OpenAI Launches 'Daybreak' to Help Build Secure By Design Software

A broad set of cybersecurity stories: vendors and researchers are rolling out AI-driven defensive tools (OpenAI Daybreak, Anthropic Claude Security) while attackers increasingly leverage AI for offensive work (AI-assisted zero‑day development, LLMs used in OT attack). Multiple supply‑chain and repository compromises were reported (Mini Shai‑Hulud tainting TanStack npm and spreading to PyPI; malicious/typosquat Hugging Face repos; trojanized Daemon Tools). New Linux kernel flaws—two vulnerabilities reported and chained as “Dirty Frag”—are high‑severity and prompted rushed patches. Active threat