Loading vulnerability…
Unauthenticated Privilege Escalation in Azure OpenAI
CVE-2026-119990 is a critical missing-authentication vulnerability in Microsoft Azure OpenAI that allows an unauthenticated attacker to elevate privileges over a network. A crafted internet request could permit access to orchestration information, alteration of workflows, or service disruption. The specific vulnerable function and resulting privilege level have not been disclosed. The vulnerability has a CVSS v3.1 base score of 9.8. Microsoft completed service-side remediation and requires no customer action.
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 could compromise the confidentiality, integrity, and availability of affected Azure OpenAI resources through unauthorized privilege escalation. Potential consequences include disclosure of orchestration information, unauthorized workflow modification, and service disruption. The exact access level and resource scope have not been disclosed. Microsoft reported no exploitation at the time of the original advisory.
If you can’t patch tonight, do this now.
Microsoft has fully mitigated the vulnerability within the Azure OpenAI service. No customer-side workaround or additional mitigation is required.
Patch, then assume compromise.
Microsoft has completed service-side remediation. Customers do not need to install updates or change configurations. Advisory revision 1.1, dated November 10, 2026, clarified that the vulnerability was fully mitigated and no customer action was required. No customer-installable fixed version is specified.
No public exploits tracked yet. Mallory keeps watching.
20 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.