caveman-stats

SkillMonitoring & ops

caveman-stats is a skill for Claude Code that cuts token usage by about 65% by rewriting responses in short, simple language. Once added, your agent gets its point across with far fewer tokens on every reply.

Available today. Use it from your connected AI after setup.

Add the skill, then ask your agent something and it will reply in short, simple language using fewer tokens.

Then ask your AI: use the caveman-stats skill

What your AI can do with it

  • Rewrite responses in short, simple language
  • Cut token usage by about 65%
  • Shorten replies before they are sent
  • Swap wordy phrasing for plain words

What this skill tells your AI

The instructions your AI receives, as published by juliusbrussee/caveman in skills/caveman-stats/SKILL.md and read by ahel’s review.

This skill is delivered by hooks/caveman-stats.js (read by hooks/caveman-mode-tracker.js on /caveman-stats). The model does not need to do anything when this skill fires — the hook returns decision: "block" with the formatted stats as the reason. The user sees the numbers immediately.

Output also includes Est. rule overhead and Est. net lines wherever a savings estimate exists with a known turn count. Rule overhead is the estimated per-turn INPUT-token cost of the injected caveman rules (default 1,250 tokens/turn, override with CAVEMAN_RULE_OVERHEAD_TOKENS) times the turn count. Net is savings minus that overhead — when negative, the output says so plainly and suggests turning caveman off for that workload, rather than hiding the net-negative regime behind a gross-savings number (see docs/HONEST-NUMBERS.md).

Signals

GitHub stars
105k
Forks
6k
Last commit
Sep 2026
Installs
297k installs
Advanced
Catalog kind
skill
Gateway key
caveman-stats
Source
github.com/juliusbrussee/caveman