caveman-stats

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

13|3|Updated Jun 10, 2026
One-click install
npx skills add https://github.com/tassiovale/claude-code-kit --skill caveman-stats-tassiovale
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: caveman-stats
Source: https://github.com/tassiovale/claude-code-kit/tree/main/skills/caveman-stats
Command: npx skills add https://github.com/tassiovale/claude-code-kit --skill caveman-stats-tassiovale

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill provides an accurate view of the token usage and estimated savings for Claude Code sessions, removing the guesswork and providing transparent insights.

Core Features & Use Cases

  • Real-time Token Usage: Offers the actual number of input/output tokens used during a session.
  • Estimated Savings: Shows the savings compared to a non-caveman baseline.
  • Immediate Feedback: The output is integrated into the Claude Code interface via a hook, providing immediate feedback on usage.

Quick Start

Run /caveman-stats to display session token usage stats.

Frequently Asked Questions about caveman-stats

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

FAQPage Schema
How do I check token usage in Claude Code sessions?

You can check token usage in Claude Code sessions by running the /caveman-stats command, which reads session logs to report actual input and output token counts without relying on AI estimation.

What is the best way to monitor real-time token usage during AI coding?

Real-time token usage monitoring during AI coding is achieved by integrating an output hook into the Claude Code interface, which provides immediate feedback on session token consumption as you work.

Does Claude Code provide savings analysis for token consumption?

Claude Code savings analysis is available through session log reporting, which calculates estimated savings by comparing current token usage against a non-caveman baseline.

Can I see actual token consumption instead of AI estimates for Claude Code?

Actual token consumption instead of AI estimates is displayed by analyzing Claude Code session logs directly, removing guesswork and providing transparent insights into input and output token usage.

Are there any dependencies required to track session token usage?

No dependencies are required to track session token usage, as the tool operates independently by parsing existing Claude Code session logs to deliver token reporting and savings analysis.