According to security alerts from the FBI and Amazon Threat Intelligence, attackers are leveraging an open-source AI penetration testing framework named ARTEX to automate cyberattacks against banks, governments, energy and healthcare providers, and other critical sectors worldwide.
Between October 3 and November 5, Amazon’s MadPot honeypot threat sensors detected two distinct ARTEX campaigns. The first enabled threat actors to steal AI credentials at an “industrial scale” to gain unauthorized access to models, while the second utilized generated deepfake videos to bypass live selfie verification checks.
In the first campaign, malicious actors used AI agents to steal API keys, feeding them into the ARTEX framework’s model gateway to fuel further agents in search of additional AI model access keys.
According to a blog post by AWS Chief Information Security Officer CJ Moses, these agents validated identities through provider token and identity services, enumerated storage, assessed permission scopes, and targeted managed AI model services, secrets, parameter stores, and application identities.
Targeted entities included LLM gateways and proxy platforms—such as One API, New API, Veloera, Sub2API, and LiteLLM—as well as AI application builders like Dify, FastGPT, Langflow, RAGFlow, Open WebUI, LobeChat, and Ollama.
This campaign also utilized compromised cloud credentials to gain unauthorized access to the models.
In the second campaign, attackers utilized ARTEX to generate real-time deepfake videos designed to bypass identity verification (KYC) checks at two cryptocurrency exchanges, alongside forged ID photos to pass document validation checks.
Moses explained that the agents reverse-engineered a vendor’s liveness protocol in detail, documenting the face-detection event stream, how it tracks face position during the check, and the coordinate system used to guide user movement.
Although the agents successfully completed the full movement sequence with the face positioned within seven pixels of the target, the vendor’s server-side verification rejected the deepfake in every attempt.
In a separate session, the agents constructed a real-time deepfake video pipeline on a single consumer-grade GPU using an open-source face animation model. The pipeline achieved 11.5 frames per second, which Moses noted is “close to what a live video check expects.”
While the ultimate objective was cryptocurrency theft, the campaign demonstrates how ARTEX operates multiple sessions simultaneously, with various agents collaborating toward the same outcome. Moses noted that one agent searched for cash-out methods, another worked on clearing account balances, one set up the deepfake video pipeline, and another executed the verification bypass.
While none of the individual attack techniques are entirely new, both campaigns highlight how AI agents can significantly enhance attack speed and persistence.
Moses added that the agents succeeded where basic security hygiene was lacking, including the use of default or hardcoded credentials, keys left in code and public artifacts, unauthenticated interfaces and inference endpoints, and unpatched software. “Getting those fundamentals right remains the most effective defense,” he stated.
The FBI echoed this sentiment in a separate alert, noting that 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 federal agency cited August attacks against South Korean financial institutions that stole personal data belonging to approximately 67,000 individuals. Additionally, the FBI highlighted a September ARTEX campaign targeting ConnectWise Cloud Backup customers in a supply chain attack that ultimately compromised over 2,250 downstream Microsoft tenants.
Following the public disclosure of these security breaches, the original ARTEX developer removed the GitHub repository in October.
However, forked and localized variations, along with a commercial version, remain in use. As of November 4, internet scanning service Censys identified 1,465 hosts running an exposed ARTEX console. According to the FBI, the total number of deployments is “likely significantly higher when including systems behind firewalls or in private networks.”
The agency urged all organizations—particularly those operating internet-exposed applications, self-service portals, cloud infrastructure, and partner or supplier systems—to implement a series of mitigation measures outlined in its Flash alert.
These measures include reducing external attack surfaces by restricting sensitive services to trusted networks, private connectivity, or dedicated administrative paths, and moving enterprise applications behind zero-trust access policies and permissions.
Also Read
- Exploring the Scientific Implications of the Trump Administration’s AI and Quantum Focus
- U.S. Measles Outbreaks Lead to Loss of Elimination Status as Child Deaths Reach 10
- Microsoft’s New Remote Lock Lets IT Secure Lost or Stolen Windows Devices Without Erasing Data
- Decades of Cosmic Surveys Challenge the Standard Model of the Universe


