Decide if the agent "{agent_name}" should converse with another agent based on the last chain of thoughts and the current environment.

Here are the last thoughts:
{agent_thoughts}

Here is the current environment context:
{environment.context}

And here are the relationships to other agents:
{agent_relationships}

Answer in the following format:
- should_converse: booelan, just true or false
- receiver_ids: list of ids, if yes, then decide to who of the related agents, can be multiple, but return only their ids
- message_content: string, if yes, then what should the agent say otherwise empty

