FIX Onboarding Validator - MCP Server

# FIX Onboarding Validator - MCP Server

An MCP server for validating FIX protocol onboarding, with a LangGraph agent for automated workflows.

Features

• FIX Validator: Validates FIX messages against onboarding rules

• MCP Server: Exposes validation tools via Model Context Protocol

• LangGraph Agent: Autonomous agent that orchestrates the onboarding flow

• LLM-Enhanced Agent: Uses LLMs for intelligent parsing and error explanation

FOM-4: LangGraph + LLM Integration - BETA v0.9

Status: Beta complete - running locally, production-ready architecture

Current Implementation (Beta):

• Model: Llama 3.1:8b via Ollama - local, free, no API key required

• 4-node LangGraph: detect_and_validate -> explain_with_llm -> report_pass/fail

• Factory pattern allows provider swapping via LLM_PROVIDER env var

Planned Production (Placeholder - Intentional):

# In get_llm() factory:
elif provider == "anthropic":
    # PLANNED - Testing in progress, requires billing
    return ChatAnthropic(model="claude-3-5-sonnet-20240620")

Why this design:

• Avoids API costs during dev while proving LLM integration works

• Extensible - same graph works with Anthropic, OpenAI, local models

• To switch to production: LLM_PROVIDER=anthropic + ANTHROPIC_API_KEY

Run Beta:

ollama pull llama3.1:8b
python langgraph_agent_with_llm.py

Related MCP server: langgraph-clean-arch

Project Structure

fix-onboarding-mcp/
    fix_validator.py
    langgraph_agent.py
    langgraph_agent_with_llm.py
    .env.example
    README.md
    requirements.txt