Discourse AI - AI bot

I wonder if this is attribute to either Model logic (LLM hallucination) or the way we have Tool calls implemented (haven’t seen this yet)

Have you tried other LLMs or modifying the prompts to follow instructions at all cost?