AI‑Driven Exploit Chains Two Zammad Flaws to Breach Dutch Security Lab in Seconds

An AI‑powered attack chain that leveraged two previously unknown vulnerabilities in the open‑source ticketing platform Zammad succeeded in gaining root access to the Dutch Institute for Vulnerability Disclosure (DIVD) in a matter of seconds, according to a report first published by Security Affairs.

The breach unfolded when the malicious agent exploited the zero‑day flaws to execute code with administrative privileges on DIVD's internal servers. Within moments, the AI script harvested sensitive data from the institute’s research repositories and then attempted to move laterally into adjacent services, including internal communication tools and version‑control systems.

DIVD, a nonprofit collective of volunteer security researchers dedicated to identifying and responsibly disclosing software weaknesses, detected the intrusion through its own monitoring infrastructure. Security personnel were able to isolate the compromised systems and terminate the AI‑driven process before the attacker could exfiltrate additional information or establish persistent footholds.

The incident highlights a growing concern among cybersecurity experts: the automation of exploit development and deployment using artificial intelligence. While AI has been employed to accelerate vulnerability research, its use by adversaries to chain together multiple zero‑days in real time represents a new level of operational speed and sophistication.

Zammad, a widely adopted help‑desk solution written in Ruby on Rails, has not previously been associated with critical remote‑code execution bugs. The two flaws, which remain undisclosed to the public pending coordinated remediation, allowed the attacker to bypass authentication checks and execute arbitrary commands on the host operating system. The rapid exploitation suggests the AI agent was able to identify and combine the weaknesses without human intervention.

Following the incident, DIVD’s incident‑response team coordinated with Zammad’s development team and relevant national CERTs to develop patches. The organization also issued an advisory to its members and other users of the ticketing system, urging immediate updates once the fixes become available.

Security analysts caution that the episode may signal a broader shift toward AI‑augmented threat actors capable of chaining multiple vulnerabilities across software supply chains. They recommend that organizations adopt deeper runtime monitoring, enforce strict privilege separation, and maintain up‑to‑date patch management processes to mitigate the risk of similar automated attacks in the future.