Passing variable to prompt on create_react_agent langgraph

[irufano]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

i have prompt, tool, and agent

VISUALIZE_PROMPT = """
{question}
Based on question if ask chart/graph do below task else just return "None".

{ai_response}
Based on AI response above, analyze data and extract to dataframe.
Then generate chart/graph based on dataframe code with matplotlib.
The chart/graph result must be in base64.

base64:
"""
# tool
python_repl = PythonREPL()
repl_tool = Tool(
    name="python_repl",
    description="A Python shell. Use this to execute python commands. Input should be a valid python command. If you want to see the output of a value, you should print it out with `print(...)`.",
    func=python_repl.run,
)

# agent
visualize_system_message = SystemMessage(content=VISUALIZE_PROMPT)
visualize_agent_executor = create_react_agent(
    llm, [repl_tool], state_modifier=visualize_system_message
)

# invoke
result = visualize_agent_executor.invoke(
        {"messages": [HumanMessage(content=question)]},  debug=True
    )

print(result)

Description

I have prompt with variable inside.
On documentation when invoking agent just pass "messages" as param and there is no vars to pass inside message.
How to pass value on variable prompt {question} and {ai_response} on create_react_agent langgraph when invoking?

System Info

langchain==0.3.1
langchain-core==0.3.7
langchain-text-splitters==0.3.0
langchain_openai==0.2.1
langchain_community==0.3.1
langgraph==0.2.32
langgraph-checkpoint==2.0.0
langchain_experimental==0.3.2

[MarcoRizk]

Not sure if I understand your need's correctly but you'd do something like

from langchain_core.prompts import ChatPromptTemplate

prompt = ChatPromptTemplate.from_messages([
    ("system", """"{question}
Based on question if ask chart/graph do below task else just return "None".

{ai_response}
Based on AI response above, analyze data and extract to dataframe.
Then generate chart/graph based on dataframe code with matplotlib.
The chart/graph result must be in base64.

base64:"""),
    ("placeholder", "{messages}"), ])

def format_for_model(state):
    # add your logic for adding questions and responses
    question = 
    ai_response = 
    return prompt.invoke({"messages": state["messages"], "question": question,"ai_response": ai_response})

visualize_agent_executor = create_react_agent(
    llm, [repl_tool], state_modifier=format_for_model
)

[ibrahimlambdatest]

I don't see an option for a state_modifier @MarcoRizk . What should i do?