FBI Warns AI Hacking Tool ARTEX Is Automating Cyberattacks at Machine Speed
The FBI has warned that hackers are weaponising an open-source artificial intelligence tool called ARTEX to carry out cyberattacks “at machine speed, from reconnaissance through data exfiltration”. In a FLASH advisory issued on Friday, November 13, the bureau said malicious actors have used the framework since at least August, targeting organisations across financial services, government, technology, defence, semiconductors, energy, healthcare, manufacturing, telecommunications and legal services. In some cases, the FBI said, intruders went from first targeting a victim to privileged access in under three hours.
Key Points
- The FBI issued FLASH advisory FLASH-20261113-001 on November 13, warning that hackers have weaponised ARTEX, an open-source AI-driven autonomous penetration-testing framework.
- Malicious actors have used it since at least August 2026 against organisations in eleven sectors, including finance, government, defence, healthcare and energy.
- Some intrusions reached privileged access in under three hours, the FBI said, citing Microsoft.
- ARTEX lets attackers with minimal skill run complex operations against many targets at once, and it is best at breaking into less-watched systems such as supplier portals.
- Linked attacks include intrusions at South Korean banks affecting about 67,000 people and the ConnectWise Cloud Backup compromise, which gave access to more than 2,250 downstream tenants.
What Happened
The FBI’s FLASH alert shares urgent threat intelligence with potential victims before more damage is done. It describes a tool built for legitimate security testing that has crossed into mainstream criminal and state-aligned use. ARTEX was designed to automate penetration testing — authorised probing to find a network’s weaknesses — but the FBI says hackers now use it without permission, and far faster than any human team.
The framework handles the whole attack chain on its own, from finding targets to taking data out. Attacks that once needed skilled operators working for days can now be launched by people with minimal expertise, against many targets at once. The tool is especially effective against the parts of a network that get the least attention: partner and supplier portals, mobile backends, legacy services and third-party integrations.
Background: From GitHub Project to Global Threat
ARTEX was written in Go and released on GitHub in July 2026 by Autumn-27, identified by The Record as Chinese security engineer Li Puhua. It gained 1,500 to 2,000 GitHub stars in ten weeks before the developer removed the repository on October 8. The removal changed little: forks and archived copies keep it available, and a commercial version, ARTEX PRO, is also in circulation. By early November, researchers had counted more than 1,600 exposed ARTEX consoles online.
Japanese media reported in November that a Chinese AI tool called Artex was used in several hacks behind Japan’s wave of data breaches, which prompted Tokyo to announce an emergency cybersecurity package. It is the clearest public confirmation yet that the tool has moved from a testing curiosity to a mainstream threat.
FBI ARTEX Warning: Why Speed Changes Everything
The most alarming detail in the advisory is the pace. Some intrusions went from initial targeting to privileged access in under three hours, the FBI said, citing Microsoft. For defenders, that compresses the window to spot and stop an attack from days to a single morning. Speed is only half the story; the other half is scale. Because ARTEX automates the work, one attacker can run complex operations against many targets simultaneously, probing dozens of organisations at once.
Attackers mostly plug Chinese AI models into the framework, led by DeepSeek’s Flash family, the FBI says, “due to their balance of capabilities, relatively low cost, and lack of observable guardrails”. The models are capable, cheap and largely unrestrained — exactly what an attacker wants. That turns hacking from a craft into an assembly line, and lowers the barrier to entry so far that skill no longer limits who can attack.
Reactions and Quotes
The FBI’s advisory describes attacks that run “at machine speed, from reconnaissance through data exfiltration”. That line captures the bureau’s concern: the tool does not just speed up one stage of an attack, it automates the entire sequence, from finding a target to stealing its data.
Attribution for the linked attacks comes from Microsoft, which tied the ConnectWise Cloud Backup supply-chain compromise to Storm-0940, a China-state-aligned group. That intrusion gave the attackers access to more than 2,250 downstream tenants, showing how one compromised supplier can expose thousands of victims at once.
Why It Matters
The warning matters because it changes who can launch a serious attack. When a powerful tool is open source, cheap to run and largely automated, the threat is no longer limited to well-funded hacking groups. Smaller outfits, and even lone actors, can now attempt intrusions that once required a team of specialists.
It also matters because of where ARTEX strikes. The systems it targets best — supplier portals, mobile backends, legacy services and third-party integrations — are the connective tissue of modern business. A bank, hospital or defence contractor can harden its own network and still be reached through a supplier’s forgotten portal; the South Korean bank intrusions exposed personal data on about 67,000 people.
What Happens Next
The threat is unlikely to fade. ARTEX remains available through forks, archived copies and the commercial ARTEX PRO version, and more than 1,600 exposed consoles were already counted online in early November. Removing the original repository did not remove the tool, and the bureau expects its use to grow.
Organisations in the named sectors should now scrutinise the systems ARTEX targets best: partner and supplier portals, mobile backends, legacy services and third-party integrations. Law enforcement and private researchers will watch for the next linked intrusions. The ConnectWise compromise and the South Korean bank breaches show the pattern — a supplier or a less-watched system, then a cascade of downstream victims.
Frequently Asked Questions
What is ARTEX? An open-source AI penetration-testing framework, written in Go and released on GitHub in July 2026, that automates the full attack chain from reconnaissance to data exfiltration.
What did the FBI warn? In FLASH advisory FLASH-20261113-001, issued November 13, the FBI warned that malicious actors have weaponised ARTEX since at least August 2026 to automate cyberattacks “at machine speed” in eleven sectors.
How fast are the attacks? Some intrusions reached privileged access in under three hours, the FBI said, citing Microsoft, and the tool lets attackers run operations against many targets at once.
Which attacks are linked to ARTEX? Intrusions at South Korean banks that exposed personal data on about 67,000 people, and the ConnectWise Cloud Backup compromise, attributed by Microsoft to the China-state-aligned Storm-0940, with access to over 2,250 downstream tenants.
Where did ARTEX come from? It was released on GitHub in July 2026 by Autumn-27, identified by The Record as Chinese security engineer Li Puhua. It gained 1,500 to 2,000 stars in ten weeks before the repository was removed on October 8, but forks, archived copies and ARTEX PRO keep it available.
Read more on US News Pulse: Japan unveils emergency cybersecurity package after wave of data breaches; Hackers hijack old Bitly links in Trump’s X posts, redirecting them to vulgar sites; and Bessent gives Nvidia 10 days to explain whether AI chips reached China.
Leave a Comment