Introduction to agentregistry

In this free, on-demand technical lab, take your first steps with the agentregistry project and learn how to curate, publish, and deploy MCP servers and agent skills.

• Initialize agentregistry: Start the agentregistry server, confirm the web UI is available, and prepare the CLI (arctl) for use.
• Publish an MCP server: Explore and publish the pokeapi-mcp-server from the registry, then run and interact with it using the MCP Inspector CLI to call tools such as getPokemonByName.
• Deploy an MCP server: Promote the server to a stable, deployable endpoint and generate a Cursor configuration for direct integration with AI clients.
• Add a new server: Register a new MCP server, including the canonical “everything” server by defining its metadata and package details directly in the UI.
• Add a skill: Register an agent skill (for example, the Anthropic “internal-comms” skill) and pull it locally using arctl for use with Claude or other AI clients.

Gain hands-on experience managing AI infrastructure with agentregistry and learn curating, publishing, and securely distributing AI artifacts across your environment.