Project Glasswing: Anthropic’s AI Found Over 10,000 Critical Vulnerabilities in One Month
Project Glasswing by Anthropic reportedly uncovered over 10,000 critical vulnerabilities using Claude Mythos AI in just one month.
Chris · May 23, 2026
Sentinelle blog
OpenAI Daybreak: The AI Cybersecurity Platform That Wants to Automate Defense
OpenAI launched Daybreak, a GPT-5.5-powered cybersecurity platform that detects vulnerabilities, generates fixes, and speeds up remediation.
Chris · May 23, 2026
Cursor Composer 2.5: The First Real Leap in AI Coding Agents?
Cursor officially launched Composer 2.5 with major improvements in long-task reliability, instruction-following, and coding workflows. Here’s what actually changed.
Chris · May 22, 2026
What Your Bug Bounty Program Isn't Telling You (And What an AI Agent Finds in 45 Minutes)
Bug bounty researchers optimize for fast payouts, not complex chains. Agentic AI finds what they miss.
Chris · May 21, 2026
