caveman-stats

Analyze Claude Code session logs for actual token usage and savings.

Updated May 17, 2026

Search Tags:#Claude Code #token usage #session analysis #stats #token saving

Ready to run

npx skills add https://github.com/galeep/plugin-place --skill caveman-stats-galeep

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: caveman-stats
Source: https://github.com/galeep/plugin-place/tree/main/plugins/caveman/skills/caveman-stats
Command: npx skills add https://github.com/galeep/plugin-place --skill caveman-stats-galeep

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

caveman-stats helps you analyze the real token usage and savings for Claude Code sessions, giving you accurate statistics on input/output tokens without AI estimation.

Core Features & Use Cases

Session Log Analysis: Reads the Claude Code session log directly for actual input/output token usage.
Estimation: Provides estimated savings compared to a non-caveman baseline.
Usage Trigger: Triggered with the /caveman-stats command.
Output Display: Shows the session count, input/output tokens, baseline usage, and savings percentage.

Quick Start

Run /caveman-stats in your Claude Code session to view detailed stats.

Frequently Asked Questions about caveman-stats

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

FAQPage Schema

How do I analyze token usage in a Claude Code session?▼

You can analyze token usage in a Claude Code session by running the `/caveman-stats` command, which reads session logs directly to report actual input and output token numbers.

Can I see how many tokens I saved during a Claude Code session?▼

Yes, you can see token savings by viewing the estimated usage difference against a non-caveman baseline, which is calculated directly from your session log data without AI estimation.

Does Claude Code provide native statistics on actual input and output tokens?▼

Claude Code session logs contain actual input and output token data, and this skill processes those logs directly to display session count, token usage, and baseline savings percentages.

What is the best way to evaluate Claude Code session performance without AI estimation?▼

Evaluating Claude Code session performance without AI estimation involves reading session logs directly to extract precise token usage metrics and calculate savings against a predefined baseline.

Do I need any external dependencies to track token usage in Claude Code?▼

No external dependencies are required to track token usage in Claude Code, as the analysis operates directly on session logs within the environment using a simple command trigger.