Papercut AI Swarm Attack Heralds Changes for Cyber Kill Chain

A likely Russian-speaking actor used hundreds of AI agents to scan for and exploit two vulnerabilities in PaperCut NG/MF instances, targeting at least 440 exposed instances across 395 organizations in 48 countries; attackers achieved RCE and then rapid Active Directory domain admin compromises (reportedly compromising at least 11 organizations) and demonstrated how agentic LLM tooling accelerates the attack lifecycle, with discussion of supply-chain and LLM hijacking risks and recommendations to rely on basic security hygiene.