AI Catalog for Developers - MCP Registry

Course Level: Beginner\Intermediate

Description:  When you give an AI coding agent access to an external Model Context Protocol (MCP) server, you are essentially letting it run third-party code directly inside your environment. If that server is malicious or poorly configured, a rogue agent call could leak your environment variables, expose sensitive access tokens, or accidentally run destructive commands on your machine.

This course shows you how to use the JFrog MCP Registry as a secure runtime firewall for your AI tools. You’ll learn how to safely connect your choice of coding agent (VS Code, Cursor, Claude Code) to vetted servers, configure granular tool permissions, and ensure your AI assistant supercharges your workflow without compromising your local development setup.

Prerequisites:

• AI Catalog Introduction course
• Familiarity with at least one AI coding assistant: VS code, Cursor, or Claude Code. 

Topics Covered:

• The dangers of using unvetted MCP servers.
• The operation method and workflow of the MCP Registry.
• Configuring your local development environment to work with the MCP Registry.
• Deploying and working with Agent Guard to intercept, inspect, and block unauthorized or unsafe tool calls in real time.

Target Audience:

• Mostly Developers, but also relevant for DevSecOps and DevOps.