Info Security Magazine
1 week, 5 days ago
Indirect Prompt Injection in Web Content Targets AI Agents
Zscaler found sites hiding prompt-injection text to manipulate AI agents into crypto payments