---
id: "c692744a-899f-4b3b-bd90-652bf2af93ee"
kind: "link"
title: "Hackers use CyberStrikeAI AI agents to take over Windows domains"
language: "en"
section: "security"
date: "2026-11-02T11:47:51Z"
updated: "2026-11-02T11:47:51Z"
canonical: "https://tldlog.com/l/hackers-use-cyberstrikeai-ai-agents-take-over-windows/"
origin: "master"
translations:
  es: "https://tldlog.com/es/l/hackers-usan-agentes-ia-cyberstrikeai-tomar-control-dominios/"
  de: "https://tldlog.com/de/l/hacker-nutzen-ki-agenten-cyberstrikeai-windows-domains/"
  fr: "https://tldlog.com/fr/l/pirates-utilisent-agents-ia-cyberstrikeai-prendre-controle/"
  it: "https://tldlog.com/it/l/hacker-usano-agenti-ia-cyberstrikeai-prendere-controllo/"
  pt-BR: "https://tldlog.com/pt/l/hackers-usam-agentes-ia-cyberstrikeai-assumir-dominios/"
  ru: "https://tldlog.com/ru/l/khakery-ispolzuyut-ii-agentov-cyberstrikeai-zakhvata-domenov/"
  zh-Hans: "https://tldlog.com/zh/l/heike-liyong-cyberstrikeai-jiemang-windows-yu/"
source:
  name: "cybersecuritynews.com"
  url: "https://cybersecuritynews.com/hackers-use-cyberstrikeai-agents-2/"
  language: "en-US"
  date: "2026-11-02T10:34:24Z"
---

# Hackers use CyberStrikeAI AI agents to take over Windows domains

## in plain words

Criminals used a tool that uses artificial intelligence to break into company computer networks. The AI chose the steps, and a human set the goal. Security researchers at Securelist found the attack during routine monitoring.

Hackers are using CyberStrikeAI, an AI-driven command framework, to break into Windows networks and take over entire domains. Education organizations in Africa and Latin America were among the confirmed victims. The tool lets a language model decide which commands to run, so the operator does not type each step by hand.

CyberStrikeAI is an open-source, Go-based platform built for penetration testing. Criminals have already turned it toward real targets, and earlier reports showed attackers using the same framework against Fortinet FortiGate devices. Analysts at Securelist identified this campaign during routine threat monitoring and described it as one of the first public cases of a language model being used inside a live intrusion.

In this setup, the human operator sets a broad goal, while the AI model on the control server picks the tools, writes the commands and reads the results. The malware on the victim machine is a plain implant with no AI inside it.

The intrusion began with a person who got into a compromised SQL Server and scanned the network. After about two weeks of inactivity, the CyberStrikeAI agent took over, set up persistence through a scheduled task and mapped the domain. It also collected data needed for Kerberoasting and AS-REP Roasting, two methods for cracking account passwords.

The final phase ran in under an hour. The agent moved across the network, dumped credentials from many machines, pulled passwords from the LSASS process, and compromised the built-in domain administrator. The whole job, from agent install to domain admin, took three to four days.

The attackers used Base64 encoding and split scripts into several files to stay hidden. Some scripts contained Chinese comments, and server restarts mostly fell within UTC+8 business hours. Securelist could not link the activity to any known group.

The methods themselves are not new, and defenders can still detect them with ordinary rules. Teams should watch for remote scheduled tasks, LSASS memory access, and WMI or WinRM commands from unusual hosts, and should patch and lock down exposed SQL Servers.

Source: [cybersecuritynews.com](https://cybersecuritynews.com/hackers-use-cyberstrikeai-agents-2/)
