Published onJuly 20, 2025Context or Command? How LLM Agents Can Be Tricked into Leaking DataSoftwareAIMCPSecurityInstruction-following isn't enough; we need agents with judgment, context, and a sense of consequence.
Published onJanuary 25, 2023Unintended Third-Party Access to Data Through Supported Azure Built-In RolessymmetryazuresecuritycloudA combination of built-in contributor permissions could allow unintended data access in Azure Lighthouse.