cloudbase-agent-ts_skill

This skill helps you build and deploy AI agents using the CloudBase AG-UI protocol in TypeScript, including adapters, servers, and web UI.

TypeScript

975

GitHub Stars

8

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill tencentcloudbase/cloudbase-mcp --skill cloudbase-agent-ts

adapter-development.md1.9 KB
adapter-langchain.md2.6 KB
adapter-langgraph.md4.9 KB
agui-protocol.md2.3 KB
server-quickstart.md3.3 KB
SKILL.md3.0 KB
ui-clients.md1.2 KB
ui-miniprogram.md4.1 KB

Overview

This skill provides a TypeScript SDK for building and deploying AI agents using the AG-UI protocol. It focuses on running agents as HTTP services, integrating with LangGraph and LangChain adapters, and delivering web or WeChat Mini Program UIs. Use it when you need a complete agent stack in TypeScript from backend to client.

How this skill works

The SDK exposes an agent-server that serves AG-UI endpoints and streams protocol events over HTTP. Adapters wrap different agent frameworks (LangGraph, LangChain) or let you implement AbstractAgent for custom logic. Clients connect with @ag-ui/client or the Mini Program UI package to send messages, receive streamed responses, and react to AG-UI events.

When to use it

Deploying TypeScript/Node agents as HTTP services with AG-UI protocol support
Building stateful workflows with the LangGraph adapter
Implementing chain-based agents using the LangChain adapter
Creating a custom adapter by implementing AbstractAgent for bespoke behavior
Connecting web or WeChat Mini Program UIs to an AG-UI-compatible agent server

Best practices

Choose the adapter that matches your architecture: LangGraph for state graphs, LangChain for chains, or custom for special flows
Expose clear /send-message or /agui endpoints and enable streaming to support progressive UI updates
Configure CORS, logging, and observability before deploying to CloudRun or other Node hosts
Keep agent state and side effects decoupled from the transport layer by implementing AbstractAgent cleanly
Test AG-UI events locally with @ag-ui/client to validate streaming and protocol behavior

Example use cases

Deploy an agent server that orchestrates multi-step workflows using LangGraph and streams progress to a web UI
Wrap an LLM chain in a LangChain adapter to expose it as an AG-UI compatible HTTP agent
Build a WeChat Mini Program chat interface that connects to the agent-server via the Mini Program UI package
Create a custom domain-specific agent by implementing AbstractAgent to handle specialized tools and actions
Integrate agent observability and request tracing before moving to production on CloudRun or similar

FAQ

No. This SDK targets TypeScript/JavaScript environments and relies on TypeScript types and Node tooling.

Can I deploy the agent server anywhere?

Yes. The server runs on Node.js and can be deployed to CloudRun, other container hosts, or any Node host; ensure network and CORS settings allow client connections.