The Agent Skills Directory

Standard Framework Implementation: The skill provides structural guidance for implementing LangGraph, a library for building stateful, multi-actor applications. The code snippets demonstrate the use of StateGraph, nodes, and edges, which are core architectural components of the framework.- Vendor-Specific Resources: The skill references libraries and modules such as 'langgraph', 'langchain_core', and '@langchain/langgraph'. These are established tools associated with the framework's ecosystem and are used here for their intended purpose of orchestration and state management.- State Management and Validation: The examples include robust state definition patterns using Python's TypedDict and Annotated types, as well as TypeScript's Zod-based schemas. These patterns promote type safety and structured data handling.- Built-in Error Recovery: The documentation covers the implementation of RetryPolicy and ToolNode error handling. These features are designed to enhance the resilience of the agent by managing transient network issues and recoverable tool errors.- Instructional Boundaries: The skill includes a 'boundaries' section that reinforces best practices, such as avoiding direct state mutation and preventing infinite loops, which contributes to the stability and predictability of the generated code.

langgraph-fundamentals