cloudbase-agent

Deploy TypeScript AI agents as AG-UI protocol HTTP services on CloudBase.

Updated May 14, 2026
One-click install
npx skills add https://github.com/study-yang/Vieb-Coding-- --skill cloudbase-agent-study-yang
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cloudbase-agent
Source: https://github.com/study-yang/Vieb-Coding--/tree/main/%E5%BE%AE%E4%BF%A1%E5%B0%8F%E7%A8%8B%E5%BA%8F-%E5%BF%AB%E8%AE%B0/rules/cloudbase-agent/ts
Command: npx skills add https://github.com/study-yang/Vieb-Coding-- --skill cloudbase-agent-study-yang

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @cloudbase/agent-server, @cloudbase/agent-adapter-langgraph, @cloudbase/agent-adapter-langchain, @cloudbase/agent-ui-miniprogram, @ag-ui/client, @langchain/langgraph, @langchain/openai.

What problem does it solve? Building and deploying AI agents that stream responses to user interfaces requires bridging agent frameworks like LangGraph or LangChain with a standardized event protocol, plus handling server deployment, session persistence, and client connectivity. This Skill provides the complete workflow for deploying AG-UI compatible agents on Tencent CloudBase. ## Core Features & Use Cases - Agent Server Deployment: Expose agents as HTTP endpoints with SSE streaming using @cloudbase/agent-server, deployed via the manageAgent MCP tool on Node.js 20 runtime. - Framework Adapters: Wrap LangGraph StateGraph workflows or LangChain agents as AG-UI compatible agents, with CloudBaseSaver checkpointers for multi-tenant session persistence. - UI Client Integration: Connect web apps via @ag-ui/client or WeChat Mini Programs via @cloudbase/agent-ui-miniprogram, including client-side tool execution. - Use Case: Build a WeChat Mini Program chatbot backed by a LangGraph agent deployed on CloudBase, where the agent can invoke client-side tools like showing toasts or navigating pages. ## Quick Start Deploy a LangGraph agent to CloudBase by installing @cloudbase/agent-server and @cloudbase/agent-adapter-langgraph with @latest tags, then use the manageAgent MCP tool with Nodejs20.19 runtime.

Frequently Asked Questions about cloudbase-agent

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I deploy a LangGraph agent to CloudBase?

Compile your StateGraph workflow with a checkpointer, wrap it in LanggraphAgent, and serve it with @cloudbase/agent-server's run() function. Deploy using the manageAgent MCP tool with runtime Nodejs20.19 and installDependency set to true.

What is the difference between the LangGraph and LangChain adapters?

The LangGraph adapter wraps compiled StateGraph workflows using ClientStateAnnotation with pre-wired messages and client tools fields. The LangChain adapter wraps agents created with createAgent() and uses the clientTools() middleware for client tool support.

Why do I get ERR_MODULE_NOT_FOUND after deploying to CloudBase?

This error occurs when dependencies are not installed during deployment. Always set installDependency=true when calling the manageAgent MCP tool so the cloud environment installs dependencies automatically.

Can I use version ranges for @cloudbase/agent packages?

No, always install @cloudbase/agent-* packages with the @latest tag. Version ranges like ^1.0.0 cause dependency errors during cloud builds because these packages do not follow traditional semantic versioning.

How do client tools work in a WeChat Mini Program agent?

The client defines tools with handlers and sends them in the request. When the agent calls a client tool, the server routes to END, the Mini Program executes the handler locally via this.agui.addTool, and sends the result back to resume the run.

When should I use CloudRun instead of HTTP Cloud Functions for agents?

Use CloudRun when you need a custom Docker image with system-level dependencies like FFmpeg, resources exceeding Cloud Function limits, persistent local file storage, or native C extensions requiring specific OS packages.