Skip to content
[MED] Data Breach #data-breach

FBI Warns Hackers Use ARTEX AI Tool Against Organizations

By Rockstar Media Editorial Original report: BleepingComputer ↗

┌─ fbi-warns-artex-ai-cyberattacks.md :: hero

Illustrative stock photo for: FBI Warns Hackers Use ARTEX AI Tool Against Organizations
Illustrative stock photo for: FBI Warns Hackers Use ARTEX AI Tool Against Organizations Photo by Brett Sayles on Pexels

The FBI has warned that threat actors are using an AI-powered penetration testing framework called ARTEX to automate cyberattacks against organisations worldwide. In a FLASH advisory released Friday, the bureau said multiple groups have weaponised the open-source tool since at least August 2026 across financial, government, technology, defence, and healthcare sectors, including an October ConnectWise Cloud Backup supply-chain breach affecting Microsoft 365 customers. Some intrusions moved from initial targeting to privileged access in under three hours. Because ARTEX targets organisations globally, Malaysian businesses should review internet-facing systems and report suspicious activity to Cyber999 (MyCERT).

Key Facts

Attribute Value
Exploited in the wild Yes — weaponised in global cyberattacks since at least August 2026
Threat actor Multiple threat actors, including China-state-aligned Storm-0940 (attributed by Microsoft), Chinese cyberespionage group Vapor Panda (reported by CrowdStrike), and a Chinese-speaking threat actor
Disclosure date November 15, 2026 (BleepingComputer report); FBI FLASH advisory released Friday

How Does ARTEX Automate Cyberattacks?

ARTEX is an autonomous penetration testing framework written in Go that connects to external large language models to run cyberattacks with minimal human oversight. Attackers supply goals and tasks, and specialised AI agents convert them into commands for reconnaissance, vulnerability discovery, exploitation, privilege escalation, and data theft.

Rather than relying on a built-in AI model, ARTEX integrates with external models including Anthropic Claude, OpenAI GPT, DeepSeek, GLM, Grok, and Kimi. The FBI noted that the framework is particularly effective at finding flaws in overlooked assets such as partner portals, employee applications, legacy services, and third-party integrations, while enabling simultaneous multi-target operations.

"ARTEX significantly reduces the time required to chain attack techniques, with some intrusions progressing from initial targeting to privileged access in under three hours. The framework enables threat actors with minimal technical expertise to conduct complex, multi-target operations simultaneously, fundamentally accelerating attack velocity and expanding operational scale."

FBI FLASH advisory, quoted by BleepingComputer

Although ARTEX's automated scanning generates high volumes of suspicious traffic that can trigger security alerts, its speed leaves defenders little time to intervene before a breach occurs.

Which Threat Actors and Organisations Have Been Hit?

The FBI advisory links ARTEX to multiple campaigns, including an October ConnectWise Cloud Backup supply-chain attack affecting Microsoft 365 customers, breaches at South Korean banks exposing data of roughly 67,000 people, and cyberespionage targeting over 200 organisations across 14 countries, attributed to groups such as Storm-0940 and Vapor Panda.

In the October ConnectWise Cloud Backup (formerly SkyKick Cloud Backup) incident, attackers compromised cloud backup application identities to authenticate into customer tenants. Huntress identified at least 162 compromised tenants, while Microsoft attributed the campaign to a China-state-aligned actor tracked as Storm-0940, reporting attempted authentication against more than 2,250 downstream Microsoft 365 tenants, with some progressing to credential modifications and persistence attempts.

In South Korea, CrowdStrike found exposed infrastructure from a Chinese-speaking threat actor combining ARTEX with Claude Code agents against Shinhan Bank, KB Kookmin Bank, and Hana Bank. Separately, South Korean police arrested seven suspects linked to intrusions at over 90 organisations—including hospitals, educational institutions, and businesses—and the theft of approximately 8.2 million data records, with ARTEX reportedly used in most cases. CrowdStrike also reported that Chinese cyberespionage group Vapor Panda used ARTEX against more than 200 organisations across 14 countries, compromising at least nine.

Publicly reported incidents show ARTEX has been used in both financially motivated data theft and state-aligned cyberespionage across multiple regions.

Is ARTEX Still Available and Detectable?

Yes, ARTEX remains accessible despite its developer, known as Autumn-27, removing the original GitHub repository on October 8 after abuse reports surfaced. The framework continues to circulate through community forks, archived copies, localised versions, and a commercial edition called ARTEX PRO, with over 1,600 publicly exposed consoles identified in early November.

Indicators of compromise, including command-and-control infrastructure and suspicious network activity, are published in the original 13-page FBI advisory, though the bureau cautions that listed IP addresses may have changed since the attacks occurred. Researchers believe the actual number of ARTEX deployments is significantly higher than the exposed consoles found online.

On Friday, cybersecurity researcher MalwareMustDie also published an open-source detection signature pack and playbook covering suspicious network requests, identity abuse, cloud activity, malicious files, and automated attack patterns, categorising indicators by confidence level to warn against outdated infrastructure indicators and false positives.

Removing the primary GitHub repository did not halt ARTEX's spread, leaving defenders reliant on behavioural monitoring and current threat indicators.

What Should You Do?

Because ARTEX is an attack framework rather than a single software flaw, no single vendor patch resolves this threat; organisations should apply the hardening and monitoring measures recommended by the FBI:

  • Secure internet-facing applications: audit and lock down partner portals, employee applications, legacy services, and third-party integrations.
  • Restrict cloud administration access: limit access to cloud administration interfaces and review granted application permissions.
  • Enforce API rate limits: implement rate limits on sensitive APIs to slow automated reconnaissance and credential attacks.
  • Monitor logs and anomalies: review authentication logs for unusual bursts of automated requests, suspicious cloud activity, unexpected credential or privilege changes, and large data transfers.
  • Report suspected intrusions: share findings with the FBI or its Internet Crime Complaint Center (IC3), and in Malaysia report security incidents to Cyber999 (MyCERT).

Common Questions

Does ARTEX use its own artificial intelligence model?

No. ARTEX does not rely on an in-house model; instead, the Go-based framework integrates with external large language models including Anthropic Claude, OpenAI GPT, DeepSeek, GLM, Grok, and Kimi to convert attacker goals into automated commands.

How were Microsoft 365 environments compromised in the ConnectWise breach?

Attackers compromised application identities belonging to ConnectWise Cloud Backup, formerly SkyKick Cloud Backup, then used previously granted permissions to authenticate into customer tenants. Microsoft attributed the campaign to Storm-0940, which attempted logins against over 2,250 downstream tenants.

Should defenders rely solely on IP addresses from the FBI advisory?

No. The FBI cautions that IP addresses in its advisory may have changed since the attacks took place, and researcher MalwareMustDie notes that infrastructure indicators can become outdated or cause false positives, making behavioural and log monitoring essential.

Sources and Methodology

This dispatch is based solely on reporting by BleepingComputer, published on November 15, 2026.

$ ls ./tags

  • #Data Breach
  • #AI Security

From the network