State in Condition

This example demonstrates how to access the run context in the evaluator function of a condition step

This example shows:

1. How to use run_context in a Condition evaluator function
2. Reading and modifying run_context.session_state based on condition logic
3. Accessing user_id and session_id from run_context.session_state
4. Making conditional decisions based on run_context.session_state

1. Create a Python file

   access_session_state_in_condition_evaluator_function.py

   from agno.agent import Agent
   from agno.models.openai import OpenAIResponses
   from agno.workflow.condition import Condition
   from agno.workflow.step import Step, StepInput, StepOutput
   from agno.workflow.workflow import Workflow
   from agno.run import RunContext
   
   
   def check_user_has_context(step_input: StepInput, run_context: RunContext) -> bool:
       """
       Condition evaluator that checks if user has been greeted before.
   
       Args:
           step_input: The input for this step (contains workflow context)
           run_context: The run context object
   
       Returns:
           bool: True if user has context, False otherwise
       """
       print("\n=== Evaluating Condition ===")
       print(f"User ID: {run_context.session_state.get('current_user_id')}")
       print(f"Session ID: {run_context.session_state.get('current_session_id')}")
       print(f"Has been greeted: {run_context.session_state.get('has_been_greeted', False)}")
   
       # Check if user has been greeted before
       return run_context.session_state.get("has_been_greeted", False)
   
   
   def mark_user_as_greeted(step_input: StepInput, run_context: RunContext) -> StepOutput:
       """Custom function that marks user as greeted in session state."""
       print("\n=== Marking User as Greeted ===")
       run_context.session_state["has_been_greeted"] = True
       run_context.session_state["greeting_count"] = run_context.session_state.get("greeting_count", 0) + 1
   
       return StepOutput(
           content=f"User has been greeted. Total greetings: {run_context.session_state['greeting_count']}"
       )
   
   
   # Create agents
   greeter_agent = Agent(
       name="Greeter",
       model=OpenAIResponses(id="gpt-5.2"),
       instructions="Greet the user warmly and introduce yourself.",
       markdown=True,
   )
   
   contextual_agent = Agent(
       name="Contextual Assistant",
       model=OpenAIResponses(id="gpt-5.2"),
       instructions="Continue the conversation with context. You already know the user.",
       markdown=True,
   )
   
   # Create workflow with condition
   workflow = Workflow(
       name="Conditional Greeting Workflow",
       steps=[
           # First, check if user has been greeted before
           Condition(
               name="Check If New User",
               description="Check if this is a new user who needs greeting",
               # Condition returns True if user has context, so we negate it
               evaluator=lambda step_input, run_context: not check_user_has_context(
                   step_input, run_context
               ),
               steps=[
                   # Only execute these steps for new users
                   Step(
                       name="Greet User",
                       description="Greet the new user",
                       agent=greeter_agent,
                   ),
                   Step(
                       name="Mark as Greeted",
                       description="Mark user as greeted in session",
                       executor=mark_user_as_greeted,
                   ),
               ],
           ),
           # This step always executes
           Step(
               name="Handle Query",
               description="Handle the user's query with or without greeting",
               agent=contextual_agent,
           ),
       ],
       session_state={
           "has_been_greeted": False,
           "greeting_count": 0,
       },
   )
   
   
   def run_example():
       """Run the example workflow multiple times to see conditional behavior."""
   
       print("=" * 80)
       print("First Run - New User (Condition will be True, greeting will happen)")
       print("=" * 80)
   
       workflow.print_response(
           input="Hi, can you help me with something?",
           session_id="user-123",
           user_id="user-123",
           stream=True,
       )
   
       print("\n" + "=" * 80)
       print("Second Run - Same Session (Skips greeting)")
       print("=" * 80)
   
       workflow.print_response(
           input="Tell me a joke",
           session_id="user-123",
           user_id="user-123",
           stream=True,
       )
   
   
   if __name__ == "__main__":
       run_example()
   
2. Set up your virtual environment

   Mac

   uv venv --python 3.12
   source .venv/bin/activate
   

   uv venv --python 3.12
   .venv\Scripts\activate
   
3. Install dependencies

   uv pip install -U agno openai
   
4. Export your OpenAI API key

   Set OpenAI Key

   Set your OPENAI_API_KEY as an environment variable. You can get one from OpenAI.

   Mac

   export OPENAI_API_KEY=sk-***
   

   setx OPENAI_API_KEY sk-***
   
5. Run Workflow

   python access_session_state_in_condition_evaluator_function.py