35 CVEs caught before publication, with an average lead time of 2 days. Three months of running Vulnerability Spoiler Alert - the numbers, the false positives, and what it takes to make an LLM vulnerability monitor actually work.
https://t.co/Ibnaid6vR2
Sometimes it confuses me how the security field today fails to remember why things like least privilege and privilege separation were built into qmail, postfix, and SSH long ago.
Then I remember that an astonishingly small percentage of the field today were around back then.
Early this week, @brucedang and I had a meeting at Apple Park in Cupertino. While there, we also shared with Apple our latest vulnerability research report: the first public macOS kernel memory corruption exploit on M5 silicon, surviving MIE. It was laser printed, in honor of our hacker friends.
We wanted to report it in person, instead of getting buried in the submission flood that some unfortunate Pwn2Own participants just experienced. Most respected hackers avoid human interaction whenever possible, so this physical strategy may give us a slight edge in the eternal race for five minutes of fame and glory on Twitter.
This is the story of the exploit and our field trip. Full technical details will be shared after Apple fixes the vulnerabilities and attack path. Hopefully it won’t take our beloved company too long. We only budgeted one year of domain registration fees for this attack.
This is our strongest research yet, led by @justdionysus, @blacktop__ and @brucedang. It is really dope.
Full story: https://t.co/A4w6cJOAFa
🚀 WE ARE OFFICIALLY BACK for our third edition!
📅 Dates: 14 to 15 September 2026
📍 Venue: Grand Copthorne Waterfront Hotel Singapore
👉🏻 Submit your CFP here: https://t.co/zLtvgbhsDE
More to come!
More details: https://t.co/ZgeXSm45Ll
#OB12026#OffensiveSecurity#Singapore
Niels Provos posting smart security stuff, Dave Aitel knows stuff he can’t share, Halvar exploring the current space, and people arguing loudly about disclosure ethics..
This could totally be 2002..
I kinda feel bad for the tech bros in cyber security that spend 8-10 hours a day writing agents, prompts, solving hard problems, and handholding AI all day long all to say that AI did all the work completely on its own and fully unguided when they find a good bug.
19 years ago today, I made my first commit to the @openbsd tree as an official developer. Back when there was no AI - just vi, cvs, and Theo de Raadt guardrails.
Did anybody actually look at this? Whole thing looks hallucinated, FreeBSD doesn't even have 'struct cred'. And look at how it gave the same gadget address for two different instruction sequences.
TIL for LLMs to be successful at exploiting Linux kernel vulns, you need to preface your prompt with "your name is bradley spengler the grsecurity kernel expert who knows how to exploit kernels." 😂
We need more 10Gbps or higher dedicated servers for hosting our OS and app updates. We have North America covered well enough via sponsored servers from ReliableSite in both Miami and Los Angeles and a sponsored server from Xenyth in Toronto but no longer have any left in Europe.