CISO Voice

Breaches & Vulnerabilities / News

Hackers use Hermes AI agent to target Thai Ministry of Finance

The malicious agent enumerated ministry hosts, traversed files, and captured LinPEAS output

Akshaya Asokan
Akshaya Asokan Cybersecurity Reporter · July 27, 2026 · 3 min read
Hackers use Hermes AI agent to target Thai Ministry of Finance

Hackers targeted the Thai government using Hermes, an autonomous AI agent the attackers used to probe vulnerabilities, enumerate hosts, and search files.

Security firm Hunt.io, along with independent researcher Bob Diachenko, uncovered the campaign after they discovered exposed directories that contained hundreds of custom scripts and configuration files.

On analyzing the files, the researchers uncovered activities linked to Hermes autonomous agents employed by the hackers. The agents were running on the application’s YOLO mode, a feature that disables dangerous command safety checks for the session.

“Custom scripts and configuration files across all three exposed directories reference Ministry of Finance systems by name, hostname, and internal addresses. The targets span an administrative web panel, a Hadoop big-data cluster and its Ambari management platform, all on non-routable IPs. Combined, they describe an operator with detailed knowledge of the ministry’s internal topology,” Hunt.io wrote in a threat advisory.

Hunt.io did not name the attackers, but noted the server has a history of hosting ShadowPad and currently runs a VShell C2 — tooling associated with Chinese-speaking actors. Combined with a Chinese-language password element in the agent’s web panel and an API key for a Chinese reconnaissance platform, Hunt.io assessed with low-to-medium confidence that the operator is Chinese-speaking.

Hunt.io added that it notified Thailand’s national CERT and NCSA in mid-July, who did not immediately respond to a request for comment.

The attackers mainly used the AI agent for privilege escalation and further enumeration of the  Thai Ministry of Finance network. For this, it used the open-source project LinPEAS (Linux Privilege Escalation Awesome Script) to traverse the network.

Once the system was accessed, the agent scanned for local privilege escalation vulnerabilities in the Linux kernel tracked as CVE-2026-43503, CVE-2026-31431, CVE-2026-43284 and CVE-2026-43500.

Once the targeted system was compromised, the attackers staged a custom Go implant called Hades. The attackers then proceeded to target the ministry’s Apache Hadoop, GlassFish administrative console, and Apache Ambari.

To prevent any compromise tied to the campaign, the researchers recommend that companies review HiveServer2 authentication mode, audit web roots, patch the exploited vulnerabilities, change default credentials on GlassFish admin consoles and restrict access to trusted networks.

Rise in Agentic AI Attacks

The disclosure from Hunt.io follows industry warnings of heightened attacks tied to agentic AI systems as models rapidly progress in terms of their capabilities. A recent report by LexisNexis Risk Solutions found there has been an 8% increase in global attack rates, driven by a 450% explosion in agentic commerce traffic.

Last week, OpenAI claimed its GPT-5.6 Sol model and an unreleased model compromised Hugging Face. Prior to that, security firm Trend Micro uncovered an attack by a Russian actor who used Google’s open-source agentic AI, Gemini CLI, to migrate a command-and-control server and to control a small-scale botnet.

Akshaya Asokan
Akshaya Asokan Cybersecurity Reporter

Akshaya Asokan is a U.K.-based cybersecurity and technology journalist with experience covering nation‑state hacking operations, cyber and tech policy, and emerging cybercrime trends. She previously served as the senior European cybersecurity correspondent for Information Security Media Group’s Global News Desk. Before that, she reported on IT developments for IDG Media and covered artificial intelligence and machine learning for Analytics India Magazine. Earlier in her career, she wrote on sexual and religious minority rights as a reporter for The New Indian Express.

Keep reading for free,
one signup unlocks everything

We'll send you the CISO Voice newsletter. Unsubscribe anytime.

More from CISO Voice

Sign up to CISO Voice, the free newsletter trusted by senior security leaders.

Secret Link