caveman-stats

Name: caveman-stats
Availability: InStock
Author: Mosaic-agent

Report real token usage and estimated savings from Claude Code session logs.

36|7|Updated Feb 22, 2026

Search Tags:#claude-code #token-usage #session-logs #ai-logging #model-efficiency #caveman-stats #hook-tracker

Ready to run

npx skills add https://github.com/Mosaic-agent/Mosaic-fund-agent --skill caveman-stats-mosaic-agent

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: caveman-stats
Source: https://github.com/Mosaic-agent/Mosaic-fund-agent/tree/main/.agents/skills/caveman-stats
Command: npx skills add https://github.com/Mosaic-agent/Mosaic-fund-agent --skill caveman-stats-mosaic-agent

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Show real token usage and estimated savings for the current session.

Core Features & Use Cases

Reads the Claude Code session log to report actual input/output tokens and lifetime savings.
Output is injected by the caveman-mode-tracker hook on the /caveman-stats command.
Displays results without the model performing token calculations.

Quick Start

Run /caveman-stats to see the current session token usage and savings.

Frequently Asked Questions about caveman-stats

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I track real token usage and savings in Claude Code session logs?▼

Token usage monitoring works by reading the Claude Code JSONL session log to report actual input and output tokens, along with estimated lifetime savings, without requiring the model to perform token calculations.

How do I check my Claude Code session token economics during interactive coding?▼

Checking per-session token economics requires running the /caveman-stats command during interactive coding sessions, which injects actual token usage and estimated savings data directly into the output via the caveman-mode-tracker hook.

Does the caveman-stats hook tracker require the model to calculate token usage?▼

No, the caveman-stats hook tracker does not require the model to calculate token usage. It reads the JSONL session log directly and injects the results externally, ensuring the model does not perform token calculations.

What is the best way to evaluate model efficiency using Claude Code logs?▼

Evaluating model efficiency is done by monitoring per-session token economics with the caveman-stats hook, which reads the JSONL session log to display real input/output tokens and estimated lifetime savings for interactive coding sessions.

Can I monitor lifetime savings from Claude Code sessions without manual token calculations?▼

Yes, you can monitor lifetime savings without manual calculations by running the /caveman-stats command, which reads the JSONL session log and outputs estimated savings automatically through the caveman-mode-tracker hook.