Ask HN: Who else got pwned by the Next.js RCE?
6 comments
This might be a hot take, but I feel like the blurring of lines between back-end and front-end apps with platforms like Vercel will lead to more and more of these exploits. I’m an experienced full-stack dev and I’m constantly confused as to “where I am” in a Next code base. Server? Client? Edge? Proponents might say “that’s the point - you don’t have to worry about there you are, it’s one code base” but these sort of issues indicate otherwise.
All platforms can be exploited I guess, but I still wonder at the complexity of the platforms we now rely on and whether it’s justified.
All platforms can be exploited I guess, but I still wonder at the complexity of the platforms we now rely on and whether it’s justified.
> All platforms can be exploited I gues
React did not have this kind of security vulnerability in 10 years. The Vercel/NextJS/RSC rugpull is responsible for that and the people that made those changes should be named. The lack of shared governance is abysmal.
React did not have this kind of security vulnerability in 10 years. The Vercel/NextJS/RSC rugpull is responsible for that and the people that made those changes should be named. The lack of shared governance is abysmal.
That touches on why I never pursued server-side React in any form. It seemed to twist what was a clean break between layers into spaghetti. I totally get that it solves other problems, but it always felt to me more like trying to force React to be something it was not. The better strategy seemed to me to use React on sites where users can handle the bulk of a front-end React app, and don't use it elsewhere.
Specific to security, keeping React 100% client-side keeps things simple: Don't trust the front-end.
Specific to security, keeping React 100% client-side keeps things simple: Don't trust the front-end.
I don't use Next.js but I'm curious as well. My impression was that most people run it under Vercel who patches quickly, but maybe that's not the case.
You had to patch manually
I'm sure a lot of people and companies got pwned and they aren't going to disclose it. There are chrome extensions that passively polls sites for the vulnerability, and since the vulnerability is so simple to exploit and leaves virtually no trace...
My gut feeling is that we are going to be feeling the consequences of simultaneous enshittification of software, the mounting complexity of our systems, and AI enslopification combine to create far more vulnerabilities in the future. The only defence is to adopt simple systems and software.
My gut feeling is that we are going to be feeling the consequences of simultaneous enshittification of software, the mounting complexity of our systems, and AI enslopification combine to create far more vulnerabilities in the future. The only defence is to adopt simple systems and software.
Fortunately this machine wasn't anything important for me and there was no sensitive data to exfil beyond AI API keys. But I imagine there's other orgs that just got catastrophically, irrecoverably pwned.
What's your story?
(RCE context: https://news.ycombinator.com/item?id=46136026 )