Security researchers have discovered 10 new indirect prompt injection (IPI) payloads targeting AI agents with malicious instructions designed to achieve financial fraud, data destruction, API key ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results