Arrakis Security’s 2026 Autonomous Agents Threat Report found that 76% of enabled AI agents had not run in 60 days, leaving them with retained access to company systems.

Earlier attacks and coverage

No earlier related attack in our archive yet.

Business impact

Organizations face potential data breaches and unauthorized access due to inactive AI agents, which can read and write sensitive data, call tools, and start workflows.

Industries affected

All industries, as the analysis covered thousands of agents in customer production environments.

Users affected

Employees and customers may have their data exposed if dormant AI agents are activated through unsecured triggers.

Sources