Security researchers have warned for at least a year that AI agents could accelerate cyberattacks, but that era has arrived. Since late August, companies, government agencies, and banks worldwide have been targeted by cybercriminals using AI-powered penetration testing tools.
These setups mimic legitimate defensive techniques. They use large language models to scan potential targets and force their way into poorly secured networks, then carry out attack plans written by human operators. Dozens of these frameworks, often called “harnesses,” link AI models to cybersecurity testing or attacks.
The tools do not introduce new ways to attack websites or servers. They are not sophisticated in that regard. Instead, they allow cybercriminals to launch attacks at speeds impossible for humans to match. Attacks that would have taken days are completed in hours.
Researchers say hackers do not need to be skilled to use these tools. Agentic frameworks have progressed from experimental use to sustained, multi-objective offensive operations that run with minimal human intervention. CJ Moses, chief information security officer at AWS, wrote this in a recent company blog post.
The most prominent tool is Artex. Written in the programming language Go, the open-source framework was first published on GitHub in July by a pseudonymous developer using the handle Autumn-27. It has spread across some hacker communities since then. Artex is used across all parts of attacks, from initial network analysis and gaining access to moving within systems and preparing data for export.
“It’s really an all-in-one product,” says Jacob Larsen, an offensive security researcher and founder of security firm Cypen. “Previous iterations of AI agent pen-testing tools would either only do reconnaissance—like the scanning and analysis [of targets]—and some would only do testing. This is really a point-and-shoot weapon.”
Artex is not an AI product itself. It is essentially a guide that points external frontier and open-source AI models to a series of existing penetration testing tools and processes. People using Artex can give simple, natural-language prompts to begin operations. Researchers at Check Point report that hackers prompted Artex with instructions like: “Launch a penetration attack on [Redacted victim] to gain access to the site’s admin panel, database, or server.”
Once instructions are issued, multiple AI agents carry them out. One agent breaks down plans into tasks, others perform these actions, and the system presents findings. Safety guardrails in underlying AI models, particularly those from frontier labs, can stop hacking activity when they detect it. Increasingly, however, free Chinese and abliterated models are getting better at breaking into systems.
“It’s fast; it’s efficient,” Guillaume Valadon, a researcher at security firm GitGuardian, says of Artex. Valadon and others say the simplicity of the tools makes it easier for unskilled wannabe hackers to try to enter the market. “It lowers the bar,” Valadon says. “I would say my grandmother can do it, except that she doesn’t have a laptop, but the bar is that low.”
As a result, a flurry of reports from cybersecurity companies links Artex use to attempted and successful hacks. These often involve Chinese-speaking cybercriminals trying to make money. They include the theft of personal information from thousands of South Korean financial institutions, potential cryptocurrency theft, and hunting for exposed login details on GitHub.
“A variety of malicious cyber actors have used Artex to conduct campaigns targeting organizations globally across multiple sectors, including financial services, government, technology, defense, semiconductor, energy, health care, manufacturing, telecommunications, and legal industries,” the FBI warned in an alert last week.
In October, China-linked hackers known as Vapor Panda used Artex to target more than 200 organizations. This included think tanks, semiconductor and defense firms in the US, and government foreign ministries in Europe and elsewhere, according to security firm CrowdStrike. Microsoft spotted Chinese hackers it tracks as Storm-0940 using Artex to try to sign in to Microsoft 365 setups following a supply chain attack. It took hackers less than three hours to gain privileged access to systems.
On October 8, the Autumn-27 developer removed Artex from GitHub. They stated in a comment that they are “no way associated with these malicious attacks” and that the tools were developed for “educational and research purposes.” An email WIRED sent to an address linked to the developer did not receive a response.
Despite the developer pulling Artex, multiple versions of the framework have been copied and continue to be developed and used, alongside other similar tools. Valadon says GitGuardian has documented 45 different AI attack frameworks so far. CyberstrikeAI, Hexstrike, CAIRN, and PentAGI are frequently mentioned by researchers. They all work in similar ways, with some slightly different technical capabilities.
One silver lining for people defending against attacks using such tools is that they are currently easy to spot. “These tools are not stealthy right now,” says Silas Cutler, principal security researcher at internet security firm Censys. Many attacks create a lot of noise in systems—making thousands of probes in quick succession. For instance, agentic coding platform Dosu said Artex was used to send 80,000 requests to its systems in just 90 minutes.
Researchers say basic security fundamentals can help defend against many approaches from those using Artex. Patching software, not using default credentials, not exposing secrets, and monitoring for unusual network activity are key.
Pinpointing the number of cybercriminals using Artex is not easy. However, thousands of instances of the systems being set up have been exposed to the internet, according to analysis by Censys and the Shadowserver Foundation. Perhaps hinting at the sophistication of those using Artex, researchers have found multiple instances of the tool left unsecured and fully open to the internet. This allows them to inspect how operators are using it.
Last week, Cutler says, two exposed Artex servers appeared to show Chinese native speakers targeting other Chinese entities. “In one of the cases, it was a university that was being targeted with a password spraying attack, and it has all of that vibe of a student trying to change their grades,” he says.
“There is a level of this activity that’s occurring that is likely organized, sophisticated, criminal activity,” Cutler says. “There is also, in parallel, likely a subset of the activity that falls more into that juvenile crime realm.”
What it means
For defenders, the immediate priority is to stop relying on default credentials and to monitor network activity for sudden spikes in automated requests. The tools themselves are loud and leave traces. For the attackers, the barrier to entry has dropped sharply, allowing less skilled individuals to execute complex campaigns that previously required significant expertise.




