What problem does it solve? Debugging LangChain and LangGraph agents requires digging through LangSmith Studio manually to find traces, errors, and tool call failures. This Skill brings trace fetching and analysis directly into the terminal so you can diagnose agent behavior, errors, memory operations, and performance without leaving your workflow. ## Core Features & Use Cases - Recent Trace Debugging: Fetch traces from the last N minutes and get a structured report of successes, failures, tool calls, durations, and token usage. - Deep Trace Analysis: Investigate a specific trace ID to see the full execution flow, identify the exact failure point, and receive root cause analysis with suggested fixes. - Error Detection & Session Export: Scan recent traces for error patterns and export organized debug sessions with traces and threads for sharing or archiving. - Use Case: Your agent suddenly stops responding. Ask the assistant to check recent traces; it runs langsmith-fetch, finds a Neo4j connection timeout in the failing trace, and recommends adding retry logic. ## Quick Start Ask the assistant to debug your agent by fetching and analyzing the LangSmith traces from the last five minutes.
