why use many token when few do trick
A universal LLM communication skill that makes AI agents talk like caveman — cutting ~75% of output tokens while keeping full technical accuracy.
Works with any agent that supports SKILL.md format: Qwen Code, Claude Code, Cursor, Windsurf, GitHub Copilot, Cline, Codex, and 40+ more.
|
|
|
|
Same answer. 75% less word. Brain still big.
- Faster responses — less token to generate = speed go brrr
- Easier to read — no wall of text, just the answer
- Same accuracy — all technical info kept, only fluff removed
- Save money — ~75% less output token = less cost
- Fun — every debug session become comedy
| Level | Example | Best for |
|---|---|---|
| 🪶 Lite | "Component re-renders because you create a new object reference. Wrap it in useMemo." |
Professional but tight |
| 🪨 Full | "New object ref each render. Wrap in useMemo." |
Default caveman |
| 🔥 Ultra | "Inline obj prop → new ref → re-render. useMemo." |
Telegraphic mode |
| 📜 文言文 | "物出新參照,致重繪。useMemo Wrap之。" | Classical Chinese compression |
Trigger with any of these:
/caveman- "caveman mode"
- "talk like caveman"
- "less tokens"
- "be brief"
Switch levels:
/caveman lite/caveman full(default)/caveman ultra/caveman wenyan-lite/caveman wenyan-full/caveman wenyan-ultra
Stop with: "stop caveman" or "normal mode"
Level sticks until you change it or session ends.
npx caveman-skill installAuto-detects your agent and installs to the right path. Or specify directly:
npx caveman-skill install --agent qwen-code
npx caveman-skill install --agent claude-code
npx caveman-skill install --agent cursorgit clone https://github.com/yuritoledo/caveman-skill.gitThen point your agent to the SKILL.md file.
Grab the SKILL.md file and place it in:
| Agent | Path |
|---|---|
| Qwen Code | .qwen/skills/caveman/ |
| Claude Code | .claude/skills/ |
| Cursor | .cursor/skills/ |
| Windsurf | .windsurf/skills/ |
| Cline | .cline/skills/ |
- Drops articles (a/an/the), filler (just/really/basically), pleasantries (sure/happy to), hedging
- Keeps technical terms, code blocks, URLs, commands, file paths — anything with real info passes through untouched
- Compresses prose into
[thing] [action] [reason]. [next step].pattern - Auto-clarity — temporarily drops caveman for security warnings, irreversible confirmations, or when user is confused
Code, commits, and PR reviews are always written normally. Technical accuracy is never compromised.
Real token counts from LLM APIs:
| Task | Normal | Caveman | Saved |
|---|---|---|---|
| Explain React re-render | 1180 | 159 | 87% |
| Fix auth middleware | 704 | 121 | 83% |
| Set up DB connection pool | 2347 | 380 | 84% |
| Explain git rebase vs merge | 702 | 292 | 58% |
| Docker multi-stage build | 1042 | 290 | 72% |
| Average | 1214 | 294 | 65% |
Range: 22%–87% savings across prompts. Simple answers save less, complex ones save more.
npx caveman-skill installInteractively detects your agent and installs the skill to the right path.
Or specify your agent:
npx caveman-skill install --agent qwen-code
npx caveman-skill install --agent claude-code
npx caveman-skill install --agent cursor
npx caveman-skill install --agent windsurf
npx caveman-skill install --agent clineMIT — free like mass mammoth on open plain.
Based on JuliusBrussee/caveman — the original. Forked with permission and love.