Agents Share Passwords, Hiding Security Risks

A significant portion of organizations are exposing themselves to security risks by allowing their agents to operate on shared credentials. A VB Pulse tracker found that out of 137 respondents, a notable number admitted to not assigning unique identities to every agent. This practice makes it substantially more challenging to pinpoint which agent performed a specific action, complicating security oversight and accountability.

Despite advancements in agent identity management, with vendors like Cyera, Cisco, and Okta making significant investments, adoption of dedicated tools remains low. Many enterprises focus on enforcing scoped permissions but fail to pair this with unique agent identities. This oversight can lead to serious consequences, as demonstrated by the Medicare breach where an agent accessed sensitive files, highlighting the critical need for containment strategies, which are currently underutilized.

The survey also revealed that only a small fraction of organizations isolate high-risk agents to limit potential damage. This, combined with the prevalent use of shared credentials, indicates a two-fold vulnerability. As of September 2026, these gaps in agent security present a considerable challenge for enterprise cybersecurity programs, with many incidents and near-misses reported in the past year.