What happened
Attackers suspected of ties to China used Cairn, an open-source framework that coordinates autonomous AI agents for penetration testing, to uncover a previously unknown path traversal flaw in MONAI Label and seize full control of a server, BleepingComputer reported. Amazon Threat Intelligence, which investigated the activity, said the operators largely told the agents to get shell access and let them work out the rest.
According to Amazon, the agents went from mapping the application's API to exploiting the bug in 11 minutes. The flaw let them place files anywhere on the server. In a later session they planted a Python reverse shell in sitecustomize.py, which runs when Python starts, then called an API endpoint to gain root. The operator next gave the agents a proxy into a second university network running the same software.
The campaign ran from September 8 to 10, 2026, mostly against educational institutions in East Asia outside China. Not every attempt worked. A 16-hour effort involving 148 actions against a Redis server failed, and more than 30,000 credential guesses over six hours did not break into a university GPU server. Amazon uncovered the operation because the attackers had exposed their Cairn instance online with no authentication.
Context
Cairn was created for research and for strengthening defenses, but this is not its first reported misuse. Gambit Security documented its role in a large credit card theft operation in September, and Google reported it was used against a U.S. aerospace organization. Amazon tied the latest campaign to a Chinese-nexus actor with moderate confidence, and MONAI Label's maintainers issued a fix on November 18.
The briefing
The week in tech, in one email.
The stories worth your time, every week. Unsubscribe anytime.
