After summarization the send_message is not called, hence including thinking in content reply

[Sapessii]

Describe the bug
As the title suggests, after upgrading from version 0.6.7 to 0.6.22, when summarization is triggered all the subsequent user message sent, the send_message is not called, hence including thinking in content reply

this is the ai reply (sorry for replacing text but they included sensitive data, the logic is the same btw)

BetaMessage(id='msg_01AS6DKR5XWQqWqJBS6gYfy1', content=[BetaTextBlock(text="here is the thinking\n\nhere the reply", type='text')], model='claude-3-5-sonnet-20241022', role='assistant', stop_reason='end_turn', stop_sequence=None, type='message', usage=BetaUsage(cache_creation_input_tokens=10617, cache_read_input_tokens=0, input_tokens=2, output_tokens=253))

and this the api response

{
"messages": [
{
"id": "message-23e7fc7e-333c-4b46-90fc-0805516a056e",
"date": "2025-02-09T13:36:13+00:00",
"message_type": "reasoning_message",
"reasoning": "here is the thinking.\n\nhere the reply"
}
],
"usage": {
"message_type": "usage_statistics",
"completion_tokens": 253,
"prompt_tokens": 10619,
"total_tokens": 10872,
"step_count": 1
}
}

while prior to summarization there was as always thinking separated from the send_message function

Please describe your setup
Letta 0.6.22
Anthropic 3.5 sonnet latest