Check these three things.
-
System owner: confirm each vCenter runs at least the fixed build for its release line: 8.0 U3m, 9.0.2.0200, 9.1.1.0200, or 5.2.5 in VMware Cloud Foundation 5.x.
Affected versions and fixes
- Affected, from Broadcom's advisory and the CVE record:
- vCenter 9.1.x before 9.1.1.0200.
- vCenter 9.0.x before 9.0.2.0200.
- vCenter 8.0 before 8.0 U3m.
- vCenter in VMware Cloud Foundation 5.x before 5.2.5.
- vCenter 7.0 is also affected. Broadcom offers its fix only through an extended support contract. vSphere 7 reached End of General Support on October 2, 2025.
- Telco Cloud Platform 3.0, 4.x, 5.0.x, 5.1.x and 5.2, and Telco Cloud Infrastructure 3.0, are affected. Broadcom points to its article KB463116 for their fix.
- Broadcom lists no workaround.
- Record for each vCenter:
- Installed version and build.
- Source of that record.
- Date checked.
- Affected, from Broadcom's advisory and the CVE record:
-
Network owner: check which systems can reach vCenter's directory service on TCP port 2012.
When the risk applies
- Broadcom: an attacker with network access to vCenter may bypass authentication and gain unauthorized access. No login is needed.
- Google Threat Intelligence Group says the flaw is reached through the vCenter directory service on port 2012 and gives a session as the built-in administrator, administrator@vsphere.local.
- In the case Google describes, the attackers reached vCenter from other internal systems they had already broken into. vCenter does not need to face the Internet to be at risk.
- The CVSS vector includes AV:N, meaning network-reachable, including internal networks. It is not proof of Internet exposure.
- A vCenter that ran an affected version while port 2012 was reachable was exposed for that time, even if it is fixed now.
-
Security owner: look in logs for new vCenter administrator accounts and for SSH being turned on, even if it is off now.
What the records can show
- Signs Google reported from the campaign it investigated:
- A new local administrator account in vsphere.local.
- SSH turned on through the appliance management interface (VAMI, port 5480), then turned off again.
- Tools placed in /tmp on a separate internal Linux appliance.
- Internal scans for ports 443, 2012 and 389.
- Records to review:
- vCenter single sign-on and audit logs.
- Appliance management (VAMI) logs.
- Firewall logs for connections to port 2012.
- Limits:
- Clean records do not prove no attack happened.
- Missing records leave the question open.
- Patching does not remove an attacker who is already inside.
- Signs Google reported from the campaign it investigated:
If records show a break-in, start incident response now. Update vCenter whether or not port 2012 appears restricted.