Threat video1:00
ChatGPT Lockdown Bypass via HTML Preview: Detection & IOCs
Nested iframes let ChatGPT's HTML preview open attacker popups past Lockdown Mode. 9 detection rules, 8 IOCs and a free Splunk SPL rule.
Chapters
- 0:00The hook4 attack scenarios bypass Lockdown Mode protection.
- 0:04The storyNested iframes escaped the sandbox completely.
- 0:11Attack pathAccording to Check Point, nested iframes escaped the preview sandbox.
- 0:28AttributionNo threat actor is attributed.
- 0:33Indicators of compromise1 indicator to hunt: 1 domain.
- 0:38MITRE ATT&CK techniques7 ATT&CK techniques, from initial access to collection.
- 0:43Remediation3 steps now: Confirm ChatGPT web and ChatGPT Classic clients are current.
- 0:50Detection rules9 detection rules, in SPL, KQL and Sigma.
Transcript
0:004 attack scenarios bypass Lockdown Mode protection.
0:05Nested iframes escaped the sandbox completely.
0:08Researchers disclosed it in November.
0:11According to Check Point, nested iframes escaped the preview sandbox.
0:16They stole chat history, harvested credentials, malware as updates.
0:20Lockdown Mode's protection was completely bypassed.
0:24OpenAI patched it September 18.
0:29No threat actor is attributed. Confidence: low.
0:331 indicator to hunt: 1 domain.
0:397 ATT&CK techniques, from initial access to collection.
0:433 steps now: Confirm ChatGPT web and ChatGPT Classic clients are current.
0:519 detection rules, in SPL, KQL and Sigma.
0:56Threadlinqs. The unified security engineering platform.