langchain-patterns

Overview

LangChain is the most widely used framework for building LLM applications. LangGraph adds stateful, multi-actor workflows with cycles. Together they provide chains, agents, retrieval, memory, and complex workflow orchestration.

Capabilities

• Build chains with LCEL (LangChain Expression Language)
• Create agents with tool use and reasoning
• Implement RAG with vector stores and retrievers
• Manage conversation memory and context
• Build graph-based workflows with LangGraph
• Integrate with 100+ LLM providers and tools

When to Use

Trigger phrases:

• "langchain patterns"

• "LangChain/LangGraph patterns — chains, agents, tools, memory, retrieval, graph w"

• Building LLM-powered applications (chatbots, RAG, agents)

• Needing structured chains for multi-step LLM workflows

• Building stateful agent workflows with branching logic

• Implementing retrieval-augmented generation

• Wanting a mature ecosystem with many integrations

When NOT to Use

• Task is outside your authorization scope
• You need to implement controls (use implementing-* skills)
• Task is about analysis, not action (use analyzing-* skills)
• You don't have access to target systems
• Task requires compliance expertise (consult professionals)
• Task is about defense, not offense (use defensive skills)

Pseudo Code

# Example workflow for this skill
def execute(input_data):
    # Step 1: Validate input
    if not input_data:
        raise ValueError("Input data is required")

    # Step 2: Process core logic
    result = process(input_data)

    # Step 3: Validate output
    validate_output(result)

    return result

LCEL Chain

from langchain_openai import ChatOpenAI
from langchain_core.prompts import ChatPromptTemplate
from langchain_core.output_parsers import StrOutputParser

llm = ChatOpenAI(model="gpt-4o")

prompt = ChatPromptTemplate.from_messages([
    ("system", "You are a helpful assistant."),
    ("user", "{input}"),
])

chain = prompt | llm | StrOutputParser()

result = chain.invoke({"input": "Explain quantum computing in simple terms."})