The article outlines how to register a Model Context Protocol (MCP) server using ERC-8004, a standard designed to enhance trust and discoverability for AI agents. MCP serves as a bridge between large language models (LLMs) and various applications, allowing them to interact with external data and tools. The challenge is ensuring these tools can be trusted, as they involve executing arbitrary code. ERC-8004 addresses this by providing a framework for registering and verifying these tools on the blockchain, building a reputation based on actual usage. The tutorial provides step-by-step instructions for creating a simple MCP server with FastMCP, requiring only 30 lines of code. Users can install FastMCP, define their server, and deploy it using a guide linked in the article. The process includes using Google Colab for easy execution of Python code without local setup. The tutorial emphasizes querying the MCP server for tools, resources, and prompts, highlighting a secure method for handling environment variables and API keys. A significant focus is on the trust aspect of MCP tools. The article introduces ERC-8004 as a solution for verifying the identity and reliability of these tools through a public reputation system. It mentions available resources for exploring registered agents, such as 8004scan.io, and suggests using the Agent0 SDK for programmatic interaction with ERC-8004. The article provides practical insights into both building and managing MCP servers while addressing the critical issue of trust in AI tools.