Crook used three open source agents to break into a Fortune 500 hospitality company, a major US airline and 25+ other orgs

What happened

Security researchers have detailed how a single threat actor used three open-source AI agents to compromise more than 25 organisations, including a Fortune 500 hospitality company and a major US airline. According to reporting by The Register, the attacker chained together off-the-shelf autonomous agent frameworks to scan, probe and exploit vulnerable systems with minimal manual intervention.

The most striking detail is cost: the operator's average AI spend per completed scan was just $25, underscoring how cheaply agentic tooling can now be repurposed for malicious reconnaissance and intrusion at scale.

Why it matters

This is fundamentally a story about what agentic AI now makes possible — for better or worse. The same capabilities that let enterprises automate complex, multi-step workflows cheaply and autonomously are equally available to bad actors, with no enterprise budget or bespoke tooling required. Open-source agents lower the barrier to running sophisticated, semi-autonomous operations across dozens of targets simultaneously.

For leaders driving AI and digital transformation, the incident is a concrete signal that agentic AI has crossed a cost and capability threshold where it can be weaponised at commodity prices. Any organisation deploying AI agents internally — for service automation, IT operations or customer-facing workflows — now has to reckon with an adversary who can use the very same class of tools against them, often more cheaply than traditional attack methods.

By the numbers

• $25 average AI cost per completed scan for the attacker, according to researcher findings cited by The Register.
• 25+ organisations confirmed compromised, spanning hospitality, aviation and other sectors.
• Three open-source AI agent frameworks were chained together to conduct the intrusions.
• One Fortune 500 hospitality company and one major US airline were named among the victims.

The Renascence take

Most coverage of this incident will focus on the security failure. The more useful lens for experience and transformation leaders is economic: agentic AI has quietly collapsed the cost of running complex, adaptive operations — whether that operation is a customer service workflow or an intrusion campaign. The behavioral lesson is the same one that applies to any automation: once a capability becomes cheap and repeatable, volume follows, and defences built for rare, bespoke attacks won't hold.

Cheap autonomy cuts both ways — the same agentic patterns letting a hospitality brand automate guest services are letting an attacker automate reconnaissance for $25 a scan. Operators rolling out AI agents should assume their own attack surface has just grown to match their automation ambitions, and treat agent governance, monitoring and access controls as core to the experience build, not a bolt-on security afterthought.