Define LangChain Agents and LangGraph Workflow

Initialize your LangChain environment. Define two primary agents: a 'Creative Director' agent powered by GPT-5 Pro and a 'Creative Specialist' agent powered by Claude 4 Sonnet. Construct a LangGraph workflow that orchestrates the collaboration between these agents to generate a music video creative brief. The workflow should start with the Creative Director, then delegate to the Specialist for detailed sections, and finally consolidate. ```python
from langchain_core.messages import BaseMessage
from langchain_core.prompts import ChatPromptTemplate
from langchain_core.runnables import RunnablePassthrough
from langchain_openai import ChatOpenAI
from langchain_anthropic import ChatAnthropic
from langgraph.graph import StateGraph, END # Initialize LLMs
gpt_pro = ChatOpenAI(model="gpt-5-pro", temperature=0.7, openai_api_key="YOUR_OPENAI_API_KEY")
claude_sonnet = ChatAnthropic(model="claude-4-sonnet", temperature=0.7, anthropic_api_key="YOUR_ANTHROPIC_API_KEY") # Define agent prompts (simplified)
creative_director_prompt = ChatPromptTemplate.from_template("You are a Creative Director. Your task is to conceptualize a music video brief based on the user's input. Delegate detailed content generation to a specialist. User input: {input}")
creative_specialist_prompt = ChatPromptTemplate.from_template("You are a Creative Specialist. Your task is to elaborate on specific sections of a music video brief provided by the Creative Director. Section to elaborate: {section}, Context: {context}") # Define agent runnables (simplified)
creative_director_agent = creative_director_prompt | gpt_pro
creative_specialist_agent = creative_specialist_prompt | claude_sonnet # Define LangGraph state
class AgentState(BaseMessage): input: str brief_sections: dict = Field(default_factory=dict) current_task: str = "" next_node: str = "" # Build the graph (initial nodes)
graph_builder = StateGraph(AgentState) graph_builder.add_node("creative_director", creative_director_agent)
graph_builder.add_node("creative_specialist", creative_specialist_agent) # Define entry point
graph_builder.set_entry_point("creative_director") # Add conditional edges later
```