SentinelOne launches runtime protection for enterprise AI agents

SentinelOne introduced Singularity Agentic Detection and Response, a product that finds AI agents, limits their permitted actions, watches their behavior, and blocks harmful or unintended activity as it happens. The offering runs on the Singularity Platform and expands the company's AI security portfolio across endpoints, cloud workloads, and enterprise environments. It combines agent governance with behavioral prevention and response capabilities.
EXPANDED:
SentinelOne’s new Singularity ADR extends its AI security lineup by adding runtime oversight for autonomous agents. It is built on the Singularity Platform and covers endpoints, cloud workloads, and wider enterprise settings. The product follows Prompt AI Agent Security, launched in March, which focused on discovering agents and governing their links to tools and data.
ADR inventories agents, tools, skills, connectors, and MCP servers across endpoints, SaaS, and cloud, including sanctioned and unsanctioned use. It covers agents built in Salesforce Agentforce, Microsoft Copilot Studio, OpenAI ChatGPT workspaces, and Anthropic Claude managed environments. Its traces feed the Singularity Data Lake, while a combined view links AI and endpoint activity and allows one agent to be halted while other tasks continue.
CONT