← Back to Blog

NGINX HTTP/2 Vulnerabilities: RCE Risk, Affected Versions and Protective Controls

Two critical NGINX HTTP/2 vulnerabilities pose remote code execution risks. Understand which systems are affected, how existing security controls can prevent exploitation or limit its impact, and how to validate protection while remediation is underway. Two vulnerabilities disclosed in NGINX on November 18, 2026, could allow unauthenticated attackers to execute code remotely by exploiting weaknesses in HTTP/2 processing.

CVE-2026-123007 involves an integer overflow that can lead to memory corruption, while CVE-2026-123633 involves a use-after-free condition during HTTP/2 connection handling. Both affect a widely deployed technology used to deliver websites, APIs and enterprise applications.

Although NGINX has released security updates, deploying them across enterprise environments takes time. During this period, organizations need to identify affected installations and determine whether their existing security controls can provide protection.

Vulnerability Intelligence

Primary CVE: CVE-2026-123007

Related CVE: CVE-2026-123633


AttributeDetails
ProductNGINX Open Source / NGINX Plus
VendorF5 / NGINX
Asset ClassesServer, Virtual Machine, Container; Network / Security Appliance where embedded
Technology CategoriesWeb Server, Reverse Proxy, API Gateway, Ingress Controller
ImpactPotential Remote Code Execution, Service Disruption
Vulnerability TypeHTTP/2 Memory Corruption
CWEInteger Overflow / Wraparound for primary CVE
CVSS9.2 – Critical
EPSSNot yet available
CISA KEVNot listed
Attack VectorNetwork
AuthenticationNot required
User InteractionNone
Exploitation PrerequisiteAffected NGINX version with HTTP/2 enabled and reachable
Exploit AvailabilityCurrent public exploit status to be verified
Exploitation in the WildNo confirmed status established for this article
Affected VersionsRefer to the applicable NGINX security advisory
Patch AvailableYes
Vendor WorkaroundDisable HTTP/2
Compensating ControlsCDN protocol termination and origin access restrictions (conditional)
Mitigation ValidationConfiguration, protocol and reachability checks
PublishedNovember 18, 2026
Last ReviewedNovember 19, 2026

1. Understanding the Vulnerabilities

NGINX is widely deployed as a web server, reverse proxy and application traffic gateway. Its HTTP/2 implementation manages how clients and servers exchange data, and the newly disclosed vulnerabilities affect how certain protocol operations are processed.

CVE-2026-123007 arises from an integer overflow in HTTP/2 flow-control processing that can lead to memory corruption. Meanwhile, CVE-2026-123633 involves a use-after-free condition during HTTP/2 connection handling in reverse-proxy operations.

If successfully exploited, these weaknesses could allow remote code execution with the privileges of the NGINX worker process. Depending on the deployment, an attacker could potentially compromise hosted applications, disrupt services or access information available to the affected process.

The vulnerabilities are relevant to affected NGINX installations running on physical servers, virtual machines, containers or appliances where the vulnerable software is embedded. However, exploitation requires access to the affected HTTP/2 processing path, while CVE-2026-123633 also has specific reverse-proxy prerequisites.

As a result, organizations should assess both the installed software version and the relevant configuration when determining which systems require immediate attention.

2. Which Existing Security Controls Can Protect You?

Organizations may already have security controls that protect against these NGINX vulnerabilities or limit the consequences of successful exploitation. However, it is important to distinguish between controls that prevent exploitation and those that reduce its impact.

For these vulnerabilities, disabling HTTP/2 on affected NGINX listeners is the most direct temporary mitigation. Alternatively, a CDN or reverse proxy that terminates HTTP/2 and communicates with NGINX exclusively over HTTP/1.1 can prevent the vulnerable processing path from being reached, provided direct access to the origin is restricted.

Existing NGINX and operating-system hardening controls can provide additional protection by limiting what an attacker could do following a successful compromise.

Security ControlEffectivenessProtection Offered
Disable HTTP/2 on NGINXPrevents exploitationRemoves the vulnerable HTTP/2 processing path, preventing exploitation of both CVEs through affected listeners.
CDN / Reverse ProxyPrevents exploitation (conditional)Terminates HTTP/2 and forwards HTTP/1.1 to NGINX, preventing malicious HTTP/2 frames from reaching the origin through that path.
Origin Firewall / Access ControlSupporting protectionPrevents attackers from bypassing the CDN or reverse proxy to reach an HTTP/2-enabled origin directly.
Run NGINX Workers as Non-RootLimits impactRestricts the privileges available to an attacker following successful code execution.
Restrict Filesystem PermissionsLimits impactReduces access to sensitive files, application data and configurations.
SELinux / AppArmor ConfinementLimits impactRestricts the actions a compromised NGINX process can perform, depending on the enforced security policy.
Restrict Outbound Network AccessLimits impactCan constrain attacker communication, lateral movement and data exfiltration.

Disabling HTTP/2 across all affected NGINX listeners prevents exploitation of these two HTTP/2 vulnerabilities through the vulnerable protocol implementation, even before the software is patched. However, it does not remove the underlying flaws, so upgrading to a fixed release remains necessary.

Similarly, CIS-aligned hardening measures such as non-root execution, restricted filesystem permissions and operating-system confinement can limit the damage following a compromise, but they do not independently prevent exploitation.

The practical distinction is straightforward: protocol-level controls can prevent exploitation, while system hardening can limit its consequences. Both contribute to protection during the remediation window.

3. Remediation and Temporary Mitigation

NGINX has released fixes for the affected versions, making upgrading to the applicable security release the recommended remediation.

ProductFixed Version
NGINX Open Source — Stable1.30.7
NGINX Open Source — Mainline1.31.9
NGINX PlusApplicable vendor-fixed release

Organizations using distribution-packaged versions should also check for vendor-backported security fixes, as the displayed software version may not fully reflect the installed security patches.

Where immediate patching is not possible, disabling HTTP/2 on vulnerable listeners provides a temporary mitigation for the two vulnerabilities.

For NGINX versions supporting the standalone HTTP/2 directive:

server {
    listen 443 ssl;
    http2 off;

    server_name example.com;
}

For older configurations, the http2 parameter can be removed from the listen directive. Alternatively, organizations using a CDN may enforce HTTP/1.1 communication between the CDN and NGINX while restricting direct origin access. This can preserve HTTP/2 support for website visitors, provided that all relevant origin paths are protected.

Before implementing either approach, application dependencies should be assessed. For example, services using gRPC may require HTTP/2 and could therefore need a different mitigation strategy.

Although these measures can reduce exposure during the remediation window, they address only the two HTTP/2 vulnerabilities discussed here. Applying the vendor security update remains necessary to permanently resolve the underlying weaknesses.

4. How Do You Verify Protection?

Implementing a configuration change does not necessarily establish that the vulnerable attack path has been removed. Consequently, security teams should validate both the effective NGINX configuration and the network paths through which the affected service can be accessed.

First, inspect the installed version and effective configuration:

nginx -v
nginx -T

Next, where direct origin testing is authorized, verify the protocol negotiated with NGINX:

curl -k --http2 -I \
  --resolve example.com:443:ORIGIN_IP \
  https://example.com/

If HTTP/2 has been disabled, the origin should negotiate HTTP/1.1. However, this confirms only the tested connection, so other relevant listeners, hostnames and network routes must also be assessed.

The following evidence can help establish whether the mitigation is operating as intended:

Validation RequirementExpected Evidence
Affected assets identifiedAsset inventory, product and version
HTTP/2 exposure assessedConfiguration and reachability
CDN protection validatedEdge-to-origin protocol configuration
Origin access restrictedFirewall policy and connectivity tests
Temporary mitigation implementedApproved change and post-change validation
Permanent remediation completedFixed build or vendor backport

Together, these checks help organizations distinguish systems that remain vulnerable and reachable from those protected by validated controls or permanently remediated.

SPOG.AI | Connected Risk Intelligence

SPOG.AI connects vulnerability intelligence with asset context, security configurations, control validation and remediation. This helps organizations understand newly disclosed vulnerabilities, determine whether existing defenses provide protection and verify that remediation has been effective.

Know the vulnerability. Understand the protection. Validate the outcome.

References

Assessment date: November 19, 2026. Dynamic intelligence fields, affected versions and fixed-release information require verification against authoritative sources before publication.

ARTICLES YOU MIGHT BE INTERESTED IN