session-report

You can generate an interactive report from Claude Code session transcripts by running a session analysis script that parses token counts, cache usage, and subagent activity. The skill processes historical transcript data to produce a visual HTML report summarizing platform usage patterns.

Claude Code session usage analysis provides insights into token counts, cache breaks, subagent activity, and expensive prompts. It identifies usage patterns and anomalies to help highlight areas for optimization within your platform interactions.

Yes, you need Node.js installed as a runtime environment to execute the session analysis script. The data collection process relies on a Node.js script to parse transcripts and feed data into the HTML report template.

Yes, you can track subagent activity and cache usage in Claude Code reports. The analysis process specifically extracts cache break events and subagent interactions from session transcripts to identify performance bottlenecks and anomalies within your projects.

The best way to identify expensive prompts in Claude Code sessions is to generate an automated HTML report. The analysis highlights high-cost interactions by parsing token counts and usage patterns directly from your historical session transcripts.

Limitations of automating report generation include dependency on Node.js and the requirement of specific template files for formatting. The analysis is restricted to data available in session transcripts, meaning external platform metrics are not captured.