
CVE-2026-7482 in Ollama Exposes 300,000 AI Servers to Memory Leaks
CVE-2026-7482, dubbed ‘Bleeding Llama,’ exposes 300,000+ Ollama AI servers to heap memory leaks via a crafted GGUF file. Patch to

CVE-2026-7482, dubbed ‘Bleeding Llama,’ exposes 300,000+ Ollama AI servers to heap memory leaks via a crafted GGUF file. Patch to

Unknown attackers compromised the official JDownloader website and replaced legitimate Windows and Linux installers with a Python-based remote access trojan.

A fraudulent OpenAI repository reached Hugging Face’s trending list while distributing infostealing malware targeting credentials and access tokens.

cPanel patched two CVSS 8.8 flaws including Perl code execution in WHM, as the 40,000-server CVE-2026-41940 campaign remains active.

Researchers disclosed 12 critical vulnerabilities in the widely-used vm2 Node.js sandbox library, all enabling sandbox escape and arbitrary code execution

Security researchers from Adversa AI and Mitiga disclosed a one-click RCE, silent MCP OAuth token hijacking, and a Chrome extension

Ivanti disclosed CVE-2026-6973, an actively exploited RCE vulnerability in EPMM 12.8.0.0 and earlier. CISA set a May 10 federal remediation

Three PyPI packages with 2,400+ combined downloads delivered ZiChatBot malware to developer machines, abusing Zulip’s REST API as a covert

Researchers identified a Linux variant of Quasar RAT targeting developer systems to steal source code access, CI/CD credentials, and signing

Disc Soft confirmed a build environment compromise that distributed trojanized DAEMON Tools Lite installers deploying an infostealer, backdoor, and QUIC
Subscribe to the Daily Security Review Email Briefing to stay informed on the latest threats, trends, and technology, along with insightful columns from industry experts.