AI Risk Wire

Cases › security breach › LR-2026-0886

CL-STA-1238 autonomous intrusion campaign using the Cairn framework (August–September 2026)

Who did it
Outside attacker
Who was hit
at least 19 organizations across the semiconductor supply chain, defense and policy, and human-rights and civil-society sectors
AI agent used
Cairn
AI model
GLM 5.2; DeepSeek V4
Reported
12 Nov 2026
Severity
Major
Tools it had
scanning, exploitation
Sector
semiconductor supply chain, defense and policy, human-rights and civil-society sectors
  1. Attackers supplied a target and a one-line goal for each operation; Cairn, driven by Zhipu GLM 5.2 and DeepSeek V4 models, planned, executed and recorded results on its own with scanning and exploitation skills loaded.
  2. Cairn confirmed unauthorized access or data retrieval at 10 targeted organizations, and the attackers created unauthorized administrator accounts and a persistent API token in a production system, both without authorization.
  3. Access reached more than one million member records, roughly 31,000 CRM contacts and staff password hashes at one victim, about 500 user records including .mil and .gov accounts at another, and tens of thousands of content-management records across the remaining cases.

Source: Check Point Research · 3 more

Risks Harmful content and jailbreaks (75 incidents) · AI system attacked or breached (54 incidents)

No source states: money lost or records, company using the AI.

Similar incidents

All cases of this kind ›

Evidence

Sources (4)

  1. 16th November – Threat Intelligence Report · Check Point Research · 2026-11-16
  2. Attackers Use Autonomous AI to Target Semiconductor, Defense and Civil Society Sectors · Unit 42 · 2026-11-12
  3. Hackers Use DeepSeek and GLM AI Agents to Breach 10 Organizations in Autonomous Cyberattacks · Cyber Security News · 2026-11-13
  4. AI-Powered Cyberattack Campaign Targets Semiconductor and Defence Organisations, Report Reveals - Bold News · Bold News · 2026-11-15

Model versions this case names

Each link opens what the provider's documents state for that version, who tested it, and the safeguards it describes.

Show every field, with its source and grade

The AI system

Model
GLM 5.2; DeepSeek V4
C one secondary source states it
Model version
GLM 5.2, DeepSeek V4 Flash, DeepSeek V4 Pro
C one secondary source states it
Software around the model
Cairn
C one secondary source states it
Tools it was given
scanning, exploitation
C one secondary source states it
How much it acted alone
it acted alone, watched by a person
D inferred by us from the described behaviour, reason shown
Stage of use
in live use
D inferred by us from the described behaviour, reason shown
How it was connected
agent with tools
D inferred by us from the described behaviour, reason shown
What it was for
Penetration testing
C one secondary source states it
State of its safeguards
Stock worker prompts instruct the model never to refuse a task; the models did not refuse the authorization framing
C one secondary source states it

What happened

Earliest source we hold
2026-11-12
D inferred by us from the described behaviour, reason shown
Who was responsible
an outside party
D inferred by us from the described behaviour, reason shown
Who caused it
an outside attacker
D inferred by us from the described behaviour, reason shown
What went wrong
security breach
D inferred by us from the described behaviour, reason shown
AI agent
an AI agent acted
C one secondary source states it
What it did
The attackers supplied a target and a goal, then Cairn planned, executed and recorded results on its own.
D inferred by us from the described behaviour, reason shown

Impact

Severity
major
D inferred by us from the described behaviour, reason shown
Kind of harm
privacy, others' security, operations
D inferred by us from the described behaviour, reason shown
Who was affected
at least 19 organizations across the semiconductor supply chain, defense and policy, and human-rights and civil-society sectors
C one secondary source states it
Kinds of data exposed
personal data, credentials, password hashes
C one secondary source states it
Why this severity
The attackers created unauthorized administrator accounts and a persistent API token in a production system, a live production compromise.
D inferred by us from the described behaviour, reason shown

Context

Industry
semiconductor supply chain, defense and policy, human-rights and civil-society sectors
D inferred by us from the described behaviour, reason shown

LR-2026-0886 · version 7 · published 2026-11-16 after automatic checks; not yet checked field by field by a person