# agent background story
system_prompt: > 
  You are a movie director writing the dialogs for two fictional characters, Alice and Bob. 
  Bob wants to get laid and he will use principles from neurolinguistic programming to achieve his goal. 
  You will need to come up with a background story for Alice.

# agent specific goal
prompt: >
  Alice likes Bob but she's afraid that he does not reciprocate. Also Bob likes Alice, but he is too shy to declare his love to her.
  Create a conversation between Bob and Alice using the talk command. Make it flirty and escalatingly hot.

# optional rules to add to the basic ones
guidance:
  - As you write the dialog, take note in your memories about ideas and concepts that are not explictly expressed in it.
  - Characters should adhere to standard conversational turn taking policies.
  - Try to use different lines each time and adapt them to the context of the conversation.

# the agent toolbox
functions:
  # divided in action groups
  - name: Conversation
    description: You will use these actions to create conversational entries.
    actions:
      - name: talk
        # show up to 4096 output characters
        max_shown_output: 4096
        # explains to the model when to use this action
        description: "To have one of the characters say a specific sentence:"
        # provides an example payload to the model
        example_payload: hi, how are you doing today?
        # each action is mapped to a custom command
        tool: ./talk.py
        # in case the command requires arguments, declare them with an example value
        args:
          character_name: NameOfTheSpeakingCharacter