Wiz says AI agents are doing hands-on intrusion work
Original title: The Rise of Agentic Attacks: What They Look Like, and Where They're Headed
Named cases include ARTEX and Hermes, plus three Wiz investigations that begin with an exposed Flowise service.
Wiz says its incident response and threat hunting increasingly find AI agents performing hands-on intrusion work while humans set objectives. Cited public cases include ARTEX at South Korean financial institutions and in the ConnectWise Cloud Backup compromise attributed to Storm-0940, Hermes at Thailand’s Ministry of Finance, JPCERT/CC API attacks, four CrowdStrike-documented actors, and a DIVD breach. Wiz says current agentic activity is often noisy and difficult to separate from legitimate automation.
Unconfirmed: Autonomy and the strength of evidence vary by case, and some tooling is unidentified.
- Researchers reported PII of about 67,000 people taken in South Korea.
- Storm-0940 attempted sign-ins against more than 2,250 Microsoft 365 tenants.
- Hermes ran unattended; investigators found web shells and a custom Go implant.
- Hermes initial access was unknown, with no exfiltration evidence at discovery.
- DIVD’s statement did not identify the agent or framework.
Impact: Credential theft, secret exfiltration, and personal data exposure
| Product | Affected |
|---|---|
| Flowise | — |
- Vendors
- ConnectWise
- Threat actors
- Storm-0940
- Organizations
- WizMicrosoftCrowdStrikeJPCERT/CCHunt.io
From the original article
13 min read at wiz.io
AI agents are changing how attackers operate. Explore real-world cases of agentic attacks, learn how to identify their fingerprints, and discover what they mean for the future of cybersecurity.
By Avigayil Mechtinger · wiz.io
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.