CyberStrikeAI Uses 44 AI Agents to Automate Hacking Operations Against Japan Targets

TrendAI analyzed three exposed server environments used by threat actors to conduct malicious cyber operations against organizations in Japan. These environments differed from previously published attacks that leveraged ARTEX, an autonomous AI penetration testing framework targeting entities in South Korea and Japan. The environments contained reconnaissance tools, scripts, target lists, and command histories, ranging from hand-operated workspaces to AI-agent platforms that automated entire workflows.

The first attack abused CyberStrikeAI, an open-source penetration testing tool previously used in campaigns targeting Fortinet FortiGate appliances. The activity occurred during a short period in October 2026 and leveraged 44 agents to automate various attack lifecycle stages, including reconnaissance, vulnerability discovery, exploitation, command-and-control, and data exfiltration. The system could generate Japanese-language forms and fake login pages.

Human-Led Attacks

The second case was a human-led attack with assistance from a large language model (LLM) to accelerate planning and tool development. This campaign targeted AI workflow platforms and inference services to obtain large amounts of inference capacity at low cost. One intrusion compromised a geographic information system (GIS) server with default credentials, resulting in the theft of cadastral records, Windows credentials, and email data. The campaign lasted 24 days in August and September 2026.

The third attack saw AI play a limited role, primarily in developing the operator's tools and scripts. This campaign targeted edge devices from Cisco, Citrix, F5, and Yamaha using published vulnerabilities or credentials from configuration files. The threat actor ran at least six campaigns from January to October 2026, opportunistically striking exposed VPN devices and WebLogic servers.

Researchers noted there is no evidence suggesting these three incidents share an operator. The attacks are not confined to Japan and neighboring countries, with targets identified in Taiwan, Hong Kong, the U.K., France, Italy, Qatar, and Colombia.

AI's Expanding Role

AI has increased attackers' capabilities by allowing them to try thousands of vulnerabilities simultaneously, adapt to target languages, and operate without fatigue. According to Microsoft's 2026 Digital Defense Report, AI has collapsed the median time from vulnerability discovery to weaponization to below 24 hours.

The growing demand for AI tools by threat actors has created new monetization avenues, with stolen AI provider session tokens and offensive AI tools like APEX AI, MessiahGPT, and DarkGPT being sold. This has further commoditized cybercrime and lowered barriers to entry for advanced attacks.

Microsoft recently revealed that a China-state-aligned threat actor compromised a ConnectWise cloud backup application and used ARTEX to access over 2,250 Microsoft 365 tenants. In South Korea, authorities identified seven suspects linked to a cybercrime operation that breached more than 90 organizations, with a 32-year-old mastermind alleged to have worked at a cybersecurity company between 2019 and 2022.