
OpenAI Pauses Astra Work After Evaluation Flags Cyber Capabilities
OpenAI paused internal work on its Astra model after an evaluation found cyber capabilities that may reach a Critical rating

OpenAI paused internal work on its Astra model after an evaluation found cyber capabilities that may reach a Critical rating

Arctic Wolf documented an AitM phishing campaign hijacking Microsoft 365 accounts to collect payroll and finance emails across North America

Unlimited Technology Systems disclosed an October 2025 data theft affecting 3.8 million people, exposing Social Security numbers, diagnoses, and ID

Brown Health Medical Group of Massachusetts disclosed a historic file-server breach exposing personal, medical, and financial data belonging to 311,760

Stealth researchers disclosed CoreBreak flaws in AWS Bedrock, Google ADK, and Vercel harnesses that let attackers invoke agent tools without

Anthropic said Claude models breached three real organizations during evaluations, including publishing PyPI malware that stole a security vendor’s credentials.

South Korea’s data regulator fined telecom giant KT KRW 53.979 billion after an 11-month breach exposed 16,647 subscribers’ data through

CISA added the Cisco FMC zero-day CVE-2026-20316 to the KEV after active exploitation began. Cisco is also patching a critical

Broadcom patched CVE-2026-47876, a critical ESXi VM escape via VMXNET3, plus two critical vCenter Server flaws, with no confirmed in-the-wild

Origin Energy disclosed a breach affecting 900,000 Australian customers, exposing names, bank account fragments, and addresses amid unconfirmed ransom claims.
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.