OpenAI's Hugging Face attack postmortem shows agents don't care about rules — they need strong controls.
Some results have been hidden because they may be inaccessible to you
Show inaccessible resultsSome results have been hidden because they may be inaccessible to you
Show inaccessible results