session-recap

To summarize a Claude Code session and track token usage, use the session-recap skill to fetch session details, recent commits, and key metrics like cost and context window size. It provides a concise overview of the work completed.

Yes, you can review recent commits made during a long-running AI agent task by requesting a session recap. The skill fetches session details and lists the associated commit history to help you understand code evolution.

To retrieve Claude session details and metrics, you need access to claude-view MCP tools for session listing and retrieval. This environment setup allows the skill to fetch and summarize your code sessions.

Session recap works for monitoring resource consumption by fetching key metrics from your Claude Code sessions. It displays token usage, cost, and context window information to help you track how many resources were consumed.

The best way to understand code evolution from an AI agent session is to generate a recap that combines session summaries with recent commit history. This provides a quick overview of code changes without manually sifting through logs.