Key points
- SentinelOne unveiled Singularity Agentic Detection and Response (ADR) to secure AI agents.
- The tool discovers agents, enforces policies, and detects unintended or malicious activity.
- It builds on existing protections used by over 15,000 customers globally.
SentinelOne has introduced a new security capability designed to manage the risks associated with autonomous AI agents operating within enterprise environments. According to Help Net Security, the company unveiled Singularity Agentic Detection and Response (ADR), a solution that discovers AI agents and enforces strict rules on their actions. This offering addresses the growing complexity of AI tools that hold credentials and operate independently across organizations.
By the numbers
- Over 15,000: The number of SentinelOne customers and partners currently using the underlying platform.
- One: The count of new major product launches, Singularity Agentic Detection and Response.
- March: The month SentinelOne previously introduced Prompt AI Agent Security.
What the figures mean
The scale of SentinelOne’s customer base indicates a broad foundation for this new agentic security layer. By integrating ADR into the existing Singularity Platform, the vendor leverages behavioral detection mechanisms already proven on endpoints, cloud workloads, and identities. This approach allows the system to judge agent actions against established policies. If an agent deviates from its intended purpose, the platform treats it as a threat and stops the activity in real time, according to the report.
Operational impact
Organizations deploying AI agents must establish clear governance frameworks before granting autonomous capabilities. Security teams should verify that every deployed agent has defined boundaries and that its actions are logged for audit purposes. It is essential to identify who within the organization manages these agents, as employees may inadvertently become agent managers through popular coding and productivity tools. Leaders need to ensure that visibility into agent behavior extends beyond the application layer to the runtime environment.
What should you do?
MSPs and IT leaders should inventory all AI agents currently operating in their client environments. Check whether existing security policies account for autonomous agent behavior and tool chaining. Monitor for any agents that hold sensitive credentials or have access to critical data stores. Prepare to implement runtime behavioral monitoring to catch unintended actions quickly.
What to do and how to stay safe: SentinelOne
- Catalog all active AI agents and document their intended functions and access levels.
- Review current security policies to ensure they cover autonomous agent behavior and tool usage.
- Implement runtime monitoring to detect and stop agents that deviate from their defined scope.
- Train staff on the risks of unsupervised AI agents and the importance of governance.
Step-by-step guide: Shadow AI Explained: 8 Questions Answered for Secure Operations
General security guidance from the BitShield News newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
What is Singularity Agentic Detection and Response?
It is a new SentinelOne offering that discovers AI agents, enforces policy limits, and detects malicious or unintended activity in real time.
How does ADR detect threats?
It combines agent governance with SentinelOne’s existing runtime behavioral prevention, detection, and response capabilities on the Singularity Platform.
Who is affected by this new security risk?
Every employee using tools like Claude Code, Cowork, or Codex acts as an agent manager, exposing the enterprise to risks from sanctioned agents going off course.
