Helpnetsecurity Dormant AI Agents Retain Access to Sensitive Data, Report Reveals
Article Content
- •76% of AI agents have not run in 60 days but retain access credentials.
- •44% of dormant agents have not run in 180 days, posing security risks.
- •70% of agents connect to sensitive systems with inadequate data loss prevention.
A report by Arrakis Security indicates that 76% of AI agents have not run in 60 days but still retain their credentials and connections to company systems. The analysis, based on telemetry from thousands of agents over 12 weeks, found that 44% had not run in 180 days. A quarter of these dormant agents can be activated via unsecured webhooks, allowing potential unauthorized access. Moreover, 70% of the agents connect to sensitive systems, yet 99.4% lack verified data loss prevention measures. The report highlights that many companies do not remove access for unused agents, which poses a significant risk to data security. Tal Baron, CEO of Arrakis Security, emphasized the lack of ownership over these agents as a contributing factor to the issue. The report also noted that 82% of agents carry at least one exposure, with many having excessive permissions.
Ask AI about this cluster
Answers cite the sources they use
Timeline
More articles in this cluster (3)
Following this threat?
Track Arrakis Security in your own feed — you're alerted when they show up in new reporting, leak sites or exploitation.
Free account · no card needed
Common questions
What percentage of AI agents are dormant?
How can dormant agents be activated?
What are the implications of these findings?
Continue Reading
.NET Vulnerability CVE-2026-111497 Requires Immediate Action On October 25, 2026, Microsoft released out-of-band security updates for.NET addressing CVE-2026-111497, a vulnerability in the Brotli decompression algorithm that allows unauthorized attackers to elevate privileges over a network. The flaw affects.NET versions 8, 9, 10, and 11, particularly in applications that…
Critical BeyondTrust 9.3 Vulnerability Actively Exploited by Threat Actor A critical vulnerability (CVE-2026-102764) in BeyondTrust Remote Support has been discovered and is actively being exploited by the threat group ARMORY CASTLE, reportedly linked to China. The vulnerability, which has a CVSS score of 9.3, allows unauthenticated attackers to upload files and potentially compromise host…