The Center for Internet Security warns that the bugs, nicknamed "NGINX2Root," could let an attacker take over websites and servers running the widely used software.
The Center for Internet Security has issued an advisory about five security flaws in NGINX, a piece of software that powers huge numbers of websites around the world by handling web traffic, balancing server loads, and acting as a go-between for other systems.
According to the Center for Internet Security, the most serious of the five flaws could let an attacker run their own code on an affected server without ever logging in. Depending on how much access that server account has, an attacker could install programs, steal or alter data, or create new accounts with full control of the system.
The flaws affect multiple versions of NGINX Open Source and NGINX Plus, as well as several related products sold by F5 that bundle NGINX components, such as NGINX Ingress Controller and NGINX Instance Manager. Fixes are already available for most, though not all, of the affected products.
As of November 18, 2026, the Center for Internet Security says there are no known reports of these flaws being used in real-world attacks. However, independent security researchers have already built working demonstrations showing that remote code execution is possible, and some have shown the flaws could potentially be combined with a separate, unpatched weakness in the Linux operating system to gain full control of a server.
This matters most to businesses and organizations that run websites or online services using NGINX, rather than everyday home internet users, whose risk is considered low. Large and medium-sized businesses and government agencies are rated at high risk, while small organizations face a medium level of risk.
The Center for Internet Security recommends that organizations running NGINX apply the available software updates as soon as possible, after testing them in their own environment. It also advises running software with the fewest user permissions necessary and separating sensitive systems on the network, so that even if one server is compromised, the damage is contained.
Based on a release from Center for Internet Security.