cost-audit

Analyze Claude session transcripts to identify and quantify token usage waste.

11|Updated Apr 16, 2026
One-click install
npx skills add https://github.com/berkcangumusisik/claude-code-practices --skill cost-audit-berkcangumusisik
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cost-audit
Source: https://github.com/berkcangumusisik/claude-code-practices/tree/main/skills/cost-audit
Command: npx skills add https://github.com/berkcangumusisik/claude-code-practices --skill cost-audit-berkcangumusisik

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Analyzes the current session's token usage, pinpoints where tokens were wasted, and produces concrete optimization actions you can apply in future sessions.

Core Features & Use Cases

  • Capture token costs and input/output breakdowns with /cost.
  • Detect common waste signals in transcripts (unnecessary reads, repeated reads, and large files) and surface actionable improvements.
  • Generate an actionable report with top leaks and practical fixes (e.g., add to .claudeignore or seed a CLAUDE.md scaffold).

Quick Start

Invoke /cost to capture the current session’s token usage, then review the resulting report for recommended fixes.

Frequently Asked Questions about cost-audit

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

FAQPage Schema
How do I audit token usage in a Claude transcript?

To audit token usage in a Claude transcript, invoke the /cost command to capture current session costs and identify optimization opportunities. The tool scans reads, duplicates, and large files to produce an actionable report.

What is token waste in a Claude coding session?

Token waste in a Claude coding session refers to unnecessary token consumption caused by repeated file reads, duplicate data, and large files. Identifying these waste signals allows you to apply fixes like adding paths to .claudeignore.

How do I optimize Claude costs for large transcripts?

You optimize Claude costs for large transcripts by scanning for common waste signals like unnecessary reads and large files. The tool then generates an actionable report suggesting practical fixes such as seeding a CLAUDE.md scaffold.

Can I use cost analysis on my current Claude session?

Yes, you can perform cost analysis on your current Claude session by invoking the /cost command. It captures input and output breakdowns and detects common waste signals directly from your active session context.

What are the limitations of auditing a Claude session transcript?

The primary limitation of auditing a Claude session transcript is that the analysis requires existing transcript context to function. Without an active session or large transcript history available, the tool cannot identify waste or suggest optimizations.