Exclusive: Upwind adds context scanning for AI agents as AI DR hits general availability
Cloud security startup Upwind Security Inc. today unveiled AI Agent Context Scanner, which inspects the instructions, tools and connections feeding artificial intelligence agents before those agents act on them. The company also moved its AI Detection & Response service, or AI DR, into general availability.
AI agents are no longer just coding assistants on employee laptops. They now run autonomous workflows inside cloud environments, calling tools, calling application programming interfaces and reaching into data with live enterprise credentials.
What an agent does is shaped by more than the model behind it. Skills, tools, Model Context Protocol connections, stored memory and whatever instructions happen to be active get assembled into a context window at runtime. Any one of those components can inject an instruction that redirects the agent’s reasoning and turns its permissions on something the operator never intended.
The AI Agent Context Scanner analyzes those building blocks continuously, wherever they sit, from employee endpoints through cloud infrastructure to managed AI providers. Upwind said the point is catching malicious, vulnerable or simply ambiguous instructions before an agent acts on one. Findings are cross-referenced against the rest of the platform so teams can rank them by actual runtime risk instead of working through a queue of standalone scanner alerts.
AI DR covers the production side. It baselines how each agent behaves, including which tools it calls, which APIs and data stores it reaches and what an ordinary session looks like. Drift from that baseline gets flagged as it happens. Upwind then correlates the activity against workload identity, network topology, API traffic, data sensitivity, permissions, posture and exposure.
The output is meant to be more than a bare alert. AI DR shows which agent was affected, what it did and what it touched, and it can reconstruct the path that led there. It also identifies data that may have been exposed and suggests a response, such as revoking a session or a credential, cutting permissions back or auditing downstream systems.
Upwind’s argument is that AI inventory, the running catalog of models, agents, tools, skills and MCP connections a security team maintains, is a floor rather than a ceiling. Two agents can look identical in a catalog and behave nothing alike once they are running, the company said.
AI DR arrived in December as one piece of a wider AI security suite that also brought posture management, an AI bill of materials, network visibility and MCP tracing to the company’s cloud-native application protection platform. Today’s release makes it generally available.
Upwind is a venture capital-backed startup that raised $250 million in January in a round led by Bessemer Venture Partners, at a $1.5 billion valuation. Total funding for the company stands at about $430 million.
Image: Upwind
Support our mission to keep content open and free by engaging with theCUBE community. Join theCUBE’s Alumni Trust Network, where technology leaders connect, share intelligence and create opportunities.
15M+ viewers of theCUBE videos, powering conversations across AI, cloud, cybersecurity and more** 11.4k+ theCUBE alumni**— Connect with more than 11,400 tech and business leaders shaping the future through a unique trusted-based network.
About SiliconANGLE Media
theCUBE AIand theCUBE SuperStudios — with flagship locations in Silicon Valley and the New York Stock Exchange — SiliconANGLE Media operates at the intersection of media, technology and AI.
Founded by tech visionaries John Furrier and Dave Vellante, SiliconANGLE Media has built a dynamic ecosystem of industry-leading digital media brands that reach 15+ million elite tech professionals. Our new proprietary theCUBE AI Video Cloud is breaking ground in audience interaction, leveraging theCUBEai.com neural network to help technology companies make data-driven decisions and stay at the forefront of industry conversations.