avivsinaiavivsinaiCommunityยท1 Agent Skills Included

langfuse-mcp

Debug LLM traces, exceptions, sessions, and prompts

Connects Langfuse observability data to your coding environment for debugging LLM applications. Finds exceptions, inspects traces and sessions, analyzes latency, and manages prompts and datasets without opening the Langfuse dashboard. Provides step-by-step playbooks that move from broad error discovery to a concrete root-cause hypothesis.
npx skills add avivsinai/langfuse-mcp --all -g -y
Available:

Gives the agent project context, environment variables, and operational rules for running and extending the Langfuse MCP server safely.

All Skills in This Repository (1)

Pure Emerald Level Indicators

Frequently Asked Questions

FAQPage Schema
How to install langfuse-mcp?โ–ผ

Run `npx skills add avivsinai/langfuse-mcp --all -g -y` in your terminal to install the skill globally, then add your Langfuse API keys to connect the MCP server.

How to debug LLM errors with Langfuse?โ–ผ

Ask your agent to find recent exceptions, and it will group errors by file, pull the failing trace, and show the full stacktrace and context.

Does langfuse-mcp work with Claude Code and Cursor?โ–ผ

Yes. It runs as a standard MCP server and works with Claude Code, Codex, Cursor, and any MCP-compatible client.

Can I analyze LLM latency and cost?โ–ผ

Yes. Built-in metrics tools aggregate cost, token usage, and p95 latency by model without pulling raw traces.

Is there a read-only mode for safety?โ–ผ

Yes. Enable read-only mode with a single flag to disable all write operations on prompts, datasets, and annotation queues.

Related Repositories in Software Engineering

View All in Software Engineeringโ†’