Hundreds of Artex AI hacking tools found exposed on the open web

Security researchers have found hundreds of servers hosting the penetration testing tool Artex exposed on the open internet with no access restrictions, meaning anyone who knows the address can use the hacking tool.

According to TechCrunch, the exposed instances all run the same AI-related hacking tool. The researchers' concern is not whether the tool itself is legitimate, but that these deployments lack basic access control.

Penetration testing tools are standard equipment in the security industry, where corporate teams use them to check their own defenses. The same capability hanging on the public internet without authentication turns authorized testers into anyone who walks by.

Exposures like this are usually not deliberate: a deployment misses a default, an admin panel has no password, a port is left open to the internet, or a test environment is kept running as a long-lived service.

When the tool carries AI capability the risk is amplified — the barrier to use drops, and someone without deep offensive security experience can launch scans or attempt exploitation.

The report does not identify which organizations own the affected servers, how long they were exposed, or whether any of the instances were misused. What is confirmed is the scale of the exposure and the unauthenticated access.

For operators the response is concrete: audit publicly reachable admin ports, put authentication and network isolation in front of these tools, and track the lifecycle of test instances so temporary environments do not serve the internet indefinitely.