NGINX HTTP/2 flaws CVE-2026-123633 and CVE-2026-123007 could allow remote code execution

Two critical vulnerabilities in NGINX, disclosed on 18 November 2026, could allow remote attackers to run code on vulnerable web servers without logging in, according to Cyber Security News. The flaws, tracked as CVE-2026-123633 and CVE-2026-123007, both carry a critical score of 9.2 under the CVSS 4.0 severity scale, and both require NGINX to accept HTTP/2 connections from clients.
The flaws sit in NGINX itself, not in the websites or applications behind it. Because NGINX often handles HTTPS traffic and routes requests to several applications, a single vulnerable proxy can put multiple services at risk even when their own code is secure. A successful attack could expose sensitive data, alter website content, or interrupt service, depending on what the NGINX process can reach.
What the two flaws are
In its advisory for CVE-2026-123633, F5 describes a use-after-free bug, in which a worker process can reuse memory after it has been released. The issue affects setups where NGINX accepts HTTP/2 client connections and forwards requests to a backend, and where the worker is running low on available connections. The flaw can crash the worker or, under certain conditions, let an attacker run code.
CVE-2026-123007 has a wider reach. It affects NGINX Open Source releases dating back to version 1.9.5, which shipped in 2015. In its disclosure report, Wiz said it demonstrated code execution on an NGINX server serving ordinary static files, with no backend application or third-party modules. A basic website should therefore be part of the patch review, not just API gateways and reverse proxies.
Wiz Research named Ikotas Labs, Metnir, Rapid7 and Xint as independent finders across the two disclosures. The findings show remote code execution is possible under the reported conditions, but do not establish that attackers have used either flaw against production systems.
Which versions are affected
For CVE-2026-123633, NGINX Open Source versions 1.19.7 to 1.30.6 and 1.31.0 to 1.31.8 are affected. For CVE-2026-123007, the affected range is wider: NGINX Open Source 1.9.5 to 1.30.6 and 1.31.0 to 1.31.8. The fixes for both flaws are listed for NGINX Open Source 1.31.9 on the mainline branch and 1.30.7 on the stable branch. For NGINX Plus, the minimum fixed builds differ by CVE, so check both F5 advisories and choose a build that covers both; R-series releases R34 to R36 have no fix listed and should be upgraded to a 37.x build.
These are separate from the NGINX flaws F5 disclosed in October. An organisation that applied those fixes may still need another update: F5 lists NGINX Open Source 1.31.8 as affected by CVE-2026-123633, so running a build newer than October’s security release does not mean it contains the November fixes. The two November flaws are also distinct from the NGINX HTTP/2 proxy and gRPC vulnerability patched in June, which depended on different configuration options.
How widely used is NGINX?
Netcraft counted more than 225 million sites using NGINX in its September 2026 web server survey. That figure shows how widely the software is used, not how many sites are vulnerable: exposure depends on the installed version and on whether the server handles HTTP/2. Multiple sites can also share the same NGINX deployment, so a site count is not a count of affected servers. Even so, the scale of NGINX use is why the disclosure matters well beyond large enterprises.
What administrators should do now
Start with internet-facing servers, then check internal proxies and static sites. For NGINX installed through a Linux distribution or bundled with another product, check the supplier’s security notice for the package or image that contains both fixes. After updating, verify the version actually running on every instance: a configuration reload alone does not replace the running NGINX binary, so restart the service or roll out replacement containers as appropriate.
For this update, administrators should check whether NGINX accepts HTTP/2 from clients; checking only the protocol used to reach a backend can miss the client-facing exposure. If a fixed build is not yet available, temporarily turn off client-facing HTTP/2 where the service can tolerate it, then test that the endpoint no longer accepts HTTP/2 and that websites and APIs still work. Disabling HTTP/2 reduces exposure but does not fix the flaws, so install the security update before turning the protocol back on.
Frequently asked questions
Do the NGINX flaws need a login to exploit? No. Both flaws can be reached by a remote attacker without logging in, provided NGINX accepts HTTP/2 connections from clients.
Which NGINX versions fix CVE-2026-123633 and CVE-2026-123007? NGINX Open Source 1.30.7 on the stable branch and 1.31.9 on the mainline branch fix both flaws. NGINX Plus fixed builds differ by CVE, so check both F5 advisories.
Is there evidence the flaws are being exploited? No. The research findings show remote code execution is possible under the reported conditions, but they do not establish that attackers have used either flaw against production systems.
Sources: Cyber Security News; F5 security advisory for CVE-2026-123633.