Executive Summary
An attacker ran the Hermes AI agent unattended on a rented server with permission prompts disabled and pointed it at Thailand's Ministry of Finance, where it explored hosts for root, searched file systems and continued post-exploitation on its own.
What Happened
Someone installed a popular AI assistant on a rented server, switched off the setting that makes it ask permission before running risky commands, and pointed it at Thailand's Ministry of Finance, which runs the country's treasury and tax collection. The agent then worked through the ministry's network on its own, checking hosts for ways to gain root access, hunting through file systems, and
Affected Products
- Hermes Agent
Vulnerability / Attack Technique
Unattended AI agent post-exploitation
Technical Mechanism
With approval prompts turned off, the agent autonomously probed for privilege escalation and searched files across the ministry network.
AI Security Category
AI-assisted attack
Part of the AI-Assisted Attacks topic.
Relevant Frameworks
RealGround Analysis
Turning off one safety setting turned an assistant into an intruder. RealGround analysis: government networks now face unattended agents working through them.
Recommended Controls
- Monitor for automated host enumeration inside the network
- Harden hosts against local privilege escalation
Related RealGround Service
Secure AI Agent Build · or start with the free AI Agent Security Scanner.
Sources
thehackernews.com: https://thehackernews.com/2026/07/hacker-runs-hermes-ai-agent-unattended.html
Summary and analysis written by RealGround from the source above; last updated 2026-10-18.