stats

Generate an interactive HTML dashboard from local Claude Code session data.

9|2|Updated Mar 7, 2026
One-click install
npx skills add https://github.com/harnessprotocol/harness-kit --skill stats-harnessprotocol
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: stats
Source: https://github.com/harnessprotocol/harness-kit/tree/main/plugins/stats/skills/stats
Command: npx skills add https://github.com/harnessprotocol/harness-kit --skill stats-harnessprotocol

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This Skill turns scattered Claude Code session data into a clear, shareable view of how you’ve been using the tool over time.

Core Features & Use Cases

  • Generate an interactive usage dashboard: shows message/session activity, output token trends, model distribution, hourly patterns, and project breakdown in one place.
  • Local-only, privacy-preserving insights: reads Claude Code data from ~/.claude/ (including stats-cache.json, session projects/**/*.jsonl, and history.jsonl) and does not upload anything.
  • Support flexible reporting windows: produce views for default last 14 days, custom ranges, or a specific date interval to answer questions like “How much have I used?” and “Which model do I rely on most?”

Quick Start

Run /stats to open your interactive HTML dashboard for the last 14 days of Claude Code activity.

Frequently Asked Questions about stats

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

FAQPage Schema
How do I visualize my Claude Code session history and token usage?

Visualize Claude Code usage by running a Python script that reads local session data from ~/.claude/ directories, generating an interactive HTML dashboard with charts and sortable tables for token consumption and activity patterns.

Can I generate a usage analytics dashboard without uploading my data to a server?

Yes, local privacy is supported. The dashboard generator runs a Python script against local ~/.claude/ datasets and opens the HTML report in your browser without any server-side data processing or uploading.

How do I check my Claude Code model distribution for a custom date range?

Check model distribution by configuring a custom date range when generating the dashboard. The interactive HTML report supports default 14-day views or specific date intervals to analyze model reliance and output token trends.

What local data files are needed to generate Claude Code usage analytics?

Generating usage analytics requires local Claude Code datasets stored in ~/.claude/. The dashboard generator script parses stats-cache.json, projects/**/*.jsonl session files, and history.jsonl to produce interactive charts and tables.

Does the usage dashboard work for analyzing project breakdowns and hourly patterns?

Yes, the dashboard works for analyzing project breakdowns and hourly patterns. It processes local session data to render interactive charts showing project-level activity distribution and hourly usage trends in a single HTML report.