Excited to launch this with @busf4ctor. We'll be posting some of our research over the next few weeks 😁 so make sure to follow. Really looking forward to seeing where this goes!
Today, we (@busf4ctor and @monkehack), are launching Starstrike: an AI pentesting and research startup. We'll be releasing our first few research articles over the next few weeks, detailing several bugs that helped us net over $100k in total. Follow to ensure you don't miss them!
Back when I was interning in a data centre, I wrote a Terraform provider for them. This opened my eyes to the flexibility of Terraform and how it's not limited to cloud - it's actually a general system of managing resources declaratively. I think this works pretty well with AI!
My life is now a cycle between building my startup and padding my bug pipeline with more bugs so I can work on my startup in mental peace for another quarter
The writeup is here. We achieved RCE in Minecraft Bedrock, turning a 4-byte heap overflow into complete client compromise.
@ryaagard details a universal, Bedrock-specific technique for bypassing ASLR and achieving arbitrary read / write primitives.
Bug bounty is dying a slow death in real time. This is HackerOne’s attempt to generate some value in the AI bubble, so they have a hope of getting acquired by a company gobbler like Cisco. It’s ironic that their agent will just get worse over time as they drive their best hackers away.
I'm not sure the community will like this. @Hacker0x01 will now reuse your novel techniques / exploits / old reports to look for vulns on the rest of the customer's infra. I guess they will add you as collab and give you a bounty, right? right?!
I was trying to analyse some of my ideas with Claude so I system-prompted it to be more harsh and honest. This worked well until it told me to stop asking it questions and go and speak to my design partners. 💀
Stop blaming AI slop for what bug bounty platforms did to themselves.
👇 wrote down some thoughts about the state of triage
clawd.it/posts/14-to-kara-al…
We're finally ready to talk about Flipper One — a project we've been grinding on for years and have rebuilt from scratch several times. Read blog post >>
blog.flipper.net/flipper-one…
I don’t have any particular bit of research to present at Def Con this year but I do have a lot of miscellaneous techniques in AI hacking and other areas. Should I talk about the gadgets?
Amaze! Amaze! Amaze! Orange Tsai (@orange_8361) of DEVCORE Research Team (@d3vc0r3) was able to exploit Edge with a sandbox escape! If confirmed, we wins $175K. He's off to the disclosure room to explain how he did it. #Pwn2Own#P2OBerlin
wtf. the tanstack attack just went live.
we flagged this exact chain 25 days ago. april 16. All Depi clients were alerted.
if you're using @tan_stack check your manifest files right now 1.166.12, 1.166.15, 1.169.5, 1.169.8 are malicious. clean your cache. rebuild.
Shoutout to @7urb01’s channel. He’s probably one of the best client side guys in the business and you get to witness his unfiltered thought process on his channel as he reviews research and gets sidetracked. This is a MUST follow resource: youtube.com/@7urb0one
He made the windows hug and now the LLM no longer bullies him by rolling to refuse to cooperate when triage tries to reproduce the bug. Thanks doc. Healing the world one iframe at a time. Research Review. youtu.be/2ZvHGtZuWPU
This time we have a guest blog from @xssdoctor, showcasing a new technique in AI hacking to achieve more consistent exploitation. This was initially a research collision, but XSSDoctor masterfully exploited this in the wild.
Link below 👇