Agent Skill
Ultracave skill
ultracaveJuliusBrussee/cavemanplugins/caveman/skills/ultracave/SKILL.md
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".
always on
54 tokname + description
once used
536 tokSKILL.md body
bundled files
0
repo stars
114kGitHub
Install the Ultracave skill
Skills CLI (Claude Code, Codex, Cursor and others)
npx skills add https://github.com/JuliusBrussee/caveman --skill ultracave
Claude Code, by hand (all projects)
git clone --depth 1 https://github.com/JuliusBrussee/caveman.git /tmp/caveman mkdir -p ~/.claude/skills cp -R /tmp/caveman/plugins/caveman/skills/ultracave ~/.claude/skills/ultracave
For one project only, copy the folder into .claude/skills/ in that project instead. It is marked disable-model-invocation, so the agent won't pick it on its own: you run it as /ultracave.
What it costs your context
Agent Skills load in three steps. Tokens counted with the o200k tokenizer; your model's count will differ a little.
| When | What loads | Cost |
|---|---|---|
| Every session | Name and description | 54 tok |
| When the agent uses it | The SKILL.md instructions | 536 tok |
| Only if opened | No bundled files | — |
Until it is used, ultracave costs about 25× less than the median MCP server we measured (1,362 tokens on every turn). Installing all 28 skills from JuliusBrussee/caveman would load 1,303 tokens of descriptions. Skills vs MCP study
Checks
Our checks found nothing to flag: it follows the Agent Skills specification's naming and size guidelines and doesn't pipe downloads into a shell.
Checked 2026-10-22 · method
What's inside
The headings of its SKILL.md. Read the full file on GitHub.
- ultracave
- Persistence
- Floor
- Rules
- 1. Fragments
- 2. One word when one word is enough
- 3. Each fact once
- 4. Cut conjunctions only when order survives
- 5. Tool runs
- 6. Never perform
- When to break the rules
- Pre-send check
More skills from JuliusBrussee/caveman
- cavemanJuliusBrussee/caveman
Terse caveman voice: answer first, fluff gone, every technical fact kept. Use for /caveman, "caveman mode", "talk like caveman", "be brief", "less tokens". Stays on until "stop caveman" or "normal mode".
586k62 tok idle · 1,086 in use - caveman-commitJuliusBrussee/caveman
Write a Conventional Commits message compressed to intent only. Use for "write a commit", "commit message", /commit or /caveman-commit.
411k38 tok idle · 564 in use - caveman-reviewJuliusBrussee/caveman
Compressed code review - one line per finding with location, problem and fix. Use for /caveman-review, "review this PR", or "review the diff".
407k40 tok idle · 585 in use - caveman-compressJuliusBrussee/caveman
Compress a memory file such as CLAUDE.md or a todo list into caveman format to save input tokens, keeping a readable backup. Sends the file's text to the configured model provider. Trigger: /caveman-compress.
407k754 tok idle · 1,247 in use - caveman-helpJuliusBrussee/caveman
Quick-reference card for the three caveman skills and their commands. Trigger: /caveman-help or "caveman help".
398k32 tok idle · 667 in use - cavecrewJuliusBrussee/caveman
When to delegate to `cavecrew-investigator` (locate code), `cavecrew-builder` (1-2 file edit) or `cavecrew-reviewer` (diff review) instead of working inline or using `Explore`. Their output is compressed, so main context lasts longer.
354k65 tok idle · 869 in use - caveman-statsJuliusBrussee/caveman
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. Trigger: /caveman-stats.
353k42 tok idle · 331 in use - caveman-statsJuliusBrussee/caveman
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. Trigger: /caveman-stats.
353k42 tok idle · 352 in use