AI-Powered PLC Attacks, GitLab Exploit & npm Backdoors: Weekly Recap

A wave of AI-driven attacks and supply-chain compromises dominated this week's cybersecurity headlines, underscoring how quickly threat actors are weaponizing automation and trusted ecosystems. U.S. federal agencies issued an urgent advisory warning that threat actors are actively using AI to generate exploit scripts targeting internet-exposed Siemens S7 Series programmable logic controllers (PLCs) deployed across water treatment, energy, and manufacturing infrastructure. Attackers are leveraging legitimate scanning services like Censys and ZoomEye to identify poorly segmented or directly exposed controllers, then deploying AI-crafted payloads disguised as monitoring tools. The agencies stressed that exploitation could trigger process disruption, equipment damage, and cascading failures across interconnected industrial environments, confirming this is an active, not theoretical, threat.

In the enterprise software space, a critical code injection flaw in GitLab—tracked as CVE-2026-19478—came under active exploitation within days of disclosure, according to watchTowr researchers. The unauthenticated vulnerability carries a CVSS score of 9.4 and allows attackers to modify or delete publicly accessible GitLab projects without requiring credentials or user interaction, making it trivial to weaponize at scale. Organizations running self-hosted GitLab instances should prioritize patching immediately and audit project integrity for signs of tampering. Security teams can also use a port scanner to identify any inadvertently exposed GitLab endpoints that may not have been patched yet.

The supply-chain threat surface expanded further with the discovery of 14 trojanized npm packages masquerading as benign calendar and streak-tracking utilities. Once installed, these packages silently deploy RedC2 4.0, an AI-powered Linux backdoor capable of remote command execution and persistent access. Developers who recently pulled utility packages from npm should audit their dependency trees immediately and rotate any secrets that may have been exposed on affected hosts. A comprehensive privacy checkup can help identify lingering exposures, while a email breach checker can confirm whether developer credentials tied to compromised systems have surfaced in known leaks. As AI lowers the barrier for both vulnerability discovery and payload development, defenders must treat every exposed endpoint and third-party dependency as a potential initial access vector.