sweetkoffi/tuto-mcp-tshark

🧠 What This Project Does

This project connects a Large Language Model (LLM) to system-level tools like tshark, using a lightweight custom protocol called MCP (Message Control Protocol).
It allows an AI agent to request network traffic analysis from a Python-based microservice — and receive structured results it can reason about.

It’s composed of two main modules:

🧠 mcp-agent

Acts as the front-end interface for the LLM.

• Uses fast-agent-mcp to:
  • Format user input into MCP-compatible JSON-RPC messages
  • Send those to the MCP server over HTTP
  • Interpret structured responses into natural language replies
• Can operate as part of a chatbot, terminal agent, or notebook workflow.

🛠️ mcp-srv-tshark

A FastAPI server that acts as the backend executor.

• Receives MCP calls and currently supports one tool:

🔧 Included Tool: tshark_query

This tool executes tshark on a given .pcap file located in the pcaps/ directory.
It performs offline analysis , and returns a summarized result based on optional parameters ---currently only protocol filters and packet limits.

Launching tsark_mcp

This guide shows how to launch the tsark_mcp FastAPI server using exact terminal commands and sequence.

Step 1: List project contents to confirm structure

ls

You should see:

mcp-agent  mcp-srv-tshark README.md

Step 2: Access mcp-srv-tshark

cd mcp-srv

Step 3: Create Python virtual environment

python -m venv env-mcp 

Step 4: Activate the Python virtual environment

source venv-tshark-mcp/bin/activate

Step 5 Install dependancies

pip install -r requirements.txt

NB : make sure to add your own apikey to the fastagent.secrets.yaml file

Step 6: Run the server

python server.py

If successful, you'll see output similar to:

INFO:     Started server process [PID]
INFO:     Waiting for application startup.
INFO:     Application startup complete.
INFO:     Uvicorn running on http://0.0.0.0:5001 (Press CTRL+C to quit)

Access the API

Open your browser and go to:

http://localhost:5001/docs

This opens the Swagger UI for interacting with the API.

Launching Fast-Agent

This section explains how to launch the Fast-Agent and connect it to the tsark_mcp server.

Step 1: Navigate to the Fast-Agent directory

cd mcp-agent

Step 2: Create a virtual environment

python -m venv env-agent

Step 3: Activate the virtual environment

source env-agent/bin/activate

Step 4: Install dependencies

pip install -r requirements.txt

Step 5: Add your api key

Create a new file for you api key : fastagent.secrets.yaml

openai : 
  api_key: sk-.......

Step 6: Check Fast-Agent setup (optional)

fast-agent check

Step 7: Launch Fast-Agent

fast-agent go --model=openai --servers=tshark_mcp

That’s all you need to launch and operate both tsark_mcp and Fast-Agent.