Particle.news

ARTEX Makes AI-Driven Hacking Accessible and Has Been Used in Real Breaches

U.S. authorities warn the code speeds intrusions by letting users give plain-language commands

Overview

  • The framework was published in July and removed from the original GitHub repository in October, but forks and archives have kept it available and researchers found more than 1,600 publicly exposed control consoles by early November.
  • The FBI said on November 13 that the tool has been used in multiple intrusions, including attacks that hit South Korean financial institutions and the ConnectWise Cloud Backup incident that affected thousands of customer organizations.
  • Security firms and researchers say the software runs AI 'worker agents' that scan targets, try exploits, escalate privileges, and steal data after receiving simple English instructions from an operator.
  • Experts warn the tool lowers the technical barrier to high-impact hacks because it can be run from a smartphone and can complete privilege gains in hours, which leaves little time for human-led incident response.
  • Agencies and firms are urging immediate defenses such as rapid secrets rotation, API rate limits, strict zero-trust access controls, and automated detection and response while publishing indicators of compromise and hardening guidance.