Promptware Attacks Against LLM-Powered Assistants Are Practical and Dangerous | Not Hacker News!