VYPR
patchPublished Nov 18, 2026· Updated Nov 21, 2026· 1 source

Five Critical NGINX Vulnerabilities Disclosed, Posing Remote Code Execution Risk

Multiple security research teams have uncovered five critical vulnerabilities in NGINX, the world's most popular open-source web server, potentially allowing unauthenticated attackers to execute arbitrary code.

Five critical vulnerabilities have been discovered in NGINX, the ubiquitous open-source web server that powers a significant portion of the internet's websites. These newly disclosed flaws, collectively rated with a critical severity of 9.2 on the CVSS 4.0 scale, could permit unauthenticated attackers to achieve remote code execution (RCE) on affected servers, granting them full control.

The vulnerabilities were independently reported by several security research teams, including Google Cloud-owned Wiz, Ikotas Labs, Metnir, Rapid7, and Xint. Notably, some of these researchers leveraged AI models, in collaboration with OpenAI, to identify these critical security weaknesses. The disclosures highlight the increasing role of artificial intelligence in both discovering and potentially exploiting software vulnerabilities.

Wiz researchers detailed two specific vulnerabilities: CVE-2026-123633, a use-after-free flaw dubbed 'Drainbreak' by Metnir, which can lead to memory corruption, and CVE-2026-123007, an integer underflow issue within NGINX's HTTP/2 flow control. Both of these vulnerabilities, which affect NGINX versions dating back to 2015 and 2020 respectively, require HTTP/2 to be enabled for successful exploitation.

F5, the maintainer of NGINX, has already released patches for these issues in NGINX Open Source versions 1.30.7 (stable) and 1.31.9 (mainline). Administrators unable to apply these updates immediately are strongly advised to disable HTTP/2 on their NGINX servers as a temporary mitigation measure.

In addition to the HTTP/2 related flaws, Xint's research, collectively named 'NGINX2Root,' uncovered three further vulnerabilities. CVE-2026-123634 affects specific configurations involving the resolver, proxy_pass, and mirror directives, potentially leading to memory handling exploits. CVE-2026-121810 targets NGINX servers using HTTP/3 (QUIC) with upstream proxy connections, offering another path to RCE. Lastly, CVE-2026-113425 impacts NGINX Mail Proxy deployments using specific proxy and auth_http directives, presenting a distinct attack vector.

These three additional vulnerabilities are also addressed in the NGINX 1.31.9 release. While Xint confirmed that proven exploits exist for all five disclosed vulnerabilities, the precise configuration requirements and affected versions vary. Although no evidence of active exploitation in the wild was reported at the time of disclosure, the age of some of these flaws means that exploitation cannot be ruled out.

The disclosure of these critical vulnerabilities underscores the ongoing security challenges associated with widely deployed open-source software. NGINX's extensive use makes it a prime target for attackers, and the discovery of multiple RCE flaws highlights the need for continuous vigilance and prompt patching by system administrators worldwide.

Synthesized by Vypr AI