Back to skills

Agent Skill

Caveman Stats

caveman-stats

Show real token usage and estimated savings for the current session. Reads directly from the Claude Code session log — no AI estimation. Triggers on /caveman-stats. Output is injected by the mode-tracker hook; the model itself does not compute the numbers.

JuliusbrusseeToolsJavaScriptAiAnthropicCavemanClaudeClaude-codeLlmMemePrompt-engineeringSkillTokens

128K installs

juliusbrussee/caveman

by Juliusbrussee

Score

8.9

/ 10

Installs

128K

Repo Stars

75.7K

Last Updated

9d ago

Fresh

Quality Ratio

99%

Description

Verified

Language

JavaScript

First Published

Apr 2026

Summary

The Caveman Stats agent skill provides developers with real-time token usage and estimated cost savings for their current session, directly from the Claude Code session log without AI estimation. This agent skill is particularly valuable for developers and teams using AI agents powered by Claude models who need clear visibility into their session's token consumption and potential cost efficiencies. This is a skill with 2K installs, indicating a notable level of adoption among the agent skill registry. To use it, a developer simply triggers the `/caveman-stats` command within their session. The skill operates via a backend hook, which directly reads session logs and injects the token usage data and savings estimates into the output, meaning the AI model itself does not process the request or compute these statistics, ensuring immediate results without additional token expenditure.

Skill Definition

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.

How to Use

Use in O-mega

Claude Code

npx skills add juliusbrussee/caveman caveman-stats
Caveman Stats | Agent Skills | o-mega