I want the agent to dynamically generate the first message with LLM. However, I have some problems with the "firstMessageMode": "assistant-speaks-first-with-model-generated-message" mode. When I'm using this mode, the agent repeats the introduction twice. That sounds strange because with the static message everything's ok. Tried different models and settings. Here is the settings I'm using: "firstMessage": "", "firstMessageMode": "assistant-speaks-first-with-model-generated-message", Call id's : 1702764f-892a-4836-9635-c119668bad63 ae4c6820-5e91-47f3-8c96-d554c7451e38 82f3a295-76c4-4074-96c3-49f4f8cb518f Can you please advise on how to fix the repeating intro phrases?