Loading vulnerability…
Elevation of Privilege in Azure AI Foundry M365 published agents
CVE-2026-35435 is a critical elevation-of-privilege vulnerability in Azure AI Foundry M365 published agents caused by improper access control. The flaw affects the hosted service capability used to publish AI agents into Microsoft Teams and Microsoft 365 Copilot. Insufficient authorization enforcement in agent-related operations allows an unauthorized attacker to access functionality or resources beyond intended permission boundaries over the network. Microsoft classifies the issue as an exclusively hosted cloud-service vulnerability with changed scope and high confidentiality impact.
What it means and what to do now: patch path, interim mitigations, and the assume-compromise checklist.
What an attacker gets, and what they’ve been doing with it.
Successful exploitation allows a remote, unauthenticated attacker to elevate privileges across a changed security scope boundary. The documented impact is high confidentiality loss, enabling access to data or resources beyond the vulnerable component’s intended authorization boundary. Microsoft’s published scoring indicates no direct integrity or availability impact.
If you can’t patch tonight, do this now.
No customer-side mitigation is required. The affected service was fully mitigated by Microsoft, and no additional user or administrator action is necessary according to the vendor guidance.
Patch, then assume compromise.
Microsoft remediated the vulnerability server-side in the affected Azure-hosted service by tightening access control checks for agent-related operations. No customer patch, KB installation, or tenant-side upgrade is required.
No public exploits tracked yet. Mallory keeps watching.
Products and vendors Mallory has correlated with this vulnerability. Open in Mallory to drill down to specific CPE configurations and version ranges.
11 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.