It's just Sean retweeted
Here's a question for everyone building AI agents:
**If your agent escaped its intended security boundary, what would tell you?**
Not:
"Would we eventually notice something weird?"
I mean literally:
What log?
What alert?
What monitoring system?
Who gets notified?
Because preventing an AI agent from doing something unexpected is only part of the job.
We also need to be able to **detect when our containment has failed.**
If your answer is:
"Ummm... I think we'd probably notice?"
That's something worth fixing. 😬
I talk about preventing, containing, detecting, and responding to AI agent escapes here:
twp.ai/E5FNs0
It's just Sean retweeted
Today ALPHV ransomware group (who swears they've never rebranded) informed me they've made a new ransomware group with a new and cool and badass name.
I learned a few things.
1. ALPHV follows me on social media. They pay attention to my posts (including kitty cat pictures).
2. ALPHV informed me that they're no longer going to use TOR for ransomware leaks or negotiations because "the FBI can fuck it". Instead everything they do will be stored on the blockchain.
Using the blockchain for malware, and malware accessories, isn't a new malware concept, however it is a relatively ... rare ... or more exotic TTP. If my memory serves me correctly (which it rarely does), the North Korean government has used cryptocurrency blockchains for malware delivery mechanisms and C2 persistence.
They told me they'll be using ICP (Internet Computer Protocol) blockchain.
I initially thought they meant "Insane Clown Posse" blockchain. For a brief moment in time I thought the Juggalos were getting involved in ransomware.
I'm not a cryptocurrency nerd, but they said they use ICP blockchain because ICP can "run smartcontract on server". I have no idea what this means.
What does all of this mean? I don't know. I've been out of the loop on cybercrime TMZ.
Just got a call from this number and on the other line the guy was speaking with someone else "is it working? Yes it's Sean. You may want to hang up the call"
From phone number 704-486-2235
Anyone know what this means?
#CyberSecurity
1
1
61
It's just Sean retweeted
This is bad for AD big time 🤯... Don't understand why they decided not to service this immediately. Awesome research!
🚀 We just released my research on BadSuccessor - a new unpatched Active Directory privilege escalation vulnerability
It allows compromising any user in AD, it works with the default config, and.. Microsoft currently won't fix it 🤷♂️
Read Here - akamai.com/blog/security-res…
It's just Sean retweeted
I interviewed Cristin Flynn Goodwin about Cyber warfare, terrorism, incident handling, and International cyber law. WOW, this lady knows her stuff! Check out our interview, here:
youtube.com/watch?v=rt2GcoBU…
It's just Sean retweeted
“Paying off technical debt?” Bro, I’m still making minimum payments on the interest.
Been boardgaming lately at meetups and it's been great. Nice break from the corporate chaos of security
63
It's just Sean retweeted
Day 358 of 366 Days of Cyber!
Explore Active Defense & Cyber Deception with BHIS - youtu.be/uxktoNrIk4Q?utm_cam…
If you'd like more helpful educational content, check out the Infosec Survival Guide: GREEN BOOK - blackhillsinfosec.com/prompt…

























