Chinese-speaking attackers used the AI framework Cairn to run autonomous intrusions against 19 organisations
- Severity
- High
- Sources
- 1
What happened
Unit 42 tracked a China-linked campaign in August and September 2026 that used a customised fork of the open-source Cairn framework, hosted in Hong Kong with a US egress and staging node, driven by Zhipu GLM 5.2 and DeepSeek V4 Flash/Pro. The attackers ran 25 operations against at least 19 organisations, achieving access at 10, with unauthenticated cloud storage access, persistent admin accounts, and a separate flaw enabling arbitrary customer logins. The report notes authorization.md files used as a pretext, and a shared-blackboard fact-intent architecture. No CVE identifiers are given in the article. Defenders should monitor for unauthorised changes to critical records and privileged access, and test their ability to verify and restore trusted data.
For security analysts · Written by AI from 1 source · Wrong?
What to do
- Monitor for unauthorised changes to critical records and privileged access.
- Test your ability to verify and restore trusted information after an incident.
Written by AI from the sources, checked against them
Full story timeline · 1 update
Who reported · 1 outlet
Who and what is in it
Companies