FBI warns ARTEX AI hacking tool is powering global cyberattacks
ARTEX is an autonomous penetration-testing framework written in Go. It uses AI agents to automate steps attackers may take during an intrusion, from reconnaissance and vulnerability discovery to exploitation, privilege escalation, and data theft. Its importance is the speed and scale it can bring to attacks, with little human intervention. Attackers give ARTEX goals and tasks, which it turns into commands for specialized agents. Rather than relying on its own AI model, it connects to external large language models, including Paradox, Axiom, Pramana, Cosmos, Vector, and Nyaya. The agents can work across multiple organizations at once. The FBI says some intrusions have moved from initial targeting to privileged access in under three hours. ARTEX was released publicly in July 2026, and its original GitHub repository was removed in October after reports of abuse. Copies, community forks, localized versions, and a commercial edition remain available.
What is ARTEX, and how does it use AI agents to automate cyberattacks?
ARTEX is an autonomous penetration-testing framework written in Go. It uses AI agents to automate steps attackers may take during an intrusion, from reconnaissance and vulnerability discovery to exploitation, privilege escalation, and data theft. Its importance is the speed and scale it can bring to attacks, with little human intervention.
Attackers give ARTEX goals and tasks, which it turns into commands for specialized agents. Rather than relying on its own AI model, it connects to external large language models, including Paradox, Axiom, Pramana, Cosmos, Vector, and Nyaya. The agents can work across multiple organizations at once.
The FBI says some intrusions have moved from initial targeting to privileged access in under three hours. ARTEX was released publicly in July 2026, and its original GitHub repository was removed in October after reports of abuse. Copies, community forks, localized versions, and a commercial edition remain available.
How quickly can an ARTEX intrusion reach privileged access, and how many Microsoft 365 tenants did attackers try to reach?
The FBI says some ARTEX intrusions progressed from initial targeting to privileged access in under three hours. That pace matters because defenders may have little time to investigate alerts and respond before attackers gain deeper access. ARTEX can also support operations against multiple organizations simultaneously.
In the ConnectWise campaign, attackers first compromised application identities used by the cloud backup service. Microsoft attributed the campaign to China-state-aligned actor Storm-0940, which used ARTEX to accelerate reconnaissance, password spraying, and abuse of cloud application permissions. After compromising the identities, attackers attempted authentication against more than 2,250 downstream Microsoft 365 tenants.
Those figures describe different stages: under three hours is the reported speed of some intrusions, while more than 2,250 is the number of tenants targeted for authentication in this campaign. Huntress reported at least 162 compromised tenants. The FBI warns that automated attacks can create suspicious traffic, but their speed can still leave organizations little time to act.
What could attackers do after compromising the application identities used by ConnectWise's cloud backup service?
The ConnectWise incident shows how attackers could misuse application identities belonging to a cloud backup service. After compromising those identities, they could authenticate to customer Microsoft 365 tenants using permissions that had already been granted to the applications. This exposed customer environments to unauthorized access.
ConnectWise’s cloud backup service was formerly known as SkyKick Cloud Backup. Microsoft attributed the campaign to Storm-0940, a China-state-aligned threat actor. The attackers used ARTEX to speed up reconnaissance, password spraying, and abuse of cloud application permissions. They then attempted authentication against more than 2,250 downstream tenants.
Huntress reported identifying at least 162 compromised tenants at the time. The source does not specify what data or services attackers accessed in each tenant, so the confirmed consequence is unauthorized access, not a particular theft or change. The case illustrates why organizations should review application permissions and authentication logs, especially when applications can access cloud customer environments.
How can a breach at one cloud service provider put many of its customers at risk?
A cloud provider’s applications may have permissions to access customer environments. If attackers compromise the identities those applications use, they may be able to authenticate to customer tenants with permissions already granted. That creates a path from one provider-side breach to risk across multiple customers.
In the ConnectWise Cloud Backup incident, attackers compromised application identities used by the service, formerly known as SkyKick Cloud Backup. They then attempted authentication against more than 2,250 downstream Microsoft 365 tenants. Huntress reported at least 162 compromised tenants. Microsoft attributed the campaign to Storm-0940, which used ARTEX to speed up reconnaissance, password spraying, and abuse of cloud application permissions.
This example shows why organizations should review application permissions and authentication logs, including activity involving third-party services. The FBI also recommends securing internet-facing applications and restricting access to cloud administration interfaces. The article does not say that every customer of a breached provider is affected; it describes a campaign in which attackers targeted many tenants, with some reported as compromised.
Why can overlooked systems, such as partner portals and old applications, become entry points for attackers?
The FBI says ARTEX is particularly effective at identifying vulnerabilities in overlooked systems. Examples include partner portals, employee applications, legacy services, and third-party integrations. These systems can become entry points when they contain weaknesses that attackers can discover and exploit.
ARTEX automates reconnaissance and vulnerability discovery, then can move on to exploitation and privilege escalation. That lets attackers examine multiple targets and chain techniques quickly. The FBI says some intrusions reached privileged access in under three hours, so even an overlooked application can matter if it provides a way into a wider environment.
The source does not explain why each system was overlooked or claim that all such systems are vulnerable. It does warn that these types of systems can be effective targets for ARTEX. Organizations can reduce risk by securing internet-facing applications, reviewing application permissions and authentication logs, and watching for unusual bursts of automated requests or unexpected credential and privilege changes.
What can organizations monitor or change to spot automated attacks and limit the damage?
Organizations can look for signs that activity is automated or unusually large. The FBI recommends monitoring for bursts of automated requests, suspicious cloud activity, unexpected credential or privilege changes, and large data transfers. Such activity may trigger security alerts, although ARTEX’s speed can leave defenders little time to respond.
Defenders can also limit opportunities for attackers. The FBI advises securing internet-facing applications, restricting access to cloud administration interfaces, reviewing application permissions and authentication logs, and applying rate limits to sensitive APIs. These steps address exposure and access, including risks involving cloud applications and third-party integrations.
The FBI’s advisory includes indicators of compromise such as IP addresses, command-and-control infrastructure, and suspicious network activity. MalwareMustDie also released a detection signature pack and playbook covering network requests, identity abuse, cloud activity, malicious files, and automated attack patterns. Its playbook sorts indicators by confidence. The FBI cautions that IP addresses may change, while the researcher warns infrastructure indicators can become outdated or cause false positives.
What is penetration testing, and how is authorized security testing different from an attack?
Penetration testing is a planned assessment in which security testers try to find and demonstrate weaknesses in systems. This definition comes from established cybersecurity practice; the source article describes ARTEX as a penetration-testing framework but does not define the term. Such testing is intended to help an organization understand and address security risks.
Authorization is the key distinction. In an authorized test, the organization has agreed to the testing, and the work is limited to approved systems and conditions. An attack lacks that permission. The article describes threat actors using ARTEX to target organizations, exploit systems, escalate privileges, and steal data, rather than conducting approved tests.
ARTEX can automate reconnaissance, vulnerability discovery, exploitation, and other steps with little human intervention. The FBI says threat actors have used it since at least August 2026 and that some intrusions reached privileged access in under three hours. The framework’s testing capabilities do not make every use authorized; the user’s permission and the activity’s scope matter.
This brief was written by AI from the original reporting and checked by other models. Names, figures and quotes come from the source; read it for full context.
Read more in the JupiteX app
Pulse is free. New stories every 4 hours, each one broken into the questions that explain it.
Or read more news on the web