ThreatsDay: GPT-6.2 Jailbreak Flaws, BubbleLoader PNG Malware, $500M Ransom Demands + 17 More Stories
AI Summary
The article details multiple cyber threats, including the discovery and exploitation of jailbreak vulnerabilities in OpenAI's GPT-6.2 Astra models, the deployment of the BubbleLoader malware loader via malicious PNG files embedded in legitimate software, and a ransomware campaign by the North Korean group Kimsuky using a malicious Hangul document to deliver Funhouse ransomware. Additionally, a new AI-powered phishing kit called TheMailer uses generative AI to dynamically rewrite and test phishing emails until they bypass spam filters, enabling highly evasive social engineering attacks. These threats highlight the increasing sophistication of AI-driven attack techniques and ongoing malware distribution campaigns.
AI-extracted · verify before operational use
Extracted Entities 1 found
Indicators of Compromise 1 extracted
| Type | Value | Detail |
|---|---|---|
| Domain | recruitersglobal[.]net | Details → |