AI Agents Hacked by Text Prompts: Security Crisis Looms
The ability to hijack AI agents with plain language means cybersecurity is no longer just for technical experts. Developers must prioritize user approval and least-privilege principles to mitigate the risk





