What providers does the TanStack AI skill support?

It supports a wide range of providers including OpenAI, Anthropic, Google Gemini, and local models via Ollama through provider-agnostic adapters.

How does it ensure type safety during tool execution?

By leveraging Zod schemas for tool definitions, it ensures that all inputs and outputs are validated and typed across the entire network boundary.

Is this skill compatible with modern web frameworks?

Yes, it is designed for production use with React, Solid, Next.js, and TanStack Start, providing framework-specific hooks like useChat.

Can I prevent agents from performing unauthorized actions?

Yes, the skill supports a 'needsApproval' flag for tools, which forces the system to pause and wait for explicit user confirmation before executing sensitive logic.

How does this skill handle real-time streaming?

The skill utilizes Server-Sent Events (SSE) on the server and specialized fetch adapters on the client to ensure low-latency, responsive chat experiences.

TanStack AI Integration

Name: TanStack AI Integration
Author: secondsky

bysecondsky

•

データサイエンスとML

Builds type-safe, provider-agnostic AI chat interfaces and agentic workflows using TanStack AI's streaming and tool-calling capabilities.

概要

This skill integrates TanStack AI into Claude Code, providing a production-ready framework for building complex AI applications. It enables developers to switch seamlessly between LLM providers—including OpenAI, Anthropic, Gemini, and Ollama—while maintaining full type safety via Zod. The skill excels at implementing responsive streaming chat UIs, managing hybrid client-server tool execution, and enforcing human-in-the-loop approval flows for agentic cycles, making it ideal for developers building robust AI features in React, Solid, or Next.js applications.

主な機能

End-to-end type safety using Zod for tool schemas and model options
Provider-agnostic LLM orchestration for OpenAI, Anthropic, Gemini, and Ollama
21 GitHub stars
Built-in streaming support via Server-Sent Events (SSE) and connection adapters
Advanced agentic loop management with iteration caps and manual approval flows
Hybrid tool architecture separating client UI logic from server-side operations

ユースケース

Implementing complex agentic workflows that require human-in-the-loop validation
Building real-time, streaming AI chatbots with React or Solid frontends
Creating isomorphic tools that bridge backend data fetching with frontend UI mutations

概要

主な機能

End-to-end type safety using Zod for tool schemas and model options
Provider-agnostic LLM orchestration for OpenAI, Anthropic, Gemini, and Ollama
21 GitHub stars
Built-in streaming support via Server-Sent Events (SSE) and connection adapters
Advanced agentic loop management with iteration caps and manual approval flows
Hybrid tool architecture separating client UI logic from server-side operations

ユースケース

Implementing complex agentic workflows that require human-in-the-loop validation
Building real-time, streaming AI chatbots with React or Solid frontends
Creating isomorphic tools that bridge backend data fetching with frontend UI mutations