Cairn agents exploited a MONAI Label zero-day to take over a server
Original title: Hackers use autonomous AI tool Cairn to find zero-day, take over server
Amazon ties the September 8–10 activity to a Chinese-nexus actor with moderate confidence and published two related IPs.
Amazon Threat Intelligence says a publicly exposed Cairn instance showed agents, told only to obtain shell access, finding a path-traversal zero-day in MONAI Label. Within 11 minutes they mapped the API and exploited the flaw, later planting a Python reverse shell that yielded root. The same bug was aimed at a second university. Maintainers shipped a security update on November 18.
Unconfirmed: Chinese-nexus attribution is moderate confidence, based on prompt language, timestamps and target choice.
- Targets were mainly East Asian educational institutions outside China.
- A 16-hour, 148-action attempt against a Redis server failed.
- Another run tried over 30,000 credentials on a university GPU server and failed.
- Associated addresses were 142.4.98.145 and 142.4.107.34.
Impact: root access on a server via path traversal
| Product | Affected |
|---|---|
| MONAI Label | — |
- 1.Install the MONAI Label security update released on November 18.
- 2.Review connections from 142.4.98.145 and 142.4.107.34 between September 8 and 10.
- Products
- MONAI Label
- Malware
- Cairn
- Organizations
- Amazon Threat Intelligence
- Countries
- China
Indicators of compromise
auto-extracted · verify before use · export allAll| Type | Indicator | Context |
|---|---|---|
| ipv4 | 142.4.107.34 | , observed between September 8 and 10, 2026: 142[.]4.98.145 142[.]4.107.34 Amazon warned that AI agents can find and exploit vulnera |
| ipv4 | 142.4.98.145 | th the campaign, observed between September 8 and 10, 2026: 142[.]4.98.145 142[.]4.107.34 Amazon warned that AI agents can find and |
From the original article
2 min read at bleepingcomputer.com
Threat actors used the open-source AI penetration-testing framework Cairn to find a previously unknown vulnerability in a medical imaging application. [...]
By Bill Toulas · bleepingcomputer.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.