The value of an agent is not just what it can do, but whether it can be Large language model trusted in production. Under the Secure Future Initiative, continuously strengthening this foundation remains a company-wide priority. “Microsoft Execution Containers (MXC), integrated with OpenShell, provides a policy-driven foundation for private, on-device agents on Windows.”
Containment bounds what agents can access and do, so non-deterministic behavior doesn’t translate into uncontrollable risk. The same infrastructure serving millions of human users now becomes the foundation for the next generation of intelligent work. An AI AI Leadership Moves Put Enterprise Deployment in Focus agent interacting with a GUI requires the same fundamental resources as any user—compute, network, identity and policy controls.
Process isolation provides fast, lightweight containment within the user’s environment for scenarios like running model-generated code within a dedicated process boundary that restricts access to files and network domains outside defined policy. This means organizations can automate complex workflows, scale operations without adding headcount and unlock new productivity gains—all while maintaining enterprise-grade security and compliance. Access policies can be applied to Windows session isolation so agents run independently with controlled local access and full lifecycle governance managed through Microsoft Entra and Intune in the cloud. These enhancements go beyond simply running AI agents on Cloud PCs—they optimize how agents are provisioned, managed and controlled, while maintaining enterprise-grade security and compliance. Together, these investments help provide the secure foundation on which trustworthy agentic computing is built. When that foundation is in place, organizations can scale agent adoption while maintaining control and trust.
Session isolation
GitHub Copilot CLI has adopted MXC process isolation to constrain what dynamically generated and executed code can do. Windows supports a range of containment options so that guardrails can match the nature and risk of the workload. The same policy model and SDK can map to different isolation constructs depending on the workload and containment requirements. Developers define what to constrain in their apps and agents, and Windows enforces those constraints consistently at runtime through MXC. We’ve previously shared the principles guiding how we secure agent workflows on Windows. Security for agents must be built into the foundation by design so they can be developed, deployed and governed with confidence.
- These enhancements go beyond simply running AI agents on Cloud PCs—they optimize how agents are provisioned, managed and controlled, while maintaining enterprise-grade security and compliance.
- We’ve previously shared the principles guiding how we secure agent workflows on Windows.
- Is there a way to provide the desirable properties of process isolation like low overhead with a stronger isolation boundary?
- This agentic security model runs on a Windows platform designed to reduce risk by default.
- This mitigates UI spoofing, input injection and cross-session data leakage, and is suited for sustained workflows that run alongside the user’s own work.
Windows as the platform for intelligent work
This innovation gave organizations the flexibility to scale computing resources instantly, reduce IT complexity and strengthen security—all while empowering employees to work from virtually anywhere. Windows continues to evolve so developers and organizations can move fast on AI while maintaining trust and security. Developers need control over what an agent can access and trust that those controls will hold,” said Dillon Rolnick, CEO of Nous Research. Windows assigns a local ID or a cloud provisioned identity backed Generative artificial intelligence by Entra and attributes all activity from the container to that identity, so you can clearly differentiate human from agent.
The Microsoft Execution Containers (MXC) SDK
- Windows 365 for Agents, now generally available, extends containment beyond the local device.
- Windows assigns a local ID or a cloud provisioned identity backed by Entra and attributes all activity from the container to that identity, so you can clearly differentiate human from agent.
- Developers are building agents that read files, invoke services, modify environments and chain operations together at increasing speed.
- The same policy model and SDK can map to different isolation constructs depending on the workload and containment requirements.
- Integrating MXC via OpenShell provides developers with an easy-to-deploy package for autonomous, always-on agents safely.
Developers are building agents that read files, invoke services, modify environments and chain operations together at increasing speed. From elastic resource pools to human-in-the-loop safeguards, these innovations help organizations automate complex tasks, reduce idle costs and ensure trust in autonomous operations. By extending the same Artificial neural network trusted environment from human users to agent workloads, businesses can accelerate innovation and improve employee productivity.
- As agents evolve, we are continuing to expand MXC containment capabilities and invite developers and the broader ecosystem to share feedback, including through engagement with the project on GitHub.
- GitHub Copilot CLI has adopted MXC process isolation to constrain what dynamically generated and executed code can do.
- In 2021, Microsoft introduced Windows 365, reimagining the PC as a cloud service that streams a Cloud PC—a complete, secure, personalized Windows experience to any device, anywhere.
- As they become persistent participants in how software runs, they introduce new risk to control and trust, challenging the security assumptions that have defined computing for decades.
Policy-based controls
Sessions in Windows separate the agent’s execution from the human user’s environment, such as the interactive desktop, clipboard, UI, input devices and active sessions. For developers, Windows will provide the building blocks needed to implement agents that are more secure on both consumer and enterprise systems. Containment, identity and manageability are built as foundational primitives in Windows, extending security beyond the app and model into the OS. As they become persistent participants in how software runs, they introduce new risk to control and trust, challenging the security assumptions that have defined computing for decades. Microsoft AI Microsoft Security Dynamics 365 Microsoft 365 Microsoft Power Platform Microsoft Teams Microsoft Copilot Small Business Azure Microsoft Developer Microsoft Learn Support for AI marketplace apps Microsoft Tech Community Microsoft Marketplace Software https://aicurated.com/ companies Visual Studio Careers About Microsoft Company news Privacy at Microsoft Investors Diversity and inclusion Accessibility Sustainability In just four years, Windows 365 has become Microsoft’s flagship software-as-a-service (SaaS) solution for delivering secure, managed, enterprise-grade computing globally, helping businesses lower costs, simplify management and accelerate productivity.
Start building secure agents today
AI in pharma: Accelerating development with Microsoft The Microsoft Cloud Blog – Microsoft
AI in pharma: Accelerating development with Microsoft The Microsoft Cloud Blog.
Posted: Mon, 14 Sep 2026 07:00:00 GMT source
We are partnering with leading innovators in the industry like Hermes, Manus, NVIDIA, OpenAI and OpenClaw, to ensure the containment we are building supports real developer needs. With the future addition of MXC integration, Windows 365 for Agents will scale from lightweight local isolation to stronger hardware-backed boundaries – through a single SDK and policy model. Windows 365 for Agents, now generally available, extends containment beyond the local device.
