliusBrussee, from JuliusBrussee/caveman. Tags: cli, copy-editing, local-workflow, writing.
What it does
Caveman at maximum compression: fragments, one word when one word is enough, each fact once. Invoke only with /ultracave or /caveman ultra. Stays on until "stop caveman" or "normal mode".
Install
With the skills CLI, which installs into Claude Code, Codex, Cursor and other agents:
npx skills add JuliusBrussee/caveman --skill ultracaveOr copy the skill folder into Claude Code's skills directory by hand (~/.claude/skills for every project, or .claude/skills inside one):
git clone --depth 1 https://github.com/JuliusBrussee/caveman
cp -r caveman/plugins/caveman/skills/ultracave ~/.claude/skills/ultracaveSafety box score
Grade A, minimal risk: 5 of 100 risk points, rated 2026-10-17. The grade rates what the skill and its scripts can reach on the machine of whoever installs it, across eight categories from shell execution to secrets access. How the score works.
This skill only changes the assistant’s speaking style and session behavior; it does not ship scripts, make network calls, or touch files. Main risk is prompt-level control: it persists for the whole chat until a stop phrase and constrains how the agent responds.
| Category | Level | Why |
|---|---|---|
| Code & shell execution | 0 (none) | No shell commands, code execution, or runtime scripts in `SKILL.md`. |
| Network access | 0 (none) | No downloads, API calls, or remote transmissions are defined. |
| Filesystem reach | 0 (none) | No file reads, writes, or deletes are requested beyond the chat session. |
| Secrets & credentials | 0 (none) | No credential access, secret handling, or requests to reveal tokens/passwords appear. |
| Privilege & persistence | 0 (none) | No sudo, global installs, services, or OS/git setting changes are present. |
| Instruction hijack surface | 1 (low) | It explicitly changes agent behavior across the session: `Invoke only with /ultracave`, `Every response, whole session`, and `The user's language` are prompt-level control rules. |
| Irreversible actions | 0 (none) | It does not send messages, deploy, pay, book, delete, or otherwise act on the user's behalf. |
| Transparency | 0 (none) | The skill is plain markdown; no binaries, encoded payloads, or hidden runtime content are present. |
Before installing
- Read `SKILL.md` to confirm you want a session-wide terse-response mode.
- Try it in a throwaway chat first; it persists until `stop caveman` or `normal mode`.
- Check whether any downstream workflows depend on full-form explanations, because this skill strips them.
- Verify the repo contains only this markdown skill and no hidden scripts or installers.
Rated 2026-10-17 by an automated review of 1 file from the skill's repository.
Source
- Repository
- JuliusBrussee/caveman (all skills from this repository)
- Path
- plugins/caveman/skills/ultracave/SKILL.md
- Branch
- main
- Collection
- caveman
- Updated
- 2026-10-05
Related skills
- caveman — Terse caveman voice: answer first, fluff gone, every technical fact kept. Use for /caveman, "caveman mode", "talk like caveman", "be brief", "less tokens".
- megacave — Caveman in Classical Chinese: 文言文 register, far fewer characters, technical terms verbatim. Invoke only with /megacave or /caveman wenyan.
- caveman-stats — Show recorded output and cache-read token usage and mode attribution for the current Claude Code session, or locate the host's native usage report.
- agents-get-started — Use when a developer wants to create a new agent project or get started with AgentCore.
- copy-editing — When the user wants to edit, review, or improve existing marketing copy, or refresh outdated content.
- google-docs — Author and edit Google Docs with the google-docs tools - structured writing, character and paragraph styling, lists, tables, images, headers/footers.