Large language model hijacking attacks, also known as LLMjacking, are generating large fraudulent cloud charges, with some organizations losing more than $100,000 per day. [...]