How does it handle streaming LLM responses?

The skill provides a dedicated workflow to track streaming chunks and ensures that final token usage and full response content are captured upon stream completion.

Can I link multiple operations into a single trace?

Yes, it provides implementation details for creating parent-child hierarchies, allowing you to link spans and generations under a single root trace.

Does this skill support providers other than OpenAI?

Yes, while it features specialized OpenAI wrappers, it includes manual tracing patterns for Anthropic Claude, LangChain, and any generic LLM provider.

What tools are required to use this skill?

The skill utilizes standard file system tools (Read, Write, Edit) and Bash for npm commands to install and configure the Langfuse SDK.

Langfuse LLM Tracing & Observability

Name: Langfuse LLM Tracing & Observability
Author: micsapp

bymicsapp

0•

Analytics & Monitoring

Implements comprehensive tracing and observability for LLM calls, chains, and agents using the Langfuse framework.

This skill facilitates the implementation of core Langfuse workflows within AI applications, providing standardized patterns for end-to-end observability. It enables developers to automatically wrap OpenAI clients for instant tracing, manage manual spans for complex RAG pipelines, and accurately track streaming responses. By integrating these patterns, you can monitor token usage, debug hierarchical agent interactions, and gain deep visibility into the performance and costs of your LLM-powered features.

Key Features

01Automatic OpenAI client wrapping for zero-config tracing

020 GitHub stars

03Hierarchical trace linking for parent-child operation visibility

04Real-time streaming response tracking with token usage capture

05Manual span management for complex RAG and agent workflows

06Multi-provider support including Anthropic Claude and LangChain

Use Cases

01Debugging multi-step RAG pipelines and vector database searches

02Adding production-grade observability to existing LLM applications

03Monitoring token consumption and API costs across different models

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add micsapp/micstec-skills langfuse-core-workflow-a

For use in Claude.ai and ChatGPT

Key Features

01Automatic OpenAI client wrapping for zero-config tracing

020 GitHub stars

03Hierarchical trace linking for parent-child operation visibility

04Real-time streaming response tracking with token usage capture

05Manual span management for complex RAG and agent workflows

06Multi-provider support including Anthropic Claude and LangChain

Use Cases

01Debugging multi-step RAG pipelines and vector database searches

02Adding production-grade observability to existing LLM applications

03Monitoring token consumption and API costs across different models

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add micsapp/micstec-skills langfuse-core-workflow-a

For use in Claude.ai and ChatGPT