MITRE has released version 20 of its ATT&CK framework, broadening coverage of adversary tradecraft that targets AI systems, multi-factor authentication, and operational technology environments. The organization describes the update’s main focus as expanded documentation of recent attacker activity.
AI and Authentication Bypass
The Enterprise domain gains nine new techniques and sub-techniques. Prompt injection is now split into two distinct techniques: one covering malicious instructions entered directly into an LLM interface, and another for instructions hidden inside content the model processes. A separate technique addresses attackers abusing an organization’s own AI agents to exfiltrate data, while new sub-techniques cover poisoning web content to manipulate AI search and assistant outputs, and interfering with LLM-assisted malware analysis.
On the authentication side, a new technique tracks theft of application-specific passwords, which let attackers log in without triggering an MFA prompt. Another new technique covers CAPTCHA bypass, with sub-techniques for automated solving and for avoidance. The existing sub-technique on compromised software dependencies now also accounts for tampered IDE and coding agent configuration files.
Water Utility and ICS Updates
The ICS domain adds a campaign tracking attacks on internet-exposed PLCs at more than 30 water and wastewater utilities in Minnesota, which left operators locked out of their own systems. MITRE notes that attribution for these intrusions remains unconfirmed, despite reports linking the broader wave of activity to Iranian-affiliated actors. A second new campaign documents an AI-assisted breach of Mexican government institutions that spanned IT and OT environments and extended to a municipal water utility.
ICS also gains a Command and Scripting Interpreter technique with six interpreter-specific sub-techniques, replacing deprecated Command-Line Interface and Scripting entries. Two additional sub-techniques cover command and control traffic concealed in DNS and publish/subscribe protocols.
Threat Actors and Detection
New threat actor entries include Nimbus Manticore, an IRGC-associated group targeting aerospace, defense, and telecom sectors, along with The Gentlemen, a ransomware-as-a-service operation. Existing entries were updated too: Salt Typhoon’s profile now reflects exploitation of network edge devices, and North Korea’s APT38 entry reflects a pivot from financial theft toward software supply chain attacks.
Defenders gain 15 new detection strategies, backed by 38 analytics, with new ICS analytics designed to filter out routine engineering and maintenance activity. On Mobile, MITRE reinstated a technique for delivering malicious apps outside official app stores and added NGate, malware that relays payment card data over NFC to enable fraudulent ATM withdrawals.
