Skip to content
worth noting Security

Framework Artex speeds up cyberattacks by connecting LLM models with penetration tools

only one source so far

The open-source tool Artex connects LLM models with penetration techniques, and since late August cybercriminals have been massively using it to attack companies, government offices and banks; according to the FBI and security firms, it speeds up attacks from days to hours, even without advanced skills.

The open-source framework Artex, written in Go, was published on GitHub in July by a pseudonymous developer going by the name Autumn-27. Since late August, according to researchers from several security firms (Check Point, GitGuardian, CrowdStrike, Censys, Cypen) and according to an FBI warning, cybercriminals have begun using it extensively to attack companies, government institutions and banks around the world. Artex itself is not an AI model but a "guide" that connects external frontier and open-source LLM models available via API with existing penetration tools, and allows covering the entire attack from gathering information about the network through gaining access to preparing data for exfiltration.

According to security firms, the tool does not bring new attack techniques, but significantly speeds up their execution — from days to hours — and allows attacks even by people without advanced technical skills. The FBI stated that Artex has been used against organizations in the financial sector, government administration, technology, defense, the semiconductor industry, energy, healthcare, manufacturing, telecommunications and legal services. In October, a China-linked group called Vapor Panda, according to CrowdStrike, hit over 200 organizations including think tanks and ministries of foreign affairs; according to Microsoft, the hacking group Storm-0940 used Artex to gain privileged access to Microsoft 365 systems in under three hours.

On 8 October, developer Autumn-27 removed the framework from GitHub, claiming to have nothing to do with the attacks and that the tool was created for "educational and research purposes." Nevertheless, according to sources, copies of Artex continue to spread and be developed alongside other similar tools — GitGuardian has documented a total of 45 different AI attack frameworks (including CyberstrikeAI, Hexstrike, CAIRN, PentAGI). Sources cite one advantage for defenders: the attacks generate a large volume of network traffic and are therefore still easily detectable — for example, the Dosu platform recorded 80 000 requests within 90 minutes.

What changed

Why it matters

The tool does not introduce new hacking techniques, but dramatically lowers the barrier to entry — attacks that used to take days can now be carried out by less experienced attackers in hours thanks to AI orchestration. This increases the frequency and scope of intrusion attempts against corporate and government networks across sectors. At the same time, sources state that these attacks are currently still easily recognizable by an unusually high volume of traffic, which gives defense teams a real chance to catch them in time.

Relevant practical impact

What this means

01

For a business

Companies in the financial, government, technology and other sectors face a real risk of faster and less costly attacks, because the tool allows less experienced attackers to carry out in hours what used to take days; according to security firms, these attacks are still easily detectable due to the large volume of traffic, which gives defenders an advantage, provided they have functional…

Risks and compliance
What to decide Check whether corporate monitoring can detect an unusually high volume of automated requests (thousands of probes in a short time) and whether systems are properly patched and access is secured.
More business impacts →
AI agents ARTEX Cybersecurity LLM open-source Penetration Testing

Check the original

Event sources

only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.

1
Wired — AI section independent context · first detected Artex Is Criminals’ New Favorite ‘Point-and-Shoot’ AI Hacking Tool