New Technique High TD-2026-0050 For SOC / IR, CISO, AppSec, AI Eng AI-assisted
Elastic: infostealer-stolen credentials used within 8 minutes; AI apps and signed RMM tools abused
Elastic Security Threat Command reports that attackers logged in with stolen credentials an average of eight minutes after infection across 27 confirmed infostealer cases. The report also documents misuse of AI coding assistants and ChatGPT desktop to run commands, and signed remote monitoring tools used to maintain access.
Why it matters
Infostealer-harvested passwords and session data let attackers reach cloud and internal accounts before defenders react. Trusted, signed or approved tools make malicious activity look like normal employee or IT activity.
Do this
- Hunt for ClickFix-style activity: users running copied commands from verification or troubleshooting pages, and unexpected shell or PowerShell child processes of browsers.
- Review which AI coding assistants and desktop AI apps have command execution and access to credential stores or internal documents; restrict or monitor them.
- Hunt for unapproved or unexpected ScreenConnect and other remote monitoring and management installs, especially those running as Windows services.
- Enforce phishing-resistant MFA and block legacy authentication protocols that accept passwords without a second factor.
- Review Microsoft Entra device-code sign-in flows and newly registered virtual devices; revoke sessions and tokens after suspected infostealer exposure.
Signals
- NoCISA KEV
- 34/100Priority
- 80%Confidence
Evidence
“Attackers used stolen credentials to log in an average of eight minutes after infection across 27 confirmed infostealer cases”
“These cases involved misuse after attackers obtained access. They do not establish that the applications contained a vulnerability or that their providers were compromised.”
“ClickFix-style verification pages were the most common entry point, appearing in 12 cases.”
Sources
Affected / related
Infostealer credential theft, ClickFix, Adversary-in-the-middle (AiTM) phishing, Device-code phishing, ScreenConnect, Claude Code, Cursor, ChatGPT desktop application
Published 2026-11-09 05:00 UTC · updated 2026-11-09 14:36 UTC · AI-assisted summary, checked against quoted sources · How we score