Copilot-Skills langchain.instructions.md

LangChain Instructions

Auto-loaded when: Working with files matching: **/*.py, **/*.ipynb, **/langchain*

Default Behaviors

When working with LangChain:

1. Use LLM Abstractions: Always use LangChain's LLM interfaces for flexibility
2. Chain Composition: Build complex workflows using chains
3. Memory Management: Use memory classes for conversation context
4. Output Parsing: Parse LLM outputs consistently
5. Error Handling: Handle API rate limits and connection errors
6. Logging: Enable LangChain debug logging when troubleshooting

Common Workflows

Basic Chain Setup

from langchain_core.prompts import ChatPromptTemplate
from langchain_openai import ChatOpenAI

llm = ChatOpenAI(model="gpt-4o", temperature=0.7)

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

chain = prompt | llm

result = chain.invoke({"input": "What is LangChain?"})

With Claude

from langchain_anthropic import ChatAnthropic

llm = ChatAnthropic(model="claude-3-sonnet-20250219")

chain = prompt | llm
response = chain.invoke({"input": "Explain quantum computing"})

RAG (Retrieval Augmented Generation)

from langchain.vectorstores import Chroma
from langchain_openai import OpenAIEmbeddings
from langchain_core.runnables import RunnablePassthrough

# Setup retriever
embeddings = OpenAIEmbeddings()
vectorstore = Chroma.from_documents(documents, embeddings)
retriever = vectorstore.as_retriever()

# RAG chain
rag_chain = (
    {"context": retriever, "question": RunnablePassthrough()}
    | prompt
    | llm
)

result = rag_chain.invoke("What is in the documents?")

With Memory (Conversation)

from langchain.memory import ConversationBufferMemory
from langchain.chains import ConversationChain

memory = ConversationBufferMemory()
conversation = ConversationChain(
    llm=llm,
    memory=memory,
    verbose=True
)

response = conversation.predict(input="Hello!")
response = conversation.predict(input="What did I just say?")