EN ▾
SNAPE REP RT Front page →
ConfirmedHacks & tools18 Nov, 16:55 UTC

Researchers using AI find five critical NGINX flaws that allow unauthenticated server takeover

The Register's illustration for its report on AI-found NGINX vulnerabilities
The Register

Wiz, Ikotas Labs, Metnir, Rapid7 and Xint independently reported the bugs, all rated 9.2; one HTTP/2 flaw dates to 2015, and F5 shipped fixes in NGINX 1.31.9 and 1.30.7.

Read at Infosecurity Magazine Share

Why it matters

NGINX serves about a third of websites, roughly 225 million, so the bugs extend the wave of AI-assisted discoveries that has already hit PHP, Apache and Java. The Register says most of the teams found the flaws with AI help. Xint chained one bug with an unpatched Linux kernel flaw to gain root, and Rapid7 published a working proof of concept. Xint saw no exploitation at publication. Administrators who cannot upgrade are told to disable HTTP/2 until they patch.

Key facts

  • Five critical CVEs, each rated 9.2 (CVSS 4.0)
  • Fixed in NGINX 1.31.9 (mainline) and 1.30.7 (stable)
  • CVE-2026-123007 affects versions back to 1.9.5 (2015)
  • NGINX runs about 225 million websites
  • Found independently by Wiz, Ikotas Labs, Metnir, Rapid7 and Xint
  • Xint chained a bug with a Linux kernel flaw to get root

Sources

More on Hacks & tools
More from Snape ReportFront page →