AI DAILY W.O.D. DROP — #034 — APRIL 9, 2026

Today's W.O.D.: AI Just Hacked Every Major OS and

Do the course now: https://www.skool.com/aititans/classroom/54cee558?md=50449f11585945e5886b9bbf28f9a046

Anthropic's new AI model just found critical security vulnerabilities in every major operating system and web browser using a technique called Project Glasswing. The same prompting method that's hunting zero-days could be scanning your company's code right now.

Project Glasswing uses Anthropic's Claude to systematically analyze code for security flaws using a specific multi-step prompting framework. The AI found previously unknown vulnerabilities across Windows, macOS, Linux, Chrome, Safari, and Firefox. Anthropic released the exact prompt templates that made their model think like a penetration tester.

**15-MIN WORKOUT:**

Test Glasswing prompts on your own code:

1) Go to Claude.ai

2) Upload a code file from your current project

3) Use this prompt: 'Analyze this code as a security researcher. First, map all user inputs and data flows. Second, identify potential injection points. Third, trace how untrusted data moves through the system. Fourth, flag any validation gaps.'

4) Review the vulnerabilities it finds and share with your dev team.

Who's testing this today? 👇

0:29

1 comment