Autonomous AI Agent Exploits Ticketing Software Vulnerabilities in Security Nonprofit Attack

The Dutch Institute for Vulnerability Disclosure disclosed that attackers leveraged two zero-day flaws in the Zammad ticketing system to breach its network using an autonomous AI agent. The vulnerabilities enabled session hijacking, remote code execution, and privilege escalation to root level, with the AI-driven attack completing data exfiltration in seconds. Zammad users are advised to upgrade to version 7 or take systems offline immediately.

The attack represents a significant shift in breach methodology. Rather than following a predetermined script, the autonomous AI system made real-time decisions about which systems to target and how to exploit them, completing the entire compromise—from initial access through data theft—in seconds. This speed and autonomy made traditional security monitoring difficult, though the attacker's detailed logging of its reasoning actually aided forensic investigation.

Zammad serves a diverse customer base spanning retail, nonprofit, and technology sectors. The vulnerability chain is particularly concerning because it requires no user interaction; attackers can move from unauthenticated access to complete system control through automated exploitation. The fact that network segmentation prevented lateral movement underscores how infrastructure design can limit damage even when perimeter defenses fail.