Access Control - 1

2026 (8)

September (2)

github: Require proof of presence for high-impact actions

GitHub

2026-09-24

GitHub Enterprise Cloud now supports requiring proof of presence before members perform high-impact actions. This adds an interactive re-authentication or MFA challenge on top of the existing sudo-style control.

github: Workflow execution protections in GitHub Actions generally available

GitHub

2026-09-17

Workflow execution protections in GitHub Actions are now generally available for GitHub Enterprise, organizations, and repositories. The feature lets you define an allowlist to control who can execute workflows.

August (2)

azure: [Launched] Generally Available: Azure Bastion shareable link expiration

Azure Updates

2026-08-26

Azure Bastion shareable links now support an administrator-defined expiration date and time. After expiration, the link can no longer be used to connect.

commandline: Stop restricting the agent. Start restricting its environment.

Microsoft Command Line Blog

2026-08-21

Azure SRE Agent argues that AI safety is better achieved by constraining the runtime environment rather than the agent itself. This approach aims to preserve autonomy while reducing risk compared with human review or standing credentials.

July (3)

github: Manage GitHub Copilot app access with a dedicated policy

GitHub

2026-07-27

GitHub Copilot now has a dedicated policy for controlling app access at both the enterprise and organization levels. This separates Copilot app access from the previous broader access controls.

azure: [In preview] Public Preview: Protect sensitive generative AI telemetry in Application Insights and Microsoft Foundry

Azure Updates

2026-07-20

Application Insights now stores generative AI content in a dedicated GenAIContent table (AppGenAIContent in Log Analytics), enabling access controls for sensitive AI telemetry. The update is in public preview.

security: Least privilege for AI agents: Identity, access, and tool binding

Microsoft Security Blog

2026-07-16

Microsoft Security published a post on applying least-privilege principles to AI agents, focusing on identity, access control, tool binding, and auditing as agents become more autonomous.

June (1)

github: Restrict issue creation to collaborators only

GitHub

2026-06-29

Repository admins can now limit issue creation to collaborators with write access. This adds control over who can open issues and helps reduce unwanted issue creation.