Honeypot captures agent that hijacks exposed AI models
The captured prompt tells a stolen model to scan, exploit web apps, steal secrets and mine, for up to 100,000 steps.
64
Product
go2rtc
- Mentions · 7d0
- Mentions · 30d2
- All mentions4
- First seenAug 12, 2026
- Last mentioned7d ago
Timeline
4 articles · newest firstSymantec finds AI agent hijacking exposed models for autonomous attacks
A honeypot captured a five-layer jailbreak and a loop of up to 100,000 steps that scans, steals data, and installs XMRig.
68
ZDI-26-561: (Pwn2Own) Home Assistant Green go2rtc Command Injection Remote Code Execution Vulnerability
ZDI disclosed a Pwn2Own command injection flaw (CVSS 7.5) in Home Assistant Green's go2rtc, enabling network-adjacent attackers to execute arbitrary code via localhost access.
28
ZDI-26-560: (Pwn2Own) Home Assistant Green go2rtc Command Injection Remote Code Execution Vulnerability
ZDI disclosed another Pwn2Own command injection flaw (CVSS 7.5) in Home Assistant Green's go2rtc, enabling network-adjacent code execution after obtaining localhost access.
28
Appears with
- Home Assistantvendor · 2
- Zero Day Initiativeorganization · 2
- Home Assistant Greenproduct · 2
- Zero Day Initiativevendor · 2
- Jupyterproduct · 1
- XMRigproduct · 1
- Djangoproduct · 1
- XMRigmalware · 1
- Symantecvendor · 1
- Symantecorganization · 1
- Laravelproduct · 1
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.