Find patterns in your Claude Code sessions.
Analyzes recorded Claude Code sessions to reveal usage patterns, failures, and opportunities to improve.
When to use it
Use for task retrospectives, debugging failures, weekly productivity reviews, finding skill opportunities, or understanding token usage.
Give it a project with recorded sessions; it returns a session summary with tool, agent, and skill usage.
What you provide
Runs the analysis script with Python; no version is specified.
Uses uv to run the Python analysis command.
Requires Braintrust tracing to be enabled through braintrust-claude-plugin.
This skill
Braintrust
Reads your Braintrust session traces