Home icon

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

Security Blog



This article details Amazon Threat Intelligence's investigation into a Chinese-nexus threat actor using Cairn, an autonomous AI penetration-testing framework, to conduct hacking operations with minimal human involvement.

  • Threat actor deployed Cairn AI agents with goal to "obtain shell access" against universities in East Asia
  • Agents discovered a zero-day path traversal vulnerability in open-source medical imaging application within 11 minutes
  • Agents exploited zero-day to write reverse shell payload and gain root access to server
  • Against hardened targets, agents demonstrated persistence through 16+ hours of continuous attempts without human direction
  • Agents conducted credential sprays testing 30,000+ combinations and mapped internal network topology via SSRF
  • AWS notified project maintainers; vulnerability was patched and GitHub Security Advisory released
  • Autonomous AI agent frameworks now adopted by both espionage and financially motivated threat actors

Organizations must prioritize automated response to high-confidence detections to match the operational speed of autonomous AI agents in offensive operations.



Go to article

The AWS News Feed is currently looking for gold sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.

Related articles

Nov 19
2026
Real-time AI agents with Confluent Cloud and Amazon Bedrock AgentCore
Nov 13
2026
Amazon Threat Intelligence observes threat actors running cybercrime campaigns with Artex AI agents
Nov 12
2026
Inside an enterprise AI agent: A governed, observable loop architecture
Nov 17
2026
Runtime governance for AI agents with Noma Security and AgentCore Gateway

The AWS News Feed is currently looking for silver sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.