Cybersecurity & Exploits

Global: SecureCAI Framework Cuts Prompt Injection Success in Security‑Focused Language ModelsResearchers have unveiled SecureCAI, a defense framework designed to protect large language models (LLMs) used in Security Operations Centers from…

Global: SecureDyn-FL Boosts IoT Intrusion Detection with Privacy-Preserving Federated LearningBackgroundA team of researchers has introduced SecureDyn-FL, a federated learning framework designed to improve intrusion detection across diverse Internet of Things…

Global: AI-Driven System A1 Shows High Success Rate in Generating Profitable Smart Contract ExploitsA new AI-powered system called A1 has demonstrated the ability to autonomously discover and exploit vulnerabilities in…

Global: TabPFN Outperforms Traditional Ensembles in Low-Data Malware Detection, Study Finds A research paper posted to arXiv on Jan. 12, 2026 reports that a learning‑free model called TabPFN delivers higher…

Global: Truebit Smart Contract Exploit Leads to $26.4 Million LossAn attacker exploited a vulnerability in an Ethereum‑based smart contract, resulting in the theft of 8,535 ETH, valued at approximately $26.4 million. The…

USA: NIST’s CAISI Seeks Input on Securing AI Agent SystemsThe U.S. Department of Commerce’s National Institute of Standards and Technology (NIST) has issued a Request for Information (RFI) on January…

Global: Agentic AI Redefines Cybersecurity Landscape with Dual-Use PotentialA new survey posted on arXiv outlines how autonomous AI systems are reshaping both defensive and offensive cybersecurity operations, highlighting the technology’s…

Global: Study Presents Multi-turn Jailbreak Attacks and FragGuard Defense for Multi-modal LLMsResearchers have unveiled a comprehensive framework, MJAD-MLLMs, that systematically examines both multi-turn jailbreaking attacks and corresponding defense strategies for…

Global: New Global Honeynet Dataset Offers High-Resolution Insight into Cyberattack BehaviorResearchers have released a high‑resolution honeynet dataset captured over a continuous 72‑hour period from June 9 to 11, 2025, on…

Global: Adaptive Jailbreak Attack Highlights Weaknesses in LLM Safety FiltersResearchers have unveiled iMIST, an interactive multi‑step progressive tool‑disguised jailbreak technique that targets large language models (LLMs). The approach, described in…