Threat reportThreat IntelligenceTL-2026-3719
MITRE ATT&CK v20 Adds 9 Enterprise Techniques Covering AI Exploits, MFA Bypass, and CAPTCHA Evasion
MITRE ATT&CK v20 Adds 9 Enterprise Techniques Covering AI (TL-2026-3719), also tracked as ATT&CK v20, is a info-severity tracked intrusion set, first published 2026-10-16. It has no confirmed attribution, affects MITRE ATT&CK Framework (Enterprise, ICS, Mobile), maps to 6 MITRE ATT&CK / ATLAS techniques (AML.T0051.000, AML.T0051.001, T1111), and is covered by 9 detection rules and 2 indicators of compromise.
- Severity
- INFOAssessed severity
- CVEs
- 0None referenced
- Techniques
- 6MITRE ATT&CK / ATLAS
- Actors
- 0Not attributed
- Detection rules
- 9SPL · KQL · Sigma
- IOCs
- 2Indicators of compromise
Key facts for TL-2026-3719
- Threat ID
- TL-2026-3719
- Also known as
- ATT&CK v20, MITRE ATT&CK October 2026 release
- Severity
- INFO
- Status
- ACTIVE
- Category
- THREAT_INTEL
- First published
- Last reviewed
- Attribution confidence
- LOW
- Motivation
- UNKNOWN
- Target sectors
- technology, government administration, finance, industrial, telecoms
- Target regions
- Global
- Detection rules
- 9
- Indicators of compromise
- 2
Malware and tooling in MITRE ATT&CK v20 Adds 9 Enterprise Techniques Covering AI
Malware and tooling: NGate
How MITRE ATT&CK v20 Adds 9 Enterprise Techniques Covering AI works
MITRE released ATT&CK v20 on October 13, 2026, adding nine Enterprise techniques: five AI-related (Direct Prompt Injection, Indirect Prompt Injection, AI-Enabled Exfiltration, AI Result Poisoning, Impair Autonomous Defenses) plus Steal Application Specific Passwords (T1697) and Anti-Automation Evasion (T1696, with T1696.001 and T1696.002). The release also updates ICS and Mobile and adds new detection strategies and analytics.
This is a framework and taxonomy release, not an exploited vulnerability or an attributed campaign. No CVE, CVSS score, IOC, malware family or actor attribution is stated in the sources, and no active exploitation is reported. It is tracked because the new technique IDs change detection-coverage mapping for defenders.
ATT&CK v20 was released on October 13, 2026. The Enterprise matrix gains nine techniques. T1699 (Direct Prompt Injection) covers malicious instructions submitted directly to an AI system's interface. T1698 (Indirect Prompt Injection) covers malicious instructions embedded in content an AI system processes. T1700 (AI-Enabled Exfiltration) covers directing enterprise AI agents to extract sensitive organizational data. T1608.007 (Stage Capabilities: AI Result Poisoning) covers manipulating web content to influence AI search results and recommendations. T1685.007 (Disable or Modify Tools: Impair Autonomous Defenses) covers interfering with AI-assisted security tools to disrupt malware analysis. T1697 (Steal Application Specific Passwords) covers obtaining app-specific credentials to bypass standard MFA requirements. T1696 (Anti-Automation Evasion) covers circumventing bot-detection mechanisms, with sub-techniques T1696.001 (Automate CAPTCHA Completion) and T1696.002 (Avoid CAPTCHA).
The release adds new detection strategies and analytics. One source reports 15 new detection strategies and 38 analytics. The official MITRE update page reports 6 new Enterprise detection strategies for AI-related threats, 9 new ICS detection strategies for command and scripting abuse, and over 100 new ICS analytics plus dozens of Enterprise analytics. These counts differ between sources and are recorded as reported. ICS gains seven new Command and Scripting Interpreter techniques (T1701 per the MITRE page) covering PowerShell, Python, Unix Shell, JavaScript, Visual Basic and Windows Command Shell. DNS and Publish/Subscribe protocol variants are added to Standard Application Layer Protocol. On Mobile, 'Deliver Malicious Application via Other Means' (T1476) was un-deprecated (v2.0) after last appearing in v11, covering direct downloads, SMS, messaging apps, QR codes and third-party stores. NGate malware was added and 28 existing mobile entries were revised. The release encompasses 971 software entries, 182 groups and 61 campaigns across domains.
Mapping note: the platform's local ATT&CK catalog is v19.2 and does not yet contain the v20 IDs. The mitre[] entries therefore use the existing parent techniques (T1608, T1685, T1111, T1567) and ATLAS prompt-injection IDs. The v20 IDs are recorded as technique IOCs and tags and should be added to the catalog before detection-coverage mapping.
MITRE ATT&CK / ATLAS techniques used in TL-2026-3719
Execution
AML.T0051.000 LLM Prompt Injection: Direct; AML.T0051.001 LLM Prompt Injection: Indirect
Credential Access
T1111 Multi-Factor Authentication Interception
Exfiltration
T1567 Exfiltration Over Web Service
Resource Development
Defense Impairment
Affected products and versions in MITRE ATT&CK v20 Adds 9 Enterprise Techniques Covering AI
- MITRE — ATT&CK Framework (Enterprise, ICS, Mobile)
Fixed in: v20 (October 13, 2026)
Remediation for MITRE ATT&CK v20 Adds 9 Enterprise Techniques Covering AI
Immediate actions
- Update internal ATT&CK catalogs and detection-coverage maps to v20 technique IDs (T1696, T1696.001, T1696.002, T1697, T1698, T1699, T1700, T1608.007, T1685.007)
- Review which enterprise AI assistants and agents can access sensitive data, as relevant to T1700 AI-Enabled Exfiltration
Workarounds
- Restrict or disable application-specific passwords where modern MFA-capable authentication is available
- Limit AI agent tool permissions and data scope to reduce the impact of direct and indirect prompt injection
Longer-term hardening
- Adopt the new v20 detection strategies and analytics for AI-enabled exfiltration, anti-automation evasion and application-specific password abuse
- Add prompt-injection scenarios to threat models and purple-team exercises for AI-integrated systems
- Review ICS detection coverage against the new Command and Scripting Interpreter sub-techniques
Timeline of MITRE ATT&CK v20 Adds 9 Enterprise Techniques Covering AI
- ATT&CK v19 released, splitting Defense Evasion into Stealth (TA0005) and Defense Impairment (TA0112) and adding ICS sub-techniques.
- New detection strategies and analytics published, including AI-enabled exfiltration, anti-automation evasion, application-specific password abuse and ICS command interpreter misuse.
- Mobile technique 'Deliver Malicious Application via Other Means' (T1476) un-deprecated; NGate malware added and 28 mobile entries revised.
- ICS matrix gains seven Command and Scripting Interpreter techniques (PowerShell, Python, Unix Shell, JavaScript, Visual Basic, Windows Command Shell) and DNS and Publish/Subscribe variants of Standard Application Layer Protocol.
- MITRE releases ATT&CK v20 with nine new Enterprise techniques covering AI abuse, MFA bypass and anti-automation evasion.
- Cyber Security News publishes its summary of the v20 release, which triggered tracking on the platform.
Sources cited for MITRE ATT&CK v20 Adds 9 Enterprise Techniques Covering AI
Detection coverage for TL-2026-3719
As of 2026-10-16, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-3719 across Splunk SPL, Microsoft KQL and Sigma, covering 2 indicator(s) of compromise. The whole corpus is readable without an account; a free account unlocks full detection query text in Splunk SPL, Microsoft KQL and Sigma; paid tiers add raw indicator values, correlation and the MCP server. Threadlinqs MCP server · View plans.