Caveman Explainer

Caveman Explainer rewrites any concept in blunt, tiny-word English — the kind of language a caveman might grunt out around a fire. It exists for two reasons: to produce intros and marketing copy that a distracted human actually reads, and to catch the moment you realize you can't explain your own product without hiding behind jargon. If you can't say it caveman-style, you probably don't understand it yet.

When to use it

• Writing a landing-page hero line or product intro that keeps bouncing off readers.
• Drafting the "what is this" section of a README, onboarding doc, or pitch deck.
• Sanity-checking a feature before you ship: if the caveman version sounds empty, the feature might be.
• Turning a dense technical paragraph (API docs, RFC, architecture note) into something a non-engineer stakeholder gets in one pass.
• Killing buzzwords. Feed it copy stuffed with "synergize," "leverage," "seamless," and watch them disappear.

What it does

The skill applies a rewrite pass with hard constraints: words stay short (aim for one or two syllables, never more than three-to-four), sentences stay flat, and abstract nouns get swapped for concrete ones. "Authenticate the user" becomes "make sure it is you." "Distributed cache" becomes "many fast boxes that hold copies."

It works on whatever you hand it — paste raw text, point it at a file, or describe a concept and ask for the caveman version. You can ask for a single grunt-line, a short paragraph, or a side-by-side (normal vs. caveman) so you can compare and pick.

Usage example

curl -fsSL --create-dirs -o ~/.claude/skills/caveman-explainer/SKILL.md https://claudeskil.com/skills/caveman-explainer/SKILL.md

# in a session
> Caveman-explain our product: "An event-driven serverless
  pipeline that ingests webhooks and fans them out to
  downstream consumers."

Caveman:
"Thing happen out there. We hear it. We tell all
 the others who need to know. Fast. No big rock servers."

> Now give me a side-by-side of the README intro in intro.md

Tips & gotchas

• It is a clarity tool, not a style you ship blindly. Use the caveman draft to find the plain idea, then write your final copy at a normal register if the audience expects it.
• For numbers, units, and proper nouns, ask it to keep them — caveman mode can otherwise round "300ms" down to "fast," which loses real meaning.
• It exposes vagueness ruthlessly. If the output sounds like nothing, that's a signal about the source, not a bug.
• Long input drifts back toward big words. Feed it in small chunks (a paragraph at a time) for the strictest caveman discipline.

FAQ

Does it change facts? No — it simplifies wording, not claims. Always re-check numbers and named features in the output.

Can it do other voices? It's tuned for caveman-grade plainness. For "explain like I'm 5" or pirate-speak, use a general rewrite prompt instead.

Will it work on code comments? Yes, for the prose parts. It won't rewrite code itself, just the human-readable explanation around it.