THE CRUNCH

F5 has released emergency patches for 10 vulnerabilities in open source NGINX, the web server that handles HTTPS traffic and routes requests for roughly a third of all websites, around 225 million in total. Two of the flaws, tracked as CVE-2026-123633 and CVE-2026-123007, allow an unauthenticated remote attacker to run code on the server and both carry a critical 9.2 CVSS rating. Both require HTTP/2 to be enabled on the affected server. Anyone running NGINX should update to version 1.31.9 as soon as possible.

The bugs were found independently by researchers from Wiz, Ikotas Labs, Metnir, Rapid7 and Xint, most with an AI assist, and reported to F5. Tomer Schwartz, who leads exposure management and AppSec at Wiz, told The Register that the wave of AI-discovered flaws points to a new reality. "AI has changed not only which vulnerabilities can be found, but how quickly longstanding risks can become real threats," he said. The CVE-2026-123007 flaw reaches back to NGINX 1.9.5 from 2015, meaning it has been lurking for at least a decade.

Both critical flaws are memory-safety issues in how NGINX handles HTTP/2 traffic, according to Schwartz and fellow Wiz researcher Nir Ohfeld. The first bug frees a connection's memory while a request still depends on it, producing a use-after-free condition. The second stems from faulty flow-control accounting that results in an invalid memory copy. Rapid7 published a proof of concept demonstrating unauthenticated remote code execution against a vulnerable Docker image running NGINX 1.31.8.

Xint found three further NGINX flaws plus a separate Linux kernel vulnerability that can be chained with any of the five to gain root privileges on the host, a combination the firm dubbed "nginx2root". It published videos showing the attacks working on AWS, Microsoft Azure, Google Cloud Platform and bare-metal servers. The disclosure follows a string of similar AI-found flaws in foundational software, including Java4Shell, PHP2Shell and a critical Apache flaw under active exploitation.