framework-selection

Selects the appropriate LangChain, LangGraph, or Deep Agents framework layer for agent projects.

Updated May 1, 2026
One-click install
npx skills add https://github.com/ricardoo022/4dill --skill framework-selection-ricardoo022
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: framework-selection
Source: https://github.com/ricardoo022/4dill/tree/main/.gemini/skills/framework-selection
Command: npx skills add https://github.com/ricardoo022/4dill --skill framework-selection-ricardoo022

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Choosing between LangChain, LangGraph, and Deep Agents at the start of an agent project is confusing because the frameworks overlap. This Skill provides a structured decision process that maps your task requirements to the correct framework layer before you write any agent code. ## Core Features & Use Cases - Decision Guide: A question-based flowchart that routes tasks to LangChain, LangGraph, Deep Agents, or a combination based on planning, state, and control-flow needs. - Framework Profiles: Detailed capability breakdowns for each layer, including built-in Deep Agents middleware such as TodoListMiddleware, FilesystemMiddleware, and SubAgentMiddleware. - Mixing Layers: Patterns for combining frameworks, such as registering a LangGraph compiled graph as a subagent inside a Deep Agents orchestrator. - Use Case: Before building a multi-step research agent, consult this Skill to determine that Deep Agents fits the planning and file-management needs, then follow its pointers to the deep-agents-core skill. ## Quick Start Ask the AI to invoke the framework-selection skill to determine whether LangChain, LangGraph, or Deep Agents is right for your agent project.

Frequently Asked Questions about framework-selection

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I choose between LangChain, LangGraph, and Deep Agents?▼

Use LangChain for single-purpose agents and RAG pipelines, LangGraph when you need custom control flow like loops, branching, or parallel workers, and Deep Agents for long-running tasks requiring planning, file management, and subagent delegation. The frameworks are layered, so higher layers include lower ones.

When should I use LangGraph instead of LangChain?▼

Use LangGraph when your agent needs complex control flow such as retry loops, dynamic branching, parallel fan-out, human-in-the-loop approval, or persistent state across invocations. LangChain's create_agent suits simple fixed tool loops.

Can I combine Deep Agents with LangGraph in one project?▼

Yes, a LangGraph compiled graph can be registered as a subagent inside Deep Agents. The Deep Agents orchestrator delegates to it via the task tool without needing to know its internal graph structure.

What middleware does Deep Agents include by default?▼

Deep Agents ships with TodoListMiddleware for task planning, FilesystemMiddleware for file operations, and SubAgentMiddleware for delegation, all enabled by default. SkillsMiddleware, MemoryMiddleware, and HumanInTheLoopMiddleware are opt-in.

Does LangGraph support middleware like Deep Agents?▼

No, LangGraph has no middleware layer. You wire behavior directly into nodes and edges, giving full control over graph structure but requiring manual implementation of planning, memory, and file handling.