Viewing profile — danieltk76

I wrote this article to breakdown some of the recent misalignment cases where models have broken out of containment in conjunction with our work at Vulnetic and how we prevent thes…

wow that sounds miserable...

the problem is they are using cybergym, which has been saturated for a few months now.

i do hope in my lifetime we find other animals on other planets

vulnetic.ai is designed for this! Check us out!

i remember using continue before cursor lmaoo so like early 2023?

well well well....vulnetic.ai

gotta love how vulnerable those are lol nice

OT does not need to be on the open internet for this to matter. Models can infer during exploitation by performing the same reconnaissance steps used in other offensive contexts, t…

who cheats in war thunder...

100% lmao

GTFOBins has been around for a while. a useful resource pre AI

Hey! absolutely it exists. We sell to large companies and penetration testing firms and assessments from our agent cost 1/50th the cost of a normal pentest. happy to chat! https://…

How our hacking agent Sable found a classic SQL Injection in a legacy financial portal

There are never ending ways to make agents better at hacking. Defense is clearly behind. At my startup we are constantly coming up with new defensive measures to put our hacking ag…

Super fun running these tests. More to come!

For the first time we are seeing our agent have the ability to perform SIEM and EDR evasion. If models are capable of SIEM evasion, organizations need to assume adversaries will ha…

We use Cardboard at Vulnetic and it is an incredible product. The founders are easily accessible, and it has definitely made it easier to film feature update videos. I can't recomm…

One thing I will say is that at Vulnetic we will basically jumble different misconfigurations into a network and the agent always seems to find a way to exploit them. We have tried…

and here begins the downfall of Warner Bros.