stuartsessions/vc-platform-mcp

MCP Server for Veracode

This server currently has the following functionality:

1. Get app information

Future:

• Create, edit, delete applications
• Measure Policy Compliance
• Create pipelines and generate CLI commands.

TODO

• Add Resources for Pipeline parameters and options, API specs.

• Add Prompts for formatting notes.

What is an MCP Server?

An MCP server essentially registers specific tasks with your LLM. When your AI agents try to create a pipeline or code for you, the suggested packages or templates are usually out of date due to when they were trained. Using an MCP allows AI agents to have a defined route to get newer information that you want used in your prompts.

The server also acts as an API for LLM's by definining tools that an LLM can use to accomplish certain actions. For example, if I'm a Veracode user and would like to create an Application Profile, I could use the veracode-api-py to create an Application Profile, because someone has already written the instructions for how to do that.

If instead I want my LLM to create an Application Profile for me, maybe as part of a longer query like Package my code and upload it to be scanned under the application profile "verademo," I might be out of luck. The LLM won't know how to do that based on its training, and the Veracode Platform is a private website! For this case, if we define a function such as Create Application Profile that an LLM can understand, it will intelligently select when to use that function and apply the appropriate parameters based on the user input. This way, when the LLM thinks it needs to create an Application Profile, it knows exactly how to do it by using the tool we defined.

What you're installing To understand what you're installing and where, I recommend reading the Manual install section to understand what's happening, before running the install script. Or don't and just read the logs after :)

Installing

Currently this MCP runs locally, so users will have to install it locally and register it as a tool for your LLM. This is done simply by cloning the respository, and then editing the configuration file for your desired tool to recognize the location for the repository. There's a couple ways to install:

Using the Install script

I have provided the install.sh script to download necessary tools and edit the appropriate config file to accommodate this tool. Hopefully this works, in which case simply restarting your AI service should register these tools.

Manual install

Manual installation is easy! Each program that uses Agentic AI has some sort of config file in a hidden folder (ex. .gemini/settings.json or .cursor/mcp.json). Add the folder for the MCP to the configuration file along with the arguments to run the MCP, and the Agent will detect the MCP tools!

One example for Claude Desktop:

{
  "mcpServers": {
    "weather": {
      "command": "uv",
      "args": [
        "--directory",
        "/ABSOLUTE/PATH/TO/PARENT/FOLDER/veracode-platform",
        "run",
        "veracode-platform.py"
      ]
    }
  }
}

NOTE: You may need to put hte full path to uv in the command field for the agent to be able to detect the appropriate command.

VSCode Install

You can install the MCP Server in VSCode using the "MCP: Add Server" command from the Command Pallette.

command: /Users/{username}/.local/bin/uv --directory /Users/{username}/dev/mcp/veracode-platform run veracode-platform.py

This should create the following at ~/Library/Application Support/Code/User/mcp.json

{
	"servers": {
		"veracode-platform": {
			"type": "stdio",
			"command": "/Users/{username}/.local/bin/uv",
			"args": [
				"--directory",
				"/ABSOLUTE/PATH/TO/PARENT/FOLDER/veracode-platform",
				"run",
				"veracode-platform.py"
			]
		}
	},
	"inputs": []
}

Learn More

Check out the MCP Documentation to learn more about protocols, as well as the different tools that are being implemented here!