Security for AI and Agent Augmented Coding on Windows | OD853

From the creator

AI agents create new opportunities for Windows developers to help users do more with less effort. This session explores how agent development on Windows benefits from built‑in security, observability, and manageability at the OS level. Using real examples like sandbox in GitHub CLI , we show how de‑privileged execution and clear boundaries let developers safely move agents from experiments to production without sacrificing trust or velocity. To learn more, please check out these resources: * https://aka.ms/build26-next-steps 𝗦𝗽𝗲𝗮𝗸𝗲𝗿𝘀: * Klorida Miraj * Nazmus Sakib 𝗦𝗲𝘀𝘀𝗶𝗼𝗻 𝗜𝗻𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻: This is one of many sessions from the Microsoft Build 2026 event. View even more sessions on-demand and learn about Microsoft Build at https://build.microsoft.com OD853 | English (US) | Windows Pre-recorded | (300) Advanced #MSBuild Chapters: 0:00 - Introduction – Windows and the Rise of Agentic Software 00:01:36 - Agent Terrarium Example and Collaborative Agent Building 00:02:22 - The Role of Windows in Safe, Governable Agent Development 00:03:34 - Understanding Agent Risks: Non-Determinism and Execution Boundaries 00:05:13 - Governance Foundations: Containment, Detection, Identity, and Trust 00:06:00 - Containment Overview and Introduction of Microsoft Execution Containers SDK (MXC) 00:10:47 - Containment Spectrum: From Micro-VMs to Process Sandboxing and Cloud VMs 00:19:02 - Detection and Response with Defender and Local Capabilities Discovery 00:21:29 - Agent Identity Models a

Published 2026-06-03 · Watch on YouTube ↗
Availability and playback restrictions are controlled by the original channel.