소개
Langfuse Data Retrieval is a specialized skill designed to bring AI observability directly into the Claude Code environment. It allows developers to seamlessly fetch trace data, list recent runs, and inspect specific execution steps without leaving the terminal. By offering targeted output modes like 'prompts' for quality checks and 'flow' for latency investigation, this skill helps developers pinpoint issues in complex LLM chains, filter traces by metadata or evaluation scores, and maintain a clean, readable context during debugging sessions.