As reported by the Financial Times, Upwind Security has announced two new capabilities designed to protect the rapidly growing world of enterprise AI agents.

Upwind AI Detection and Response — now generally available — monitors AI agents in production to detect suspicious behavior in real time.

It helps security teams understand which agent was affected, how it was manipulated, what data it accessed, and how to respond.

The company also introduced its AI Agent Context Scanner, which examines the skills, tools, and Model Context Protocol (MCP) connections that shape how an AI agent operates.

By analyzing the instructions embedded across these components, Upwind can identify risky or malicious behavior before it compromises the broader environment.

Together, the new capabilities give organizations greater visibility and control as AI agents gain access to sensitive data, systems, and enterprise credentials.

This story was created with help from Omri Hurwitz Media