Threats to ARTEX Autonomous Penetration Testing Framework and how bad actors seems to be abusing it to their advantage have been unearthed by the FBI’s investigations. The framework is an AI powered Penetration Testing guideline has been weaponised to serve a different need to hackers than its intended use since August. A growing number of malicious cyber attacks are targeting organisations spanning various industries.

ARTEX has specialist properties delivering AI driven,  autonomous, multi-agent penetration to deliver machine demands.

The attackers automate cyber intrusions at machine speed across critical industries, including energy, manufacturing, and telecommunications.

In a FLASH advisory, the FBI said threat actors have used the Go-based platform since at least August 2026 to chain together attacks from initial targeting to data exfiltration, often reaching privileged access in under three hours.

ARTEX relies on external large language models rather than a built-in model, with attackers often favouring Chinese originating models such as DeepSeek Flash due to low cost, performance, and weak restrictions around penetration testing prompts.

Organisations are urged to audit exposed and peripheral assets and lock down cloud and administrative interfaces, as well as tighten identity and API controls, preserve records, and report suspicious activity to local FBI officials or internet compliant centres.

The tool’s GitHub repository was removed in October, following the reports of abuse, however the FBI reported that ARTEX remains widely available through another commercial model called ARTEX PRO.