MITRE recently unveiled version 20 of its ATT&CK framework, introducing significant updates that encompass techniques for AI abuse, multi-factor authentication (MFA) bypass, and attacks targeting water utilities. This release reflects an expanded focus on recent adversary activities and tradecraft, marking a notable evolution in the framework’s capabilities.
The latest version includes nine new entries within the Enterprise domain, which comprise both techniques and sub-techniques. Notably, prompt injection is now categorized into two distinct techniques: one for malicious instructions directly input into a large language model (LLM) interface and another for instructions embedded within the model’s training data. This distinction highlights the growing sophistication of adversarial tactics aimed at manipulating AI systems.
Additionally, a new technique has been documented that involves attackers leveraging an organization’s own AI agents to exfiltrate sensitive data. Other sub-techniques focus on poisoning web content to influence AI-driven search and recommendation systems, as well as disrupting LLM-assisted malware analysis.
In the realm of authentication, a new technique tracks the theft of application-specific passwords, enabling attackers to bypass MFA prompts. Another technique addresses the evasion of CAPTCHA checks, detailing sub-techniques for both automatic resolution and avoidance of these security measures. These updates underscore the evolving landscape of cyber threats, particularly in how adversaries circumvent established security protocols.
In the industrial control systems (ICS) sector, a new campaign has been identified that targets internet-exposed programmable logic controllers (PLCs) across over 30 water and wastewater utilities in Minnesota. This attack has reportedly locked operators out of their systems, and while some sources suggest a potential link to Iranian-affiliated groups, MITRE has stated that attribution for these intrusions remains unconfirmed.
Another campaign documented in this release involves an AI-assisted breach affecting Mexican government institutions, which extended to a municipal water utility, further illustrating the intersection of AI technology and cyber threats in critical infrastructure.
Moreover, the ICS domain now includes the Command and Scripting Interpreter Technique, which features sub-techniques for six different interpreters, replacing the previously deprecated Command-Line Interface and Scripting techniques. Additional sub-techniques have been introduced to conceal command and control traffic within DNS and publish/subscribe protocols, enhancing the stealth capabilities of threat actors.
On the threat actor front, MITRE has added Nimbus Manticore, an espionage group associated with the Islamic Revolutionary Guard Corps (IRGC), which targets aerospace, defense, and telecommunications sectors. The entry for the China-backed group Salt Typhoon has been updated to reflect its exploitation of network edge devices, while North Korea’s APT38 has shifted its focus from financial theft to software supply chain attacks. Additionally, a new entry for The Gentlemen, a group operating a ransomware-as-a-service model, has been included.
To bolster defensive measures, MITRE has introduced 15 new detection strategies—six for the Enterprise domain and nine for ICS—supported by 38 analytics. The new ICS analytics are specifically designed to disregard routine engineering and maintenance activities, allowing for more effective monitoring of genuine threats.
In the mobile domain, MITRE has reinstated a technique for distributing malicious applications outside official app stores, a feature last seen in version 11. The update also includes NGate, a malware variant that relays payment card data via a phone’s NFC capabilities, enabling attackers to withdraw cash from ATMs.
These updates in MITRE ATT&CK v20 reflect a critical response to the evolving tactics employed by cyber adversaries, particularly in the context of AI and critical infrastructure. As organizations adapt to these new threats, the framework serves as an essential resource for understanding and mitigating risks in an increasingly complex cyber landscape.
