framework-selection
INVOKE THIS SKILL at the START of any LangChain/LangGraph/Deep Agents project, before writing any agent code. Determines which framework layer is right for the task: LangChain, LangGraph, Deep Agents, or a combination. Must be consulted before other agent skills.
Security Assessment
About framework-selection
The framework-selection skill is designed to guide developers in choosing the appropriate framework layer for LangChain, LangGraph, and Deep Agents projects. It solves the problem of uncertainty when starting a new AI agent project by clarifying which framework layer—foundation, orchestration, or full agent environment—best fits the task requirements. By consulting this skill before any other agent code, developers can ensure their project architecture aligns with the complexity and scope of their task, avoiding unnecessary overhead or missing capabilities.
This skill provides a structured decision guide that evaluates the needs of a project through key questions about task complexity, control flow, state persistence, and multi-step planning. It profiles each framework layer: LangChain for simple, single-purpose agents or pipelines; LangGraph for agents requiring complex control flows, loops, and conditional branching; and Deep Agents for long-running, multi-dimensional tasks that involve persistent memory, file management, and subagent delegation. Additionally, it indicates which subsequent skills to invoke depending on the chosen framework, ensuring a smooth and coherent setup for agent development.
Typical use cases include determining the right framework for a document Q&A pipeline, a multi-step workflow requiring human-in-the-loop approvals, or an open-ended task needing persistent memory and skill delegation. Target users are developers and AI practitioners working with LangChain, LangGraph, or Deep Agents who want a reliable, systematic approach to framework selection before implementing agent logic. This skill is particularly useful for teams aiming to standardize project structure and optimize the choice of frameworks for efficiency and maintainability.
FAQ
When should I invoke the framework-selection skill?
It should be invoked at the very start of any LangChain, LangGraph, or Deep Agents project, before writing any agent code or selecting other skills.
Can I use multiple framework layers in the same project?
Yes. Higher layers like Deep Agents can incorporate LangGraph graphs and LangChain primitives, allowing you to leverage multiple layers together.
Which framework should I choose for a simple single-purpose agent?
LangChain is ideal for single-purpose agents that take input, run tools, and return results without complex control flow or persistent state.
What limitations does this skill have?
This skill only guides framework selection based on task requirements; it does not implement agent logic or provide APIs for execution.
Do I need to consult this skill if my project already uses Deep Agents?
Even if using Deep Agents, consulting the skill helps confirm that you are selecting the appropriate combination of layers and invoking the right supporting skills.
Install framework-selection
Quick Setup:
- Copy the skill folder to
.claude/skills/ - Claude will automatically detect and use the skill
Repository
langchain-ai/langchain-skills