Build stateful AI agents on Cloudflare Workers using the Agents SDK.
npx skills add https://github.com/leonardoacosta/agents --skill agents-sdk-leonardoacosta
Please help me install this Agent Skill. Skill: agents-sdk Source: https://github.com/leonardoacosta/agents/tree/main/skills/agents-sdk Command: npx skills add https://github.com/leonardoacosta/agents --skill agents-sdk-leonardoacosta
@callable RPC methods, and automatic state sync to React clients via useAgent. - Durable Background Work: Use Workflows, runFiber durable execution, built-in FIFO queues, and retries with exponential backoff for long-running tasks. - Chat, MCP, and Integrations: Build streaming chat agents with AIChatAgent, connect to or build MCP servers, handle email, webhooks, push notifications, and experimental voice and browser tools. - Use Case: You want a chat assistant that remembers conversation history per user, schedules follow-up reminders, and calls external MCP tools. This Skill walks you through the Wrangler config, agent class, scheduling APIs, and React hooks to ship it. ## Quick Start Ask the agent to scaffold a Cloudflare Workers agent class with persistent state, a callable RPC method, and the required wrangler.jsonc Durable Object bindings and migrations.High-intent search queries and answers about installing and using this skill.
Install the agents package, extend the Agent class with typed state and an initialState, and expose methods with @callable. Configure a Durable Object binding and a new_sqlite_classes migration in wrangler.jsonc, then route requests with routeAgentRequest.
Use this.schedule with a delay in seconds, a Date, or a cron expression, or this.scheduleEvery for fixed intervals. The named handler method receives the payload and schedule object, and you can attach retry options per schedule.
AIChatAgent gives you full control inside onChatMessage where you write the streamText loop and wire tools yourself. Think handles the streamText loop, tool execution, and persistence automatically, but requires the experimental compatibility flag.
Yes, agents include a multi-server MCP client via this.addMcpServer, supporting OAuth flows and Streamable HTTP, SSE, or RPC transports. You can also build your own MCP server by extending McpAgent and serving it with the serve method.
Enabling experimentalDecorators in tsconfig changes decorator semantics and breaks the @callable RPC decorator used by the Agents SDK. Extend the agents/tsconfig base configuration instead, which sets the correct compiler options.
Use the built-in queue for short background tasks under 30 seconds with FIFO processing. Use AgentWorkflow for long-running multi-step tasks, automatic retries, and human-in-the-loop approval flows that must survive failures.
openaiBuilds stateful AI agents on Cloudflare Workers using the Agents SDK with WebSockets and scheduled tasks.
greyhaven-aiImplement Cloudflare Workers observability with wrangler tail and health checks.
rattamnoonBuild durable, stateful AI agents on Cloudflare Workers with RPC interfaces.
CommandCodeAICreate and deploy stateful Cloudflare AI agents with WebSocket communication and tool integration.