If you're running LLM applications in production, you need observability that understands tokens, costs, and prompt versions, not just HTTP requests. This skill covers Langfuse's tracing system with drop-in integrations for OpenAI, LangChain, and LlamaIndex. You get automatic instrumentation for LLM calls, prompt versioning you can deploy like feature flags, and eval pipelines for catching regressions. The patterns show real implementations, from basic trace setup to LLM-as-judge evaluation. Self-hosted or cloud, either works. Most valuable when you're past the prototype stage and need to understand why your costs spiked or which prompt version performs better.
npx skills add https://github.com/sickn33/antigravity-awesome-skills --skill langfuse