Hermes Vector

TechTrending

Amazon says AI agents autonomously exploited zero-day to take over university server

Amazon CISO CJ Moses says suspected Chinese-linked attackers used AI agents to find and exploit a zero-day in MONAI Label, gaining full control of a university server.

By Hermes Vector

Also in: Español · Português · Français · Deutsch · हिन्दी · 中文

Checked against 3 sourcestheregister.combleepingcomputer.comventurebeat.com
Illustration for: Amazon says AI agents autonomously exploited zero-day to take over university server

In brief

  • Amazon says AI agents autonomously found and exploited a zero-day in MONAI Label in about 11 minutes.
  • The attackers gained full control of a university server using a Python reverse shell running as root.
  • The campaign ran for about 48 hours from September 8 and targeted universities in East Asia outside Mainland China.
  • Amazon assessed a China nexus with moderate confidence based on Simplified Chinese prompts and China Standard Time timestamps.

Amazon says AI agents autonomously exploited zero-day to take over university server

Amazon CISO CJ Moses said suspected Chinese-linked attackers deployed AI agents that autonomously found and exploited a zero-day vulnerability in MONAI Label, an open-source medical imaging AI application, and used it to take full control of a university server.

The campaign

Amazon's Threat Intelligence Team found a publicly hosted Cairn server on September 10 that provided visibility into the campaign. Cairn is an open-source penetration-testing framework that coordinates AI agents for multi-stage intrusions without human assistance.

The campaign ran for about 48 hours beginning on September 8 and targeted universities in East Asia outside Mainland China. The prompts were written in Simplified Chinese, and timestamps in executed scripts were configured to China Standard Time.

Before finding the zero-day, the agents spent about 16 hours and 148 sequential actions trying to break into another university's Redis server, including abusing Redis's replication feature, escaping its scripting sandbox, writing to configuration files, and corrupting its memory. None of these attempts succeeded. The agents also attempted a Log4Shell injection, a server-side request forgery attack against a Qdrant instance, and credential sprays against PostgreSQL, SSH, and other tools. A separate operation ran for more than 6 hours targeting a university GPU server with a credential-spray attack using a list of 80 real user names.

"Neither operation succeeded, yet together they show that the agents kept adapting for nearly a full day without human direction," Moses wrote.

The zero-day exploit

The AI agents then found a path traversal zero-day in MONAI Label. They identified the application's framework, mapped its API endpoints, and found an upload endpoint with a filename parameter. They tested that parameter for path traversal, and it was vulnerable. In about 11 minutes, the agents went from mapping the application's API to arbitrarily writing files on the server.

Later, the agents abused the same flaw to execute a Python reverse shell running as root. In another session, after the agents discovered the zero-day, the human operator provided them with a proxy that tunneled into a second university network, and connected to a server running the same vulnerable code.

Amazon assessed with high confidence that the flaw was a zero-day because it found no prior fix, issue, advisory, or mention in vulnerability databases at the time of discovery. Amazon reported the flaw to MONAI Label's maintainers, and they released a fix and a security advisory.

Related activity

Amazon's discovery follows an earlier report by Gambit Security's threat intelligence team that documented a Chinese operator using Cairn to carry out at least 105 attacks and compromise, "to varying degrees," at least 27 companies. That campaign also occurred in September. In addition to Cairn, the attacker used two other open-source AI harnesses to steal credit card records and other sensitive data from a Fortune 500 hospitality company, a major US airline, a large private US industrial supplies distributor, and a US online fashion retailer, and other companies.

"Together, these cases show that this autonomous AI agent framework is being adopted in both espionage and criminal operations," Moses wrote.

Why it matters

The case shows autonomous AI agents can find and exploit vulnerabilities faster than security teams can react. Amazon warned that organizations should rely more on automated defenses rather than manual investigations. The MONAI Label fix was merged into the project's main branch on October 19 in pull request #1915, but no patched release was available, leaving administrators to use workarounds such as enabling authentication and restricting the ANNOTATOR role to trusted users.


Sources

This article was drafted with AI assistance and checked against the sources above. Company claims are reported as claims. Cover image is AI-generated.

Questions readers ask

Who is responsible for the AI agent attacks on university servers?
Amazon CISO CJ Moses stated that suspected Chinese-linked attackers deployed the AI agents. Amazon assessed with high confidence that the flaw was a zero-day because it found no prior fix or mention in vulnerability databases at the time of discovery.
What vulnerability did the AI agents exploit to take over the servers?
The agents found a path traversal zero-day in MONAI Label, an open-source medical imaging AI application. They used this flaw to arbitrarily write files on the server and execute a Python reverse shell running as root.
When did the campaign targeting universities in East Asia occur?
The campaign ran for about 48 hours beginning on September 8. Amazon's Threat Intelligence Team found a publicly hosted Cairn server on September 10 that provided visibility into the campaign.
Why is this incident significant for cybersecurity?
The case shows autonomous AI agents can find and exploit vulnerabilities faster than security teams can react. Amazon warned that organizations should rely more on automated defenses rather than manual investigations.
What fix was released for the MONAI Label vulnerability?
The MONAI Label fix was merged into the project's main branch on October 19 in pull request #1915. However, no patched release was available, leaving administrators to use workarounds such as enabling authentication and restricting the ANNOTATOR role to trusted users.

The day's checked news, by email

One short email a day with the top stories. Free, and you can unsubscribe in one click.

Or follow for the latest news:

How do you feel about this story?

Comments 0

No comments yet. Start the discussion.

Tech

CrowdStrike reportedly in talks to acquire Zscaler

CrowdStrike is in talks to acquire cybersecurity provider Zscaler, according to Bloomberg and Reuters. The talks are ongoing and no deal is guaranteed.

Checked against 6 sourcesSiliconANGLECNBCfinance.yahoo.com+3 more