🪨 Why use many token when few token do trick?
“Caveman” is a skill for Claude Code that makes the model respond like a caveman, cutting ~75% of tokens while maintaining 100% technical accuracy.
📊 Before vs. After:
| Normal Claude (69 tokens) | Caveman Claude (19 tokens) |
|---|---|
| “The reason your React component is re-rendering is likely because you’re creating a new object reference…” | “New object ref each render. Inline object prop = new ref = re-render. Wrap in useMemo.” |
Same fix. 75% less word. Brain still big.
🔧 Install:
npx skills add JuliusBrussee/cavemanActivate with: /caveman, “talk like caveman”, “less tokens please”
🧠 Caveman not dumb. Caveman efficient:
- ✅ Code blocks: normal (caveman not stupid)
- ✅ Technical terms: exact (polymorphism stays polymorphism)
- 💀 Eliminated: “Sure, I’d be happy to help”, hedging, unnecessary filler
💰 Result: 75% less tokens = 75% less cost + responses ~3x faster
💡 Explanation in a nutshell#
AI models like Claude tend to use lots of polite filler words before giving you the actual answer. This skill teaches them to get straight to the point, like a caveman. The result: same technical information, much cheaper and faster.
More information at the link 👇
Also published on LinkedIn.
