Proof-of-Concept in 15 Minutes? AI Turbocharges Exploitation Summary An AI-powered offensive research pipeline, created by two independent Israeli researchers and dubbed “Auto Exploit”, uses prompts to a large language model,…
Boffins Build Automated Android Bug Hunting System Summary Researchers from Nanjing University and the University of Sydney have created A2, an AI-powered agent that discovers and validates vulnerabilities in Android…
ICE Has Spyware Now Summary The Trump administration has lifted a Biden-era restriction that blocked US agencies from acquiring intrusive commercial spyware, clearing the way for US Immigration and Customs…
Proof-of-Concept in 15 Minutes? AI Turbocharges Exploitation Summary An AI-driven offensive research system, dubbed Auto Exploit, was built by two independent Israeli security researchers and used an LLM pipeline (Anthropic’s…
Secretive MaaS Group ‘TAG-150’ Develops Novel ‘CastleRAT’ Summary Researchers have traced a young, secretive malware-as-a-service (MaaS) operation labelled TAG-150 that uses a loader dubbed CastleLoader and a wider service known…
Scammers Are Using Grok to Spread Malicious Links on X Summary Bad actors on X are exploiting Grok, X’s native AI assistant, to republish hidden malicious links from promoted video…
Anyone Using Agentic AI Needs to Understand Toxic Flows Summary This Dark Reading piece explains why “toxic flows” are emerging as a critical security risk for organisations deploying agentic AI.…
Critical SAP S/4HANA Vulnerability Under Attack, Patch Now Summary A critical code-injection flaw in SAP S/4HANA (CVE-2025-42957) with a CVSS score of 9.9 is being exploited in the wild. The…
Critical, make-me-super-user SAP S/4HANA bug under active exploitation Summary A critical code-injection vulnerability in SAP S/4HANA (CVE-2025-42957) — rated 9.9 — is being actively exploited. SecurityBridge discovered and disclosed the…
The crazy, true story behind the first AI-powered ransomware Summary A team of New York University researchers built a proof-of-concept they call Ransomware 3.0 that uses large language models to…
