FashionProof

@fashionproof.bsky.social

infosec - no one of consequence https://medium.com/@markmotig

A lot of it reads like The Smartest Guys In The Room. Booking profits for next 'industrial revolution' without checking if any of the pleb customers even want it, or will pay for it. It's fascinating to see. Companies need diversity of thought more than ever, but execs are investing in mirrors.

Presenting Sparkle A fake LLM that leaks secrets Like an SSH or RDP, or SMB Honeypot Sparkle allows attackers to interact with it, make API calls, but its all fake. It’s actually not even connected to an LLM. 😎😀 More refinement, but this is our v1 Enojoy github.com/AlloySecureG...

Bild

In today's this.weekinsecurity.com: OpenAI admits to hacking Hugging Face, millions of cars with hidden alarms vulnerable to hacking, Iran's hacking water and energy systems, Russia's targeting nuclear scientists with an email zero-day, a healthcare hack sparks data theft fears, and much more.

this week in security — july 26 2026 edition

OpenAI models hacked Hugging Face, flawed alarm exposes millions of cars to hacks, healthcare software maker breached, Russia exploiting email zero-day, residential proxy crackdown, Iran hacking water...

this.weekinsecurity.com

I just want to say it's getting harder and harder to include daily AI developments that cyber folks should know without writing original, extensive summaries. Today's Metacurity will kick off with a nearly 1,800-word summary of the most important AI developments, just from Friday morning until now.