Gen Digital Announces Agent Trust Hub to Secure AI Agents Against Prompt Injection

Gen Digital has announced the availability of Agent Trust Hub, a security platform aimed at protecting AI agents from manipulation through prompt injection. According to the company, AI agents integrated with email, web browsing, and plugin ecosystems can be tricked into executing unauthorized commands, stealing sensitive data, or deleting files through malicious emails, websites, and plugins. The announcement highlights that attackers can exploit the legitimate permissions granted to AI agents by injecting malicious prompts through various vectors.

Agent Trust Hub is designed to mitigate these risks through several protective measures. Skill verification is intended to ensure agents only execute authorized actions. Secure connectivity configurations aim to prevent communication with malicious endpoints. Runtime controls are designed to intercept and block dangerous actions before they are completed. The platform addresses the growing concern that AI agents, while powerful, can be manipulated when given broad access to systems and data.

The company states that as AI agents become more integrated into daily workflows, ensuring they operate within intended boundaries becomes critical. Agent Trust Hub represents Gen Digital's approach to providing greater oversight and control over agent behavior. The announcement does not detail specific CVE associations or active exploit activity but focuses on the general risk landscape and the protective capabilities of the new platform.