ThreatsDay: GPT-6.2 Jailbreak Flaws Exposed, BubbleLoader PNG Malware, $500M Ransom Operations and Broader AI-Driven Threats

This week's ThreatsDay bulletin highlights how everyday lures such as job offers and service outage messages are being weaponized alongside advanced AI techniques to steal data and deploy malware. OpenAI's GPT-6.2 Astra and Sol models show major gains in cyber capabilities but suffered a documented regression in jailbreak robustness that the U.K. AI Security Institute exploited in half the queries needed for prior versions. Google introduced Gemini 4 Neon with improved secure-coding performance on AutoBaxBench and SEC-bench Pro while extending Gemini chat retention to two months for safety reasons. A $500 million ransomware campaign linked to the 8-Digits Team resulted in the extradition of Roman Khlynovskiy, while Kimsuky deployed Funhouse ransomware and GravityRAT targeted Indian defense job seekers via fake recruitment portals. Additional incidents include Infoblox's third-party SharePoint breach, TheMailer AI phishing kit, and EU sanctions against Doppelgänger operators.

The Hacker NewsTranslated from English

Read at source

This week's ThreatsDay bulletin demonstrates how ordinary-looking lures such as job offers, service outage alerts, and familiar applications are being used to steal data, spread malware, and obtain unauthorized access. AI tools are simultaneously lowering the barrier for attackers while creating new defensive and offensive challenges for defenders.

OpenAI released GPT-6.2 Astra and GPT-6.2 Sol, reporting substantial improvements in vulnerability discovery, secure code generation, and patch quality. The U.K. AI Security Institute disclosed a universal jailbreak that succeeded against pre-release versions of GPT-6.2 Astra in half the queries required for GPT-6 Astra, prompting OpenAI to remediate the regression before release.

Google unveiled Gemini 4 Neon, which achieved state-of-the-art results on hard-tier challenges in AutoBaxBench (76%) and SEC-bench Pro (90.5%) while delivering faster and more cost-efficient secure coding than previous frontier models. The company also extended retention of temporary Gemini chats to approximately two months to enhance safety monitoring.

Roman Khlynovskiy, an alleged member of the 8-Digits Team ransomware group, was extradited from Italy to the United States to face charges related to attacks on schools, hospitals, and healthcare entities that generated over $500 million in ransom demands. The group has been linked to INC and Lynx ransomware operations across at least six U.S. states.

North Korean actor Kimsuky distributed the Funhouse ransomware variant through malicious Hangul documents disguised as resumes, reusing tools such as Quasar RAT observed in earlier campaigns. In parallel, Pakistani groups SpaceCobra and Cosmic Leopard used the GravityRAT Android surveillance tool to target Indian defense-sector job seekers via a fake recruitment portal.

Additional developments include WordPress 7.1.5 security updates, Infoblox's disclosure of a third-party SharePoint data breach, the TheMailer AI-powered phishing kit capable of rewriting lures until they bypass filters, and EU sanctions against the Doppelgänger disinformation operation and NoName057(16) hacktivist group.

Meridian digestThe week's most consequential security stories, distilled into one briefing.See the digest

More in AI & Model Risk

Researchers Show Malicious Webpage Could Hijack Meta Muse AI Agent and Extract Private Conversations
China-Linked Actors Use Autonomous AI Agents to Discover Zero-Day Path Traversal in MONAI Label and Achieve Root Access
AI & Model Risk

China-Linked Actors Use Autonomous AI Agents to Discover Zero-Day Path Traversal in MONAI Label and Achieve Root Access

Suspected China-linked hackers deployed autonomous AI agents powered by the open-source Cairn framework to identify and exploit a previously unknown path traversal vulnerability in MONAI Label, an open-source medical imaging server used by research teams. The flaw, tracked as CVE-2026-133968 and rated 9.8, affects versions 0.8.5 and earlier by allowing unauthenticated attackers to write files outside the designated storage directory when authentication is disabled by default. Amazon Threat Intelligence reconstructed the campaign after discovering an exposed Cairn control panel that revealed the agents' prompts, task settings, and execution history spanning roughly 48 hours in September. Within about 11 minutes the agents mapped the application and achieved arbitrary file writes; in a follow-up session they planted a malicious Python startup file that was loaded during a training job, granting root access on a university server. Maintainers merged a source-code fix on October 19 that enforces path resolution inside approved folders, yet no patched release has been published and the latest available version remains 0.8.5 from 2024. Amazon attributes the activity to a China-linked operator with moderate confidence based on Simplified Chinese prompts and China Standard Time settings, though no specific group or government sponsorship was confirmed.

The Hacker NewsNov 21, 2026Translated from English

Chinese-Linked Actors Use Cairn AI Platform to Discover and Exploit MONAI Label Zero-Day