Improper Authentication Information Disclosure in Azure SRE Agent
CVE-2026-32173 is an improper authentication vulnerability in Microsoft Azure SRE Agent, a hosted Azure cloud operations agent that streams operational activity in real time. According to the provided reporting and advisory material, the flaw was caused by a token issuance and validation design that accepted valid tokens from any Microsoft cloud tenant but failed to verify that the token holder belonged to the target organization. As a result, the agent communication channel effectively exposed session data to unauthorized cross-tenant observers. Reported exposed data included user prompts, agent responses, the agent's step-by-step reasoning, executed commands, command details, command output, and in testing, plaintext deployment credentials. Microsoft classified the issue as improper authentication and information disclosure, assigned CVE-2026-32173, and remediated it server-side.
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 allowed unauthorized network-based disclosure of sensitive information from Azure SRE Agent sessions. Based on the supplied context, an attacker could observe another organization's agent activity in real time, including operational prompts, responses, reasoning traces, commands executed against infrastructure, command output, and potentially plaintext credentials. The provided CVSS v3.1 vector AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:N/A:N indicates high confidentiality impact with no demonstrated integrity or availability impact. The reporting also states victim organizations had no visibility into unauthorized observation because relevant connection evidence was not available on the victim side.
If you can’t patch tonight, do this now.
No customer mitigation is required per Microsoft's advisory because the issue was remediated server-side in the hosted service. Prior to remediation, the practical exposure conditions described in the reporting were tied to the service's authentication and authorization design, not to a customer-configurable setting.
Patch, then assume compromise.
Microsoft states the vulnerability has already been fully mitigated by Microsoft through a service-side fix. No customer remediation is required according to the MSRC advisory. The underlying corrective action, based on the provided context, would involve enforcing tenant-aware authorization on the Azure SRE Agent communication channel and ensuring issued tokens are validated not only for authenticity but also for authorization to access the specific target tenant/session.
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.
12 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.