Thank you. I find in interesting that the LLM just understands intuitively from the english name of the tool/function and it’s argument names. I had imagined it might need more extensive description and specification in its system prompt, but apparently not.