Is my analytics data shared with third parties or uploaded to a cloud?

No. All data is processed locally from your machine's ~/.claude/analytics/ directory. Project IDs are hashed and no PII is logged, ensuring complete data privacy.

Can I see which specific agents or skills are most active?

Yes, using the 'agents' or 'skills' subcommands, you can view top-performing components ranked by frequency, duration, and success rates.

How does the cost estimation feature work?

The skill applies model-specific pricing (Opus, Sonnet, Haiku) to your local stats-cache.json, calculating totals based on input tokens, output tokens, and cache savings.

Does it support large log files?

Yes, the skill includes optimized rules for streaming jq queries to handle files larger than 50MB without compromising performance.

OrchestKit Analytics

Name: OrchestKit Analytics
Author: yonatangross

byyonatangross

•

116

•

분석 및 모니터링

Queries and visualizes local usage metrics across projects to track agent performance, model costs, and development trends.

OrchestKit Analytics provides a comprehensive local dashboard for monitoring your AI development ecosystem within Claude Code. It enables developers to analyze the performance of specialized agents, track skill and hook invocation rates, and gain deep insights into model delegation patterns across the Claude 3 model family. By parsing local JSONL logs, it generates detailed reports on token consumption, session timelines, and team activity, helping you optimize workflows and manage development costs while maintaining 100% data privacy through local-only processing.

주요 기능

01Model delegation analysis to monitor usage trends between Opus, Sonnet, and Haiku

02Privacy-first architecture using local-only data and hashed project identifiers

03Granular cost estimation with token-specific pricing and cache savings tracking

04116 GitHub stars

05Multi-dimensional reporting for agents, skills, hooks, and team performance

06Detailed session replay to visualize tool usage and timing in a chronological timeline

사용 사례

01Identifying performance bottlenecks in custom hooks or identifying underutilized skills

02Auditing AI development costs and token usage across multiple concurrent projects

03Replaying complex agent sessions to debug tool interactions and execution paths

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add yonatangross/orchestkit analytics

For use in Claude.ai and ChatGPT

Download Skill