Posted by Mountain View, USA: SentinelOne, the AI-powered cybersecurity company has announced an integration of its Singularity platform with Anthropic’s Claude Cowork.
This integration allows SentinelOne to expand its AI security coverage to Anthropic’s agentic work application Claude Cowork also known as Claude. The move is part of SentinelOne’s AI Security portfolio, strategy and its commitment to defend the surfaces AI runs on.
Today, more numbers of employees are using and directing AI agents to work on their daily tasks. And that further expands the AI attack surface. A lot of those AI work happens on the endpoint, in cloud workloads and across enterprise apps.
Interestingly, in the case of Claude Cowork, the majority of that agentic activity happens remains invisible to security software on the device.
However, the SentinelOne’s Singularity platform integration with Claude Cowork puts a light into that potential blind spot. It complements SentinelOne’s signature endpoint detection and response (EDR) and cloud workload security (CWS) offerings. Along with recently unveiled Singularity Agentic Detection and Response (ADR).
SentinelOne Inside the Claude Cowork VM
Claude Cowork does its work inside a virtual machine (VM) on the employee’s device. The integration will place SentinelOne’s runtime security directly inside that VM. Security teams can then see what an AI agent does on the system, receive alerts on suspicious behaviour. And autonomously stop attacks using the same runtime detection and response SentinelOne already runs across endpoints.
What the integration delivers
- Visibility inside the Cowork VM
A SentinelOne agent will run inside the VM, so security teams can see what Cowork does there instead of working around a blind spot. - Behavioural detection and agentic tracing
SentinelOne leverages the same Behavioural AI detection engine that runs on its endpoints and will apply it to activity in the VM. So analysts can be alerted to anomalous behaviour and agent drift. - Coverage across operating systems
The integration will support Claude Cowork across macOS and Windows devices. - Customer-controlled rollout
IT teams will enable the integration through device management, so customers decide when and where it turns on.
“Our integration strategy across the AI ecosystem is simple: secure AI work where it happens. Claude Cowork is one of the most popular AI applications on the market, used by millions of workers around the world, so it is exactly where defenders need to be,” said Melissa K. Smith, SVP – Global Strategic Partnerships and Initiatives, SentinelOne.
“With this integration, security teams can say yes to the way their people want to work with AI, with the confidence that SentinelOne is covering that work,” added Smith.