Their llms.txt spells it all out: https://octanejs.dev/llms.txt

But it doesn't have the repeating yourself/over-emphasising that you sometimes see in LLM instructions when someone is trying to fight the model's default assumptions