Skip to content
Markets Editorial

Amazon Says AI Agents Found Zero-Day Flaw in 11 Minutes

Amazon reports AI agents found a MONAI Label vulnerability in about 11 minutes and helped attackers gain root access to a university server in East Asia.

Amazon Says AI Agents Found Zero-Day Flaw in 11 Minutes
Ishant Sharma
Technology / Science Contributor
Published: 6 min read
Reported and edited under BigBreakingWire editorial standards.

Amazon Threat Intelligence reported on November 20, 2026, that suspected state-linked cyber operators used an AI-powered penetration-testing framework to discover a previously unknown software vulnerability and obtain root-level access to a university server.

The reported campaign took place between September 8 and 10 and primarily targeted universities in East Asia outside mainland China.

According to Amazon, the attackers used Cairn, an open-source penetration-testing framework capable of coordinating multiple AI agents to automate security testing and intrusion-related activities.

Advertisement

The findings highlight the potential for AI-assisted tools to accelerate vulnerability discovery and exploitation. However, Amazon’s report also described unsuccessful intrusion attempts, showing that the framework was not consistently effective.

AI Agents Exploit Medical Imaging Software

Amazon said it identified the campaign after the operators left their Cairn instance exposed to the public internet without authentication.

This allowed researchers to examine the framework’s activities and reconstruct parts of the reported attack sequence.

The attackers reportedly instructed Cairn agents to obtain shell access to targeted systems. The agents then identified a path traversal vulnerability in MONAI Label, an open-source medical imaging application used in research and healthcare-related environments.

A path traversal vulnerability can allow software to access or modify files outside the intended directory when file paths are not properly restricted.

Amazon reported that the agents progressed from mapping the application’s interface to exploiting the vulnerability to write arbitrary files on the targeted server in approximately 11 minutes.

During a later session, the attackers used the same weakness to establish a reverse shell with root privileges, giving them highly privileged access to the university server.

Reported MONAI Label Attack Sequence

DiscoveryAI agents mapped the application interface
~11 minutesPath traversal flaw exploited for arbitrary file writing
Later sessionReverse shell established with root privileges
Sequence reported by Amazon Threat Intelligence; timings do not represent a general benchmark for AI cyberattacks.

The operators subsequently directed the agents towards another university running the same software.

Maintainers of MONAI Label published a security advisory on November 18 following discovery of the vulnerability.

Not every attempt succeeded. Amazon described several failed intrusion efforts, including an approximately 16-hour attempt against a Redis server.

The contrast between rapid exploitation of one application and prolonged failure against another is important when evaluating the capabilities demonstrated by the framework.

FBI Warning Adds to Concerns About AI-Assisted Intrusions

The findings followed a November 13 FBI FLASH advisory concerning ARTEX, a separate AI-driven offensive framework reportedly used by cybercriminals and state-aligned operators since at least August 2026.

The FBI associated ARTEX with activity targeting financial institutions, government agencies, defence organisations and healthcare systems.

According to the account, the bureau warned that the framework could accelerate some intrusion operations from initial targeting to privileged access in less than three hours.

Advertisement

The FBI subsequently clarified that it had not issued another warning after its November 13 advisory.

The two frameworks should not be treated as the same system or as evidence of a single coordinated campaign.

US Outages Remain Unlinked to Reported Attacks

The warnings emerged alongside technology disruptions affecting US companies and government systems during November.

Reported incidents included a Microsoft service disruption on November 12, an FAA ground stop on November 13, banking access problems on November 16 and additional service disruptions on November 17.

US officials have not established a common cause linking these events. There is no confirmed evidence that the outages resulted from coordinated cyberattacks.

Microsoft reported no indication of malicious activity associated with its disruption, while the NSA attributed its incident to scheduled maintenance.

Senator Mark Warner warned that faster AI-assisted vulnerability discovery could put pressure on organisations to deploy patches before they had been sufficiently tested. He also emphasised that the reported outages had not been established as cyberattacks.

Palo Alto Networks CEO Nikesh Arora expressed concern about the concentration of disruptions, but that assessment does not establish a shared malicious cause.

Amazon’s findings demonstrate a reported instance of AI-assisted vulnerability exploitation. They do not prove that unrelated US technology outages were caused by the same tools, operators or methods. For continuing coverage of AI-powered cyberattack frameworks and cybersecurity risks, see BigBreakingWire’s Technology coverage.

BBW Analysis

What It Means

The incident illustrates how AI-assisted tools may reduce the time required to identify and exploit certain software vulnerabilities. For universities, healthcare researchers and technology providers, that increases the importance of prompt vulnerability management and monitoring of exposed systems.

The unsuccessful attacks are equally relevant. They show that reported AI capabilities should not be generalised into claims of universal or automatic compromise.

The separate US technology outages also require careful reporting. A concentration of disruptions does not establish that they share a cause, and the available information contains explicit statements contradicting some suggestions of malicious activity.

What to Watch

  • Further MONAI Label security guidance.
  • Additional technical findings from Amazon.
  • Follow-up FBI advisories about ARTEX.
  • Evidence of AI-assisted exploitation in other sectors.
  • Confirmed causes of separate US technology disruptions.

Market Impact

  • MONAI Label exploitation: Approximately 11 minutes — reported rapid vulnerability exploitation.
  • Root access: Obtained in a later session — serious reported compromise.
  • Redis attempt: Approximately 16 hours without success — capability limitation.
  • ARTEX intrusion window: Under three hours, according to FBI warning — potential acceleration.
  • Financial losses: Not disclosed — no measurable market impact established.

What did Amazon discover about AI-powered cyberattacks?

Amazon reported that suspected state-linked operators used the Cairn framework to coordinate AI agents during a campaign against universities in East Asia. The agents reportedly identified and exploited a previously unknown vulnerability in MONAI Label, contributing to privileged access on a university server.

How quickly did the AI agents exploit the vulnerability?

According to Amazon, the agents progressed from mapping the MONAI Label application interface to exploiting a path traversal weakness for arbitrary file writing in approximately 11 minutes. A later session reportedly established a reverse shell with root privileges.

What is MONAI Label?

MONAI Label is an open-source medical imaging application used in research and healthcare-related environments. Amazon reported that attackers exploited a path traversal vulnerability in the software. Its maintainers published a security advisory on November 18, according to the report.

Is ARTEX the same as Cairn?

No. The reporting describes Cairn as the framework examined in Amazon’s investigation and ARTEX as a separate framework discussed in an FBI advisory dated November 13. Both raise concerns about AI-assisted intrusions, but they should not be presented as one campaign.

Were recent US technology outages caused by cyberattacks?

No common malicious cause has been established in the reporting. Microsoft reported no evidence of malicious activity in its disruption, while the NSA attributed its incident to scheduled maintenance. Officials have not confirmed that the broader series of outages resulted from coordinated attacks.

Sources

  1. Amazon
  2. FBI
  3. Microsoft
  4. Palo Alto Networks

Sourcing and verification are governed by our editorial policy.

Advertisement
Ishant Sharma
Technology / Science Contributor

Ishant Sharma

Ishant Sharma is a Technology and Science Contributor at BigBreakingWire, covering artificial intelligence, software, semiconductors, gadgets, and emerging technologies.

Last reviewed by Ishant Sharma on November 21, 2026

Disclaimer: BigBreakingWire provides news and informational content for general educational purposes only. The information presented on this website does not constitute financial, investment, tax, or legal advice. Readers should consult qualified professionals before making any financial decisions. BigBreakingWire, its authors, and editors are not responsible for any financial losses or damages arising from the use of information on this site.

Discussion

Share your take — react and comment in seconds, no login needed.

0 comments

Loading discussion…

Be civil — abusive or promotional comments get removed. No login required.

Most Read

Install app