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

Cybersecurity researchers at TrendAI disclosed details of three cyber attacks targeting organizations in Japan in which artificial intelligence played a key role. The researchers analyzed three exposed server environments used by threat actors, ranging from hand-operated workspaces to AI-agent platforms that automated entire workflows, including command and control.

CyberStrikeAI case

The first attack abused CyberStrikeAI, an open-source penetration testing tool previously used in a campaign targeting Fortinet FortiGate appliances. The activity was concentrated in a short period in October 2026 and used 44 agents to automate reconnaissance, vulnerability discovery and validation, exploitation, command and control, persistence, data exfiltration, and automated reporting. It also enabled generation of forms and fake login pages in Japanese. At least 19 identifiable organizations, including at least 16 in Japan, were subjected to validation activity, and each had a report in the same format with an index of evidence.

Human-led and tool-building cases

The second case was human-led with assistance from a large language model to accelerate planning and tool development. Analysis indicates the campaign targeted AI workflow platforms and inference services to obtain large amounts of inference capacity at low cost. One intrusion targeted a geographic information system server exposed with default credentials, resulting in theft of cadastral and commercial records, Windows credentials, and email data. The campaign lasted 24 days in August and September 2026. The third case used AI mainly to develop the operator's tools and scripts. The threat actor ran at least six campaigns from January to October 2026 targeting edge devices from Cisco, Citrix, F5, and Yamaha using published vulnerabilities, management functions, or credentials in configuration files, and opportunistically struck exposed VPN devices and WebLogic servers.

Broader context

TrendAI said there is no evidence at this stage that the three incidents share an operator, and it is unknown which AI models were used. The first operator also targeted organizations in Taiwan and Hong Kong, while the third case included organizations in the United Kingdom, France, Italy, Qatar, and Colombia in its WebLogic discovery lists. TrendAI said AI has increased attackers' ability to try thousands of targets at once, adapt to a target country's language, and continue without fatigue. The report also described other AI-oriented campaigns, including BUGSTER using Hermes Agent with DeepSeek V4 and Cursor; CyberCartel using CARLIY0S with four Claude models; an offensive agent able to run up to 100,000 steps; and a Chinese-speaking actor using Cairn against more than 100 companies. Microsoft's 2026 Digital Defense Report said AI has collapsed median time from vulnerability discovery to weaponization to well below 24 hours.

Sources: The Hacker News