NEW: malware developers added nuclear & biological weapons text to to their spyware.
Goal? To trigger LLM safety refusals... so that their spyware wouldn't be analyzed by an AI security scanner.
Cleanest practical example I can think of for why over-indexing on first order safety alignment is risky.
When closed (and open) models ship with aggressive refusals, they will be sprinkled with second-order blindspots that attackers will discover...and exploit.
We are only in the earliest days of attackers leveraging these features, and it wouldn't surprise me if users systems that need to handle complex cybersecurity issues demand that models be less safety-blunted.
In the weeds: @SocketSecurity's post also shows why intention matters in how you design a malware analysis pipeline to avoid prompt manipulation.
H/T to colleagues that shared this with me https://t.co/f3Aj9TYxU4
🛑 OpenClaw AI agents can leak data via indirect prompt injection.
A crafted URL generated by the agent triggers Telegram or Discord link previews that silently send sensitive data to attacker domains.
China’s CNCERT warns organizations to isolate or restrict the tool.
🔗 Attack details → https://t.co/gtpGUldFrO
🚨EDR Telemetry website is live! 🥳
I hope this makes it even easier for folks to compare the telemetry of EDR vendors and visualize their visibility gaps 🙂
‣ Website🔗https://t.co/gc5UXmr8ih
‣ GitHub 🔗https://t.co/NNAI4K539x
**Telemetry results reflect the most recent updates from the EDR Telemetry project.
I and @s0nnguy3n_ have successfully reproduced CVE-2023-2825. This vulnerability allows unauthenticated malicious users to read arbitrary files on GitLab server.
This is an interesting vulnerability that is quite easy to reproduce.
Attackers deleted all of the Win event logs?
Don’t give up!
1. Take a memory dump
2. Mount with MemProcFS
3. Copy out the evtx files with something like gci -path M:\pid\ -Include *.evtx -Recurse | Copy-Item -Destination .\evtx_files
4. Run Hayabusa!
I found a vulnerability that allowed me to unlock any @Google Pixel phone without knowing the passcode. This may be my most impactful bug so far.
Google fixed the issue in the November 5, 2022 security patch. Update your devices!
https://t.co/LUwSvEMF3w
It took me 18 years to become a #cybersecurity expert. You want to be as awesome as I am? Here’s how you can do it in 13 easy steps.
1. Wake up at 4:15 am
2. Meditate for 1 hour. Clear your mind like you have 0 unread emails.