Check Point issued emergency hotfixes on November 12 for CVE-2026-123039, a local privilege escalation in Gaia OS that allows authenticated users to execute commands as root. The vulnerability resides in the DynamicCLI daemon, which listens on a world-writable socket at /tmp/dynamic_socket without authenticating connecting processes. While the zero-day provides root access, the initial foothold often comes from PHP2Shell vulnerabilities (CVE-2026-109618 and CVE-2026-109619), which were patched in October but remain active in exploited chains.

The Attack Chain and Wiper Deployment

Attackers gained entry via the Mobile Access portal using PHP2Shell flaws, then escalated privileges using the Check Point zero-day. This chain led to the deployment of webshells, the TSGA backdoor, and eventually the destroy_v6 data wiper. The wiper stage, observed in campaigns against Israeli organizations beginning October 30, explicitly rejected ransom negotiations. A second wave on November 2 targeted companies in multiple sectors but skipped the destructive step, suggesting a shift in attacker objectives or campaign phases.

Critical Patch Details and Attribution

CISA added the Check Point CVE to its Known Exploited Vulnerabilities (KEV) catalog on November 13, with a federal due date of November 16. Check Point advisory sk1000362 lists specific Jumbo Hotfix Accumulator takes required for remediation: R82.20 Take 14, R82.10 Take 52, R82 Take 132, and R81.20 Take 174. Check Point tentatively tied the tooling to PhiliKit, which ESET has linked to a Chinese threat cluster, though the attacks could not be attributed with full confidence.

Chrome and GPUStack Join the KEV List

CISA expanded the KEV list this week with two additional critical infrastructure items. Chrome CVE-2026-125350, a use-after-free in the browser's XML component, was added November 10 with a November 24 due date. Google confirmed an exploit exists in the wild, and the fix is available in version 157.0.8103.50. Separately, GPUStack CVE-2026-102895, an authentication bypass in the open-source GPU cluster manager, was listed on November 9. Prior to version 2.2.0, the manager treats loopback requests as administrator privileges by default, a flaw exploited against version 2.1.1 deployments.

AI Tools on Both Sides of the Bug Hunt

The security landscape this week highlighted AI's dual role in offense and defense. Anthropic's OSS Scanner and Google Cloud Vulnerability Research independently discovered a critical HTTP/2 remote code execution flaw (CVE-2026-119618) in Apache httpd 2.4.55 through 2.4.70. Conversely, Amazon Threat Intelligence captured AI agents using the ARTEX framework to steal credentials, feeding stolen keys back into their own model gateways to fund further attacks. This demonstrates how AI agents can autonomously scale offensive operations, though the underlying techniques remain traditional.

Key Takeaways

  • Install Check Point Jumbo Hotfix Accumulator takes immediately, prioritizing gateways exposing Mobile Access portals. If patching is delayed, stop the dynamic_server process to mitigate risk.
  • Upgrade Apache httpd to version 2.4.71 to address the critical concurrency flaw in mod_http2, which allows for remote code execution.
  • Audit AI-built applications for public exposure. Research by Claroc found that 94.7% of tested Grok Build apps were accessible without sign-in, often due to default 'Anyone with the link' settings.
  • Monitor for indicators of compromise including session files containing 'upload_progress_pwn', unexpected CPconfigdaemon processes, and unauthorized SSH connections from gateways to internal ESXi hosts.
  • Update Chrome to 157.0.8103.50 and GPUStack to 2.2.0 to address newly listed KEV vulnerabilities with imminent federal deadlines.

The Bottom Line

The real story isn't the AI finding the bugs, but the basic hygiene failures that let attackers exploit them. Patch your gateways, lock down your publish settings, and stop assuming local sockets are safe.