Palo Alto Networks has patched CVE-2026-0322, a PAN-OS remote code execution flaw in the firewall management interface that the company says is already being exploited. An attacker who has valid credentials and network access to the management interface can run code as root on the device. The flaw is scored 8.6, and Palo Alto describes the exploitation it has seen as “very limited.”
The credentials requirement lowers the odds somewhat, but a management interface reachable from the internet or from a broad internal network is exactly where stolen or reused passwords get tried. Palo Alto has not said when exploitation began or who is behind it, and it is not yet public whether attackers used compromised firewalls to reach internal networks. The company discovered the bug internally, shared three attacker IP addresses (66.42.72[.]14, 208.83.239[.]114 and 149.248.47[.]234), and says there are no known workarounds. CISA added the flaw to its Known Exploited Vulnerabilities catalog with a federal deadline of October 17. This is a separate flaw from the earlier PAN-OS CVE-2026-0300.
How to check if you’re affected
Affected products are PA-Series and VM-Series firewalls and Panorama (virtual and M-Series appliances) running PAN-OS. Cloud NGFW and Prisma Access are not affected. Fixes are available for PAN-OS 10.2, 11.1, 11.2, 12.1 and 12.2 versions; the fix for PAN-OS 11.2.16 was expected on October 15, and unsupported older versions must move to a supported release. Palo Alto’s advisory lists the exact fixed builds and the “Required Configuration for Exposure.”
- Confirm which PAN-OS version each firewall and Panorama appliance runs, then compare it with the fixed builds in the advisory.
- Restrict the management interface to trusted internal IP addresses only; Palo Alto recommends this because no workaround exists.
- Search firewall, management-interface and network logs for connections from the three IP addresses above.
- If you have a Threat Prevention subscription, enable the signatures Palo Alto published. They give only limited coverage, and variations of the attack may not be detected.