NATHANIELPETTUSCYBER INTELLIGENCE
NEW POST DAILY
BACK TO CYBER NEWS, BLOG & ANALYSIS

Critical NGINX HTTP/2 flaws CVE-2026-123007 and CVE-2026-123633 require urgent patching

F5 has patched two critical NGINX HTTP/2 memory-safety flaws that can permit unauthenticated code execution in affected configurations. Rapid7 demonstrated CVE-2026-123007 in a lab, while CERT-EU says no exploitation in the wild has been reported.

By Nathaniel PettusCybersecurity, Linux/UNIX, OSINT, and privacy-focused analysis

What happened?

Confirmed facts: On November 18, 2026, F5 published out-of-band updates for NGINX Open Source and NGINX Plus addressing multiple HTTP/2 vulnerabilities. The most serious are CVE-2026-123007, a flow-control calculation flaw that can cause a heap buffer overflow, and CVE-2026-123633, a use-after-free involving HTTP/2 connections and upstream proxying. F5 rates both Critical. Rapid7, which reported CVE-2026-123007, says it demonstrated unauthenticated remote code execution against a vulnerable NGINX build in a controlled lab. CERT-EU says no exploitation in the wild has been reported. Important limits: this is not evidence that every NGINX server is vulnerable or compromised. Exposure depends on the installed version and configuration. Public guidance identifies HTTP/2 as a prerequisite; CVE-2026-123633 also involves proxying to a backend and pressure on available worker connections. F5 says CVE-2026-123007 may lead to code execution if address-space layout randomization is disabled or bypassed. No victim breach or data theft has been confirmed as part of these disclosures.

How the technology works

Safe defensive explanation: HTTP/2 uses flow-control accounting to limit how much data can be sent at once. In CVE-2026-123007, specially formed traffic can make that accounting go below its intended range, leading NGINX to handle memory incorrectly and potentially corrupt a worker process. CVE-2026-123633 is a separate lifetime-management error: under specific proxying and connection-pressure conditions, NGINX may continue using memory after it has been released. Either condition can crash a worker; under favorable conditions, memory corruption may be turned into code execution. This article intentionally omits packet sequences, proof-of-concept details and exploitation instructions.

Who is affected?

Who is affected: operators of internet-facing NGINX Open Source or NGINX Plus systems should verify their exact running version and whether HTTP/2 is enabled. Rapid7 lists NGINX Open Source 1.31.0 through 1.31.8 and 1.9.5 through 1.30.6 as vulnerable to CVE-2026-123007, with fixes in 1.31.9 and 1.30.7. F5 lists fixed NGINX Plus releases in its advisory and notes that some older branches do not have a fix. CVE-2026-123633 requires HTTP/2 client connections and proxying to an upstream server, with the flaw triggered when worker connections are scarce. Products that bundle NGINX may have different release schedules, so use the supplier's advisory rather than assuming the package name alone proves exposure. A vulnerable version identifies risk, not proof of compromise.

What should you do?

What should you do? 1. Inventory every public NGINX endpoint, record the actual running binary version, and check whether its listeners enable HTTP/2. 2. Update NGINX Open Source to 1.30.7 or 1.31.9, or apply the fixed NGINX Plus release specified by F5 for your branch. Restart the service and verify the running process—not only the package database—reports the expected version. 3. If an update cannot be deployed immediately, CERT-EU recommends temporarily disabling HTTP/2 on exposed listeners, then testing that the change took effect. 4. Review worker crashes, unexpected child processes, privilege changes and unusual outbound connections around the exposure window. A worker crash alone is not a reliable indicator of successful code execution; correlate it with process and network telemetry. 5. Preserve relevant logs and images before rebuilding a system if suspicious activity appears. Rotate credentials or secrets from a known-clean device when evidence indicates possible compromise, and investigate downstream systems those credentials could reach.

OPINION

My analysis

My Analysis: This is the kind of vulnerability that deserves fast action without an exaggerated claim that 'NGINX has been hacked.' A lab-demonstrated path to unauthenticated code execution on a widely deployed edge server is serious, but configuration prerequisites and the absence of reported in-the-wild exploitation are equally important facts. A privacy-first response starts with precise asset inventory, rapid patching and least privilege. Organizations should minimize the secrets and backend authority available to an internet-facing proxy, so one memory-safety failure cannot become a broad data-access event. Monitoring should also be proportionate: collect narrowly useful signals such as software versions, worker crashes, unexpected process creation, privilege changes and unusual egress, retain them for a defensible period, and avoid turning an urgent patch cycle into an excuse for indefinite collection of employee or customer activity.

Why this matters

Cyber incidents often sound distant or overly technical. The important question is whether the same weakness, behavior, surveillance power, or exposure exists in systems you use. Facts and opinion are separated here so you can judge both clearly.