Upcoming
Lessons from Building Multi-Session Agents with Cursor + Claude Code SDK
Workshop
In this workshop, you'll learn how to build a long-running AI agent harness by combining Cursor and the Claude Code SDK.We'll start by using Cursor to plan out a full application through conversation, defining features, architecture, and requirements. Then, we'll take that plan and adapt Anthropic's Claude Code SDK agent architecture to execute it. You'll see how to modify the SDK's prompts and workflow to align with your specific project.By the end of the session, you'll watch a live agent handle context management, write progress logs, and maintain state across sessions, all the patterns Anthropic outlined for building agents that don't lose track of their work.You'll walk away with a working workflow you can apply to your own projects.
