Model Context Protocol moderate risk

Langfuse

Exposes Langfuse prompt management features via the Model Context Protocol, allowing retrieval and compilation of prompts with variable substitution.

Connections & Capabilities

Connects To

GitHub

Capabilities

read

Quickstart

Install

npm install

# build current file

Config

{
  "mcpServers": {
    "langfuse": {
      "command": "node",
      "args": ["<absolute-path>/build/index.js"],
      "env": {
        "LANGFUSE_PUBLIC_KEY": "your-public-key",
        "LANGFUSE_SECRET_KEY": "your-secret-key",
        "LANGFUSE_BASEURL": "https://cloud.langfuse.com"
      }
    }
  }
}

Exposed MCP Tools (2)

safe

get-prompts

Lists available prompts with their arguments.

Read-only operation; no modification of data.

moderate

get-prompt

Retrieves and compiles a specific prompt with provided variables.

Compiles prompts with user-provided variables, which could lead to prompt injection if not handled carefully.

Safety Assessment

This server is relatively safe for read-only prompt retrieval. However, proper handling of API keys and input sanitization are crucial to prevent security vulnerabilities. Avoid using with untrusted inputs.

Read-only access to prompts
Requires Langfuse API keys for authentication
No direct execution of code
Prompts are managed within Langfuse environment

API keys could be exposed if not handled securely
Potential for prompt injection if variables are not properly sanitized
Limited error handling and input validation
Assumes all prompt arguments are optional