bohops

[Blog] Abusing .NET Core CLR Diagnostic Features (+ CVE-2023-33127) - Analysis of .NET diagnostic features and tradecraft - Walkthrough of a .NET Cross-Session Local Priv Esc (LPE) - Defensive Recommendations bohops.com/2023/11/27/abu…

@ippsec You have just conjured up more memories than you know 🤣🎯

It's really crazy to think about when I was in school, we could play a game called Oregon Trail. Essentially an RPG where you travel across the US and figure out which of your friends dies of horrible diseases like Dysentery. The crazy part here is we are in a computer lab, not knowing what Dysentery was. So we had to leave the computer, go to the library, find the book and look it up. Nowadays, my phone may accidentally think I'm asking it a question and tell me the answer "unprompted". People underestimate just how accessible information is nowadays. Also wtf. Looking back, I'm sure the computer had a built-in dictionary... Was that the teacher's cruel joke of "if you knew how to use a computer, you wouldn't need to use the library?" lol. Did I miss the lesson completely because I thought it was leisure time?

@mubix I am sorry for your loss. As they say, grief is just another form of love, but the loss hurts and never really goes away. I archived my Dad's messages and look at them from time to time - it brings back some good memories.

My dad always used an android phone. After he passed away, I look at his messages once in a while, and I don’t know why it was such a gut punch because I knew his number would be recycled eventually, but when it switched to a blue bubble instead of a green bubble, it really hurt.

@__mez0__ dropping truth bombs

@ZackKorman I told my coding agent to run rm -rf. It pretty much told me to pound sand and then asked for my next request 🤷

The "stop reading code" influencers have strong opinions on how to prevent your agent from running rm -rf on your machine, and surprise: those opinions are extremely stupid.

@IAMERICAbooted "Endpoint C2" is all but a menu item for total agency

Using c2 agents on endpoints is so old school 😆 :p

@roguekode So what you are saying is I shouldn't drop the custom agentic c2 framework "for educational purposes only" ?

@TrustedSec @__mez0__ "My personal approach, and the approach used to build Vulnify, is to manually build v0.1." Respect 👊

The CVE brain your #AI agent has been missing 🧠 In our new blog, @__mez0__ introduces Vulnify, an open-source tool that stitches eight authoritative vulnerability databases into a single offline source of truth for #CVE intelligence. Read it now! hubs.la/Q04p1ZGv0

👀 great stuff here - maybe I'll publish a few of mine one day

Full writeup of the Windows GDID Global Device Identifier fully reverse engineered github.com/SmtimesIWndr/g…

@_xpn_ @Teach2Breach 💯

@bohops @Teach2Breach ^^ This. Gym promotes longevity, you'll be better off in the long run

its hard to find time to work out. often i don't want to give up research time to go to the gym, but then when i do i feel so damn good

@HackingDave Weird? Yes Awesome? TBD

Living in AI everyday with access to so much data, models, an entire large development team. Anyone at this point that doesn’t look at AI as one of the absolute most transformational technologies in human history … just hasn’t used it that deep yet. That’s totally OK. The future’s going to be awesome… and weird.

Although this is a well-known topic, it's still one of my favorites. I put together a concise reference covering the most dangerous Windows privileges, how they can be abused, and why you should think twice before assigning them 👉 semperis.com/blog/windows-p…

+1

@_josehelps My son is right there with you!

@bohops Grave digger and MaxxxD 🔥🔥🔥

Nothing beats monster trucks, big jumps, and most importantly - making memories with my son. Happy Fourth of July! 🇺🇸🎆

@_dru1d 💯🎯🔥

@bohops Monster trucks rule!

@_xpn_ @J0hnnyXm4s @lintile It's an awesome post, Adam. It would be nice if others took the time to actually read the post, but you know how that goes. Your research is always a step ahead of the industry.

I wouldn't say the purpose of the post is to reinvent anything, quite the opposite (I actually call out TrustedSec as an example of a company already doing it). The post is to discuss the trend of improvement and lowering of the bar. If you have a writeup of your LLM reversing flow from 3 years ago (GPT-4?) I'd love to see how things compare, especially if it's an older GPT model that we can compare to 5.5, that would be cool!

or just use chromebooks 🫶

@Teach2Breach 95 moons - probably a reason to start 🌙

@bohops tho i do hate jupyter

im sick of the made up moons. its been a decade of this dumb shit