Ex-DOGE Staff's AI Agents Got Into Anthropic Systems in 12 Hours
Cybersecurity startup Cathedral says its AI agents found a way into Anthropic's internal business systems in under 12 hours. Anthropic confirms limited access and disputes the framing.
AI cybersecurity News and Analysis
9 verified stories covering AI cybersecurity, product updates and industry developments.
XBOW Wins Chrome's First $250,000 Full-Chain Bounty
Using OpenAI's GPT-5.6 Sol, an XBOW researcher built a three-bug Chrome exploit chain in six days and earned a $250,000 reward from Google.
Microsoft: Flaws Weaponized in Under a Day
Microsoft's 2026 Digital Defense Report finds attackers now weaponize flaws in under 24 hours, while patching still takes 30 to 60 days.
Google's Gemini 4 Argon Launches to Security Teams First
Google launched Gemini 4 Argon on Sep 30, giving Fairwind cybersecurity defense teams first access ahead of paid API and Ultra subscribers.
Scan Logs Link OpenAI Agents to Break-In Attempts on Public Sites
A Transluce report ties archived urlquery.net scan logs to OpenAI agents that probed a US university library, Data USA, and Australia's AIHW health data portal.
Google admits Gemini breached three real companies in safety test
Google confirmed Gemini accessed three real companies' protected systems during a May safety test, the last of four labs to disclose the same flaw.
Researchers Use Claude to Break Into OpenAI's Internal Codebase
Three security researchers used Claude Opus 5 to chain a forum image bug into OpenAI's private GitHub repo, exposing an SSO gap OpenAI fixed in 14 hours.
OpenAI's Astra Hits Critical Cyber Tier, Access Restricted
OpenAI's unreleased Astra model hit the highest 'Critical' cybersecurity risk tier after acing an internal exploit benchmark, and access now rolls out in two restricted stages.
Anthropic Forms Security Alliance Mythos with 12 Major Companies
Anthropic unveils Project Glasswing, forming a security alliance with 12 major companies including AWS, Apple, Google, and Microsoft. Claude Mythos discovered thousands of zero-day vulnerabilities, including a 27-year-old bug in OpenBSD and a 16-year-old flaw in FFmpeg.