AI Framework Breaches Government Systems in Asia, Stealing Thousands of Records

A multi-agent AI framework, including Hermes and OpenClaw agents, was used to compromise government systems in Asia, resulting in the theft of over 2,500 personnel records and the cracking of 85 employee accounts. Dream Research Labs reported that the attack demonstrated the capability of coordinated AI agents to execute intrusion operations at machine speed. The operational archive discovered contained 1,395 files generated over four days of activity, highlighting the extensive nature of the breach. The attack showcases a significant shift in cyber capabilities, leveraging open-source AI tools for sophisticated intrusions. The current status of the breach is under investigation, with researchers analyzing the extent of the impact on affected government entities.

Key Points: • AI agents compromised government systems in Asia, stealing over 2,500 records. • The attack involved cracking 85 employee accounts using Hermes and OpenClaw frameworks. • A 160 MB archive was found, containing 1,395 files from four days of activity.