FBI warns AI hacking tool ARTEX is accelerating cyberattacks worldwide
The framework automates familiar intrusion techniques. Its attacks can trigger security alerts, but the FBI says their speed leaves defenders little time to respond.
Loading page…
The framework automates familiar intrusion techniques. Its attacks can trigger security alerts, but the FBI says their speed leaves defenders little time to respond.
Listen to this story
The FBI’s November 13 FLASH advisory links ARTEX, an AI-assisted penetration-testing framework, to intrusions by multiple threat actors since at least August 2026. By coordinating specialized agents through external language models, the tool can automate reconnaissance, exploitation and privilege escalation across multiple organizations, leaving defenders little time to respond. The cited incidents show attackers exploiting trusted cloud-app permissions and a former vendor’s unrevoked credential. The advisory gives organizations indicators to investigate and recommends reviewing application permissions, authentication logs and unusual automated activity.
ARTEX connects to models from Anthropic, OpenAI, DeepSeek, GLM, Grok and Kimi rather than supplying its own model.
Microsoft attributed the ConnectWise Cloud Backup campaign to China-state-aligned Storm-0940 and said ARTEX accelerated reconnaissance and password spraying.
Barracuda said an exposed credential for a former third-party service opened a route to customer-data collection credentials; it changed all customer credentials.
Hackers using an AI security-testing tool can move from choosing a target to gaining elevated access in under three hours, the FBI warns. A FLASH advisory issued November 13 describes ARTEX-assisted attacks worldwide and gives defenders signs to investigate. The danger is not necessarily stealth: these attacks can be conspicuous and still move too quickly.
The bureau says multiple threat actors have weaponized ARTEX since at least August 2026. Targets span financial services, government, technology, defense, healthcare and other industries. Its warning brings together previously reported intrusions, including the ConnectWise Cloud Backup attack, with technical indicators and guidance for organizations checking their own systems.
ARTEX is a penetration-testing framework: software designed to probe systems for security weaknesses. Rather than supply its own AI model, it connects to external language models, including those from Anthropic and OpenAI, as well as DeepSeek, GLM, Grok and Kimi. The framework turns an operator’s goals into commands carried out by specialized AI agents.
Those agents can investigate targets, find and exploit weaknesses, obtain higher permissions and steal data with little human intervention. Attackers can direct activity against multiple organizations simultaneously. The FBI says the framework is particularly effective at finding weaknesses in overlooked systems, including partner portals, employee applications, older services and third-party integrations.
That activity can generate large amounts of suspicious traffic and set off defensive alerts. But visibility is not the same as enough time to intervene. The bureau warns that ARTEX compresses the work of chaining attack techniques together, allowing some intrusions to reach privileged access—permissions beyond those of an ordinary user—in 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
The ConnectWise case shows how that automation can operate through existing trust between companies. In the October Cloud Backup attack cited by the FBI, attackers compromised application identities used by the backup service. They then used permissions customers had already granted those applications to gain unauthorized access to Microsoft 365 environments.
Microsoft attributed that campaign to Storm-0940, a China-state-aligned threat actor. It said ARTEX accelerated reconnaissance, password spraying—trying passwords across many accounts—and abuse of cloud application permissions. Some intrusions progressed to credential changes and attempts to establish persistent access, according to BleepingComputer’s account of Microsoft’s investigation.
Barracuda disclosed another ARTEX-linked attack involving an exposed credential for a third-party service. The company had not revoked it after ending work with the vendor. That access opened a route to stores of credentials Barracuda used to collect data from customer systems. Barracuda said its logs were not detailed enough to rule out access and changed all customer credentials.
ARTEX first appeared on GitHub in July 2026. Its developer, known as Autumn-27, removed the repository on October 8 after reports of abuse. The developer subsequently made the software closed-source, but the FBI says the earlier open-source version remains widely accessible through copies and community forks. Removing the original repository did not withdraw those versions.
For defenders, the FBI’s 13-page advisory supplies indicators of compromise: technical traces that may point to an intrusion. These include IP addresses, infrastructure used to control attacks and suspicious network activity. The bureau cautions that listed IP addresses may have changed since the attacks occurred, limiting their usefulness as a permanent blocklist.
Organizations that identify suspicious activity associated with ARTEX are encouraged to report it to the FBI or its Internet Crime Complaint Center. The advisory pairs its warning about faster attacks with concrete traces organizations can investigate now.
Loading discussion...
Join the conversation
Explain when restricting access helps, or when openness is worth preserving.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.