Back to all lessons
Awareness Lessons
3 days ago

AI Uncovers Critical NGINX Flaws Hiding Since 2015 — 225M Sites at Risk

Written by AI from the linked story. Not reviewed by an editor. How this works

AI-assisted research exposed ten vulnerabilities in NGINX, two of which carry a CVSS score of 9.2 and enable remote code execution on servers with HTTP/2 enabled — a configuration common across millions of deployments. These flaws existed undetected for nearly a decade, underscoring how legacy code in widely adopted open-source software can harbour critical risks that traditional manual review misses. The incident highlights the accelerating threat landscape: if AI can find these bugs defensively, adversaries can use the same tools offensively. Organizations running internet-facing NGINX instances must treat emergency patching as a core operational capability, not an exception. Waiting for routine patch cycles is no longer acceptable when a single unpatched server can expose an entire infrastructure.

Tactical Insight

Immediate actions

  • Apply F5's emergency patches for NGINX immediately, prioritising any internet-facing instances with HTTP/2 enabled
  • Audit your asset inventory to identify every NGINX deployment, including those embedded in reverse proxies, load balancers, and containerised workloads
  • Disable HTTP/2 on NGINX instances that cannot be patched immediately as a temporary risk-reduction measure

Long-term improvements

  • Establish a formal emergency patching procedure with defined SLAs for critical (CVSS 9.0+) vulnerabilities that bypasses normal change-control timelines
  • Maintain a continuously updated software bill of materials (SBOM) so open-source components like NGINX are tracked and alerted on when new CVEs are published
  • Integrate automated vulnerability scanning into CI/CD pipelines so newly deployed images are checked against known CVEs before reaching production

Detection measures

  • Deploy web application firewall (WAF) rules targeting HTTP/2 exploitation patterns as a compensating control while patches are applied
  • Monitor NGINX access and error logs for anomalous HTTP/2 request patterns that may indicate exploitation attempts
  • Subscribe to vendor security advisories and threat-intelligence feeds so zero-day and emergency disclosures trigger immediate internal alerts

Back this up with a control

All controls and vendors

Training helps people spot this. These controls make the attack fail even on the day someone does not.