Upwind Expands AI Agent Protection With Context Scanning and Runtime Detection

The security problem posed by AI agents is not really about the models themselves. It is about everything wrapped around them: the skills, the tools, the connections, and the instructions that determine what an agent actually does once it…