Practice your understanding of prompt security by crafting the shortest possible prompt that tricks a LLM into calling the forbidden function. You are given the full set of instructions sent to a large language model.