Cursor 101: Run AI Agents That Actually Ship Code

Cursor 3 replaces the old tab-and-prompt model with a full agent workspace. You’ll learn how to run parallel agents that write, test, and iterate on code autonomously. Use Composer 2, manage worktrees for multi-repo projects, and learn how the local-to-cloud handoff keeps your agents running even when your laptop closes. Explore the integrated browser for testing, the plugin marketplace, where you can extend agents using MCPs, and the autonomy slider for controlling how much your agents decide on their own.

Learning objectives
Configure the Cursor 3 agent workspace and navigate the new interface for managing local and cloud agents.
Demonstrate how to run multiple AI agents in parallel to build, test, and iterate on code simultaneously.
Apply Composer 2 to generate and refine code autonomously within a development project.
Explain how worktrees enable multi-repo development and isolated agent workspaces.
Describe the local-to-cloud agent handoff and when to use it to keep work running across environments.