SentinelOne brings runtime threat detection and response to AI agents

SentinelOne has unveiled Singularity Agentic Detection and Response (ADR), a new offering that discovers AI agents, enforces what agents are allowed to do, monitors agent behavior in depth, and detects and stops unintended and malicious activity in real time.

Bringing together visibility and governance of agent activity with SentinelOne’s signature runtime behavioral prevention, detection and response, the new offering is built directly on the Singularity Platform, giving new and existing customers a battle-tested foundation to secure AI running on endpoints, cloud workloads, and the broader enterprise. Singularity ADR is the newest addition to SentinelOne’s AI Security portfolio and a direct expansion of the company’s commitment to defend the surfaces AI runs on.

Securing the agents that power the AI enterprise

AI agents hold real credentials, chain tool calls, and work independently across the modern enterprise, and with Claude Code and Cowork, Codex, as well as other tools, every employee is now an agent manager. As recent high profile events have shown, the risk AI agents now carry is as likely to come from a sanctioned agent going off course as from an attacker. SentinelOne already protects the endpoint, identity, mobile, and cloud workloads where these agents run. Singularity ADR builds on that proven defense by focusing on agentic behavior and actions in runtime.

The new Singularity ADR benefits from the capabilities of SentinelOne’s fast growing product lines, SentinelOne Prompt Security and Purple, along with the modern endpoint, cloud workload and identity security capabilities already in use by over 15,000 SentinelOne customers and partners worldwide.

It governs which AI agents are allowed to run and what they can reach, and blocks known-bad actions at the moment of interaction. Anything that gets past policy is judged by the same behavioral detections the Singularity Platform already runs on those surfaces, so an agent acting outside its intent is caught and stopped like any other threat.

From agent governance to runtime protection

In March, SentinelOne bolstered its AI security portfolio with the introduction of Prompt AI Agent Security to discover AI agents and govern how they connect to tools and data. Singularity ADR builds on that foundation and takes agent security to the next level, pairing agent governance with the behavioral detection and response SentinelOne runs on endpoints, cloud workloads and identities, and ties every agent action to its origin. Security teams now get one view of what an agent was asked to do and what it actually did on the system, and they can stop a single agent mid-action without halting the rest of the workflow.

What Singularity ADR delivers

• Agent and MCP inventory. Every agent, tool, skill, connector, and MCP server across endpoints, SaaS, and cloud providers, sanctioned or not, with the permissions each holds and the employees behind it. Security teams know exactly what agents exist before deciding what to allow.
• Multi-platform agent coverage. Goes beyond the MCP gateway into Salesforce Agentforce, Microsoft Copilot Studio, OpenAI ChatGPT workspace agents, and Anthropic Claude managed agents. Coverage follows agents wherever the business builds them.
• Agent Traces. Every agentic event is monitored and sent to Singularity Data Lake so it can be visualized, retained, searched, and used in dashboards or detection rules.
• Unified Storyline. ADR event behavior is integrated with traditional EDR behavior in one execution graph, showing which AI components led to a process before it ran: the user request, the AI assistant, the MCP server that supplied the tool, and the file or process that followed. Analysts see the full chain in one place instead of piecing it together manually.
• Policy and execution governance. Allow and block controls for which agents run, who uses them, and what they can reach, with approval workflows for high-risk tools.
• In-line prevention. Blocks prompt injection, jailbreaks, credential exfiltration, and tool poisoning before the action executes.
• Intent drift detection. Flags the gap between what an agent was asked to do and what it did, so analysts stop reconstructing events by hand.
• Agent containment. Stops one agent mid-action while the rest of the workflow keeps running.
• Agentic Investigation and governed response. Automatically completes an investigation and executes a response within the boundaries set by customers.

“Securing AI agents goes beyond just stopping threat actors with malicious intent. As we’ve seen all too frequently, an equal if not more prominent risk comes from unintended behavior of sanctioned agents,” said Chris Corde, Chief Product Officer of SentinelOne. “By bringing to bear SentinelOne’s AI-based behavioral prevention, detection and response capabilities, we can stop rogue agentic actions using the same platform our customers already use and trust to protect their core attack surfaces.”

“AI Security has to cover everything AI touches: the surfaces it runs on, the models behind it, and now the agents acting on its behalf,” said Tomer Weingarten, CEO of SentinelOne. “Agents hold real credentials and act without waiting for anyone. Singularity ADR brings the runtime protection that already defends tens of millions of endpoints, workloads and identities to the agents doing real work inside the enterprise, so our customers can adopt AI at full speed.”