Build stateful AI agents on Cloudflare Workers using the Agents SDK.
npx skills add https://github.com/kkkaoru/dotfiles --skill agents-sdk-kkkaoru
Please help me install this Agent Skill. Skill: agents-sdk Source: https://github.com/kkkaoru/dotfiles/tree/main/.agents/skills-stroage/agents-sdk Command: npx skills add https://github.com/kkkaoru/dotfiles --skill agents-sdk-kkkaoru
@callable RPC methods, and automatic client synchronization via React hooks like useAgent. - Durable Background Work: Use Workflows, runFiber durable execution, built-in FIFO queues, and retries with exponential backoff for long-running tasks. - Chat, MCP & Integrations: Build streaming chat agents with AIChatAgent, connect to or build MCP servers with McpAgent, and handle email, webhooks, push notifications, voice, and browser automation. - Use Case: You need a chat application where each user session is a persistent agent that remembers conversation history, schedules follow-up reminders, and calls external MCP tools — all deployed on Cloudflare Workers. ## Quick Start Ask the AI to create a Cloudflare Workers agent using the Agents SDK with a Counter class that has persistent state and a callable increment method.High-intent search queries and answers about installing and using this skill.
Extend the Agent class from the agents package, define an initialState, and use this.setState to persist updates to SQLite. Register the class as a Durable Object binding in wrangler.jsonc with a new_sqlite_classes migration entry.
Extend AIChatAgent from @cloudflare/ai-chat and implement onChatMessage using streamText from the AI SDK. Always pass abortSignal and onFinish to enable proper cleanup, message persistence, and resumable streaming on reconnect.
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 MCP servers by extending McpAgent and registering tools and resources.
Enabling experimentalDecorators in tsconfig conflicts with the SDK's decorator implementation and breaks @callable RPC methods. Extend the agents/tsconfig base configuration instead, which applies the correct compiler settings.
Use the built-in queue for background tasks under 30 seconds with sequential FIFO processing. Use AgentWorkflow for long-running multi-step tasks, automatic retries, and human-in-the-loop approval flows that must survive failures.
State persists to SQLite automatically, but in-flight work does not. Use runFiber with ctx.stash checkpoints and implement onFiberRecovered to resume work after eviction, since only the stash data is restored, not the lambda.
mintuzBuild stateful AI agents on Cloudflare Workers using the Agents SDK.
aashahinBuild stateful AI agents on Cloudflare Workers using the Agents SDK.
wixelsCreate secure HTTP endpoints and webhook handlers for Convex backend applications.
jokken79Build stateful multi-actor AI agents with graph-based state management.