iCreatAI News
Skip to content
Hacker News AI· acossta·· 10 hr agoFeaturedAI score82

Chinese-nexus actor's Cairn AI agents found a zero-day on their own

Chinese-nexus actor's Cairn AI agents found a zero-day on their own

AI summary

Amazon Threat Intelligence reported that a Chinese-nexus threat actor used the open-source autonomous AI penetration-testing framework Cairn, left running without authentication on a public server, to target IP addresses tied to universities in East Asia outside mainland China over roughly 48 hours from September 8, 2026. The agents ran long, unsuccessful campaigns against an open Redis instance and a university GPU server, but against an open-source medical imaging AI application they found an unpatched path traversal flaw and used it to write a reverse shell running as root in about 11 minutes. Amazon assessed with high confidence that the flaw was a zero-day, notified the maintainers who released a fix and GitHub Security Advisory, and worked with a government partner to notify the affected non-AWS organizations.

Why it matters

It documents a real case of autonomous agents discovering and exploiting a zero-day, with concrete timelines that illustrate how agentic operations compress defenders' response windows.

Source: Hacker News AI · aws.amazon.com