A comprehensive reference for building agent workflows with LangGraph's state-based approach. Covers the essential patterns you'll need: designing state schemas with proper reducers (so your message lists actually append instead of overwrite), implementing nodes that return partial state updates, and wiring together conditional edges. The 5-step design methodology is solid for mapping complex workflows to graphs. Includes TypeScript and Python examples for StateGraph basics, Command/Send for advanced routing, and streaming patterns. Most useful when you need fine-grained control over agent orchestration beyond what basic LangChain agents provide.
npx skills add https://github.com/langchain-ai/langchain-skills --skill langgraph-fundamentals