Jul 9, 2026Confirmed
Threat actor used Hermes AI agent unattended inside Thailand's Ministry of Finance network, Hunt.io reports
AI loose in the worldUnauthorized actionDERIVED
Approved by person, Oct 11, 2026. How entries are approved
Hunt.io reported that an unidentified threat actor ran Nous Research's Hermes agent in unattended "YOLO" mode during an intrusion targeting Thailand's Ministry of Finance. Logs showed it probing for privilege escalation and searching a personnel records directory. No data exfiltration was found, and the ministry had not confirmed a breach.
What this means for youAttackers are now using AI agents that keep working on a hacked network without a person watching. Governments and organisations holding personal records should expect faster, more automated intrusions. Small businesses can check that staff accounts have only the access they need and that login activity is logged and reviewed.
Sources
- AI Incident Databasehttps://incidentdatabase.ai/cite/1669
Data from the AI Incident Database (incidentdatabase.ai), licensed CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/). McGregor, S. (2021) Preventing Repeated Real World AI Failures by Cataloging Incidents: The AI Incident Database. IAAI-21.
Source data license: CC BY-SA 4.0 (database). Report contents keep their own rights; reports.text, submissions, images and videos are excluded from the data license.
Covered by snapshot 45
da93f03b406f409940d0a6ef47b526607f82ebf457d98f69299b88aceede576a
This entry is record number 2100. It has not been edited since it was approved; the hash above would not match if it had. How to verify