What problem does it solve? Building stateful AI agents on Cloudflare Workers requires navigating Durable Objects, WebSocket routing, scheduling, MCP integration, and rapidly evolving SDK APIs. This Skill provides up-to-date retrieval sources, code patterns, and gotchas so you avoid outdated knowledge and common configuration mistakes. ## Core Features & Use Cases - Agent Implementation Guidance: Covers the Agent class, persistent state with setState, @callable RPC methods, scheduling, workflows, durable execution with runFiber, queues, and retries. - Chat & Client Integration: Documents AIChatAgent streaming chat, React hooks (useAgent, useAgentChat), server-driven messages, and human-in-the-loop approval flows. - Integrations & Experimental APIs: Includes MCP client/server setup, email handling, webhooks, push notifications, observability, plus experimental voice, browser tools, Think, and Codemode. - Use Case: You need to add a streaming chat agent with scheduled reminders to an existing Workers app. The Skill points you to the correct wrangler configuration, migration rules, and current documentation for each step. ## Quick Start Ask the AI to create a Cloudflare agent class with state, a callable method, and the correct wrangler.jsonc configuration using the agents-sdk skill.