review-session

To analyze tool usage and efficiency after a completed session, you can run a session retrospective. This process audits tool invocations, detects user correction patterns, and evaluates error recovery to produce an actionable improvement report.

A session retrospective is a review process that analyzes your conversation to identify ineffective tool choices and wasted effort. It improves workflow by highlighting correction patterns and outputting a structured report with prioritized recommendations.

Yes, you can detect user correction patterns such as rephrasing or adding missed context to find AI performance gaps. Analyzing these corrections serves as the primary signal to understand where time or effort was wasted during a task.

You generate a retrospective report with tool usage metrics by asking the AI to review the full conversation history. The output enumerates tool successes and failures, evaluates search breadth, and provides structured tables of categorized issues.

The best way to evaluate error recovery and search breadth is through a session review analysis. This approach assesses tool appropriateness, identifies batching opportunities, and highlights retry loops to prevent repeated mistakes in future workflows.