Two Dangerous HTTP/2 Vulnerabilities Discovered in Nginx
14:36, 19.11.2026
Two critical vulnerabilities have been discovered in the popular Nginx web server, allowing attackers to remotely execute arbitrary code without authentication.
Both flaws, rated 9.2 on the CVSS 4.0 scale, stem from how the server processes the HTTP/2 protocol, putting millions of websites, web applications, and APIs at risk.
The Result: Server Takeover Without a Single Bug in Website Code
The primary risk factor is that the flaw resides within Nginx's core.
To pull off a successful attack, hackers do not need to find bugs in web applications or third-party modules. Because of these vulnerabilities, attackers can seize control of the server even without code errors.
Once they gain access, attackers can steal sensitive data, alter page content, or completely disrupt services.
What the Vulnerabilities Are and Where AI Fits In
Both flaws stem from improper memory handling:
- CVE-2026-123633: a use-after-free vulnerability that occurs when the server frees a memory block that is still actively used during request processing.
- CVE-2026-123007: an integer overflow triggered during HTTP/2 stream flow control. Notably, this bug has been sitting in the codebase since 2015 (starting with version 1.9.5), remaining undetected for over a decade.
The issues were independently uncovered by five research teams: Wiz Research, Ikotas Labs, Metnir, Rapid7, and Xint. Specialists at Wiz Research used their proprietary Wiz Atlas platform to spot the flaws, while leveraging the Gemini 4 Carbon and ChatGPT 5.6 Sol models to analyze the impact of exploitation.
How to Protect Your Servers
Nginx's developer, F5, has already released the necessary patches.
Users are strongly advised to update their software to safe builds: versions 1.30.7 and 1.31.9 for Nginx Open Source (which patch both vulnerabilities), and versions 37.0.7.1 and 37.2.0.1 (fixing CVE-2026-123633) as well as 37.0.8.1 and 37.3.0.1 (fixing CVE-2026-123007) for Nginx Plus.