"We need to treat every internet-connected system as potentially vulnerable," Tomer Schwartz, Wiz’s chief technology officer of exposure management and AppSec, told The Register.
F5's emergency patch for open source NGINX
F5 on Wednesday published an out-of-band security notification addressing 10 CVEs in the open source NGINX codebase. The vendor, which maintains the open source version, released an update: if you run NGINX, update to 1.31.9 as soon as possible. NGINX powers about a third of all websites—roughly 225 million—handling HTTPS requests and routing incoming traffic to applications and APIs.
The two critical CVEs: CVE-2026-123633 and CVE-2026-123007
Two of the disclosed vulnerabilities are rated critical with a 9.2 CVSS score. Both allow unauthenticated remote code execution on NGINX servers that have HTTP/2 enabled. CVE-2026-123007 affects versions dating back to NGINX 1.9.5 from 2015. According to researchers from Wiz, Metnir, Ikotas Labs, Rapid7 and Xint, both are memory-safety issues in how NGINX handles HTTP/2 traffic: CVE-2026-123633 is a use-after-free in which a connection's memory is freed while a request still needs it; CVE-2026-123007 stems from incorrect flow-control accounting that leads to an invalid memory copy.
Rapid7 published a proof-of-concept demonstrating unauthenticated remote code execution using CVE-2026-123007 against a vulnerable linux/amd64 Docker image running nginx 1.31.8, underscoring the practical exploitability of these flaws when HTTP/2 is enabled.

Audit-ready is a season. It shouldn't be.
Evidence in spreadsheets, controls drifting between audits, frameworks multiplying on flat headcount. Nubivance runs continuous compliance on Rapid7 Cyber GRC - SOC 2, HIPAA, ISO 27001, PCI, CMMC.
End the scrambleXint's "nginx2root" chain and cloud implications
Xint reported three additional NGINX flaws and a separate Linux kernel vulnerability that can be chained with any of the five NGINX vulnerabilities to escalate to root on the host. The research team dubbed the collection "nginx2root." Xint published videos showing an attack that executes commands as the NGINX worker inside a container using CVE-2026-123007, then leverages the kernel bug to gain root on the host. The company said these chained attacks work on AWS, Microsoft Azure, Google Cloud Platform, and bare-metal servers.
Xint also described an AI-driven discovery workflow: one researcher found CVE-2026-123007 while preparing for ZeroDay Cloud 2026, and a fully automated Xint scan discovered four other issues "without human guidance."
AI-accelerated discovery: what researchers found and said
Several security teams—Wiz, Ikotas Labs, Metnir, Rapid7, Calif, and Xint—independently reported at least one of the newly disclosed CVEs to F5, with most acknowledging use of AI assistance. Schwartz and Wiz researcher Nir Ohfeld framed the wave of findings as part of a "new reality": "AI has changed not only which vulnerabilities can be found, but how quickly longstanding risks can become real threats." The Register notes one of the bugs had been hiding in code for at least 10 years.
The Register also placed these NGINX discoveries alongside other recent AI-aided finds: Java4Shell (found by Calif), PHP2Shell (Calif used AI to find and weaponize PHP flaws), and a critical Apache flaw under active exploitation. Schwartz warned that "AI is also accelerating vulnerability discovery faster than existing infrastructure was designed to absorb," leaving security teams with heavy manual work to locate vulnerable versions, test patches, deploy interim mitigations (such as disabling HTTP/2), and coordinate rollouts.
What this means for technologists, cloud providers, and security teams
- Technologists and security teams: the immediate operational task is clear in F5's advisory—upgrade to NGINX 1.31.9 and consider temporary mitigations (the companies cited disabling HTTP/2 as an interim step) while inventories and patch testing proceed.
- Cloud providers and platform operators: Xint's demonstrations show chained exploits that affect containers and hosts across AWS, Microsoft Azure, and Google Cloud Platform; operators must verify images and host kernels as part of patching and mitigation.
- Security teams and vulnerability managers: researchers warn of an accelerating cadence of AI-discovered flaws, increasing the burden of locating vulnerable versions and coordinating fixes even after vendors issue patches.
Schwartz struck a cautiously optimistic note: in the long run, he told The Register, "AI-powered vulnerability research will harden foundational software." For now, however, a steady stream of AI-discovered critical bugs—some dormant for a decade—means defenders must move quickly to apply F5's fixes, reassess configurations like HTTP/2, and reconcile the speed of discovery with the realities of deployment and testing.




