VoidLink Linux Malware Framework Built with AI Assistance Reaches 88,000 Lines of Code

[dontcryforme]

The recently discovered sophisticated Linux malware framework known as VoidLink is assessed to have been developed by a single person with assistance from an artificial intelligence (AI) model.

That's according to new findings from Check Point Research, which identified operational security blunders by malware's author that provided clues to its developmental origins. The latest insight makes VoidLink one of the first instances of an advanced malware largely generated using AI.

VoidLink Linux Malware Framework Built with AI Assistance Reaches 88,000 Lines of Code

Experts say the VoidLink Linux malware was largely built using AI, reaching 88,000 lines of code in days and highlighting faster malware development.

thehackernews.com

 

[dontcryforme]

Some IT related news if anyone was interested, if IT related news isnt allowed on the website, just let me know

 

[Herr der Verboten]

So, slop malware?

I laughed my head of yesterday at YT vid (vide coding): when to prompt and when to code

...not like you are actually coding when a LLM is doing it for you.

 

[garyc]

Looks like this may be a proof-of-concept for rapidly devolving malware to exploit vulnerabilities. Might be a taste of things to come.

 

[Herr der Verboten]

Looks like this may be a proof-of-concept for rapidly devolving malware to exploit vulnerabilities. Might be a taste of things to come.

How is this any different from 'kids' running generated scripts; like they did back then?

I suppose you could say its generative malware for task, and if that task changes, so can the generative malware, but still it's same as generated scripts being revised.

 

[RaptorSA]

How is this any different from 'kids' running generated scripts; like they did back then?

I suppose you could say its generative malware for task, and if that task changes, so can the generative malware, but still it's same as generated scripts being revised.

Speed for one thing.
And as much as people like to s**t on vibe coding, these LLM's have instant access to metadata, documentation and samples of millions of already written code libraries out there that can abstract functionality without having to write a lot of functionality from scratch.

I've seen it myself multipe times, I start working on helper classes etc. but then the LLM comes along and sh*ts on my parade by knowing about some insanely abstract library somewhere out there that you can import and drop one-liners for.

 

[Herr der Verboten]

Speed for one thing.
And as much as people like to s**t on vibe coding, these LLM's have instant access to metadata, documentation and samples of millions of already written code libraries out there that can abstract functionality without having to write a lot of functionality from scratch.

I've seen it myself multipe times, I start working on helper classes etc. but then the LLM comes along and sh*ts on my parade by knowing about some insanely abstract library somewhere out there that you can import and drop one-liners for.

I think the crunch/analyses part is great but not always the output.

 

[konfab]

Claude: Hey, mind if I grep -ohP "useEffect\(.*?\[\K[^\]]+" **/*.tsx 2>&1|tr ',' '\n'|awk 'NF{$1=$1;a[$0]++}END{for(k in a)print a[k],k}'|sort -rn|head -20

Me: ... yeah go for it dude

 

[RaptorSA]

Claude: Hey, mind if I grep -ohP "useEffect\(.*?\[\K[^\]]+" **/*.tsx 2>&1|tr ',' '\n'|awk 'NF{$1=$1;a[$0]++}END{for(k in a)print a[k],k}'|sort -rn|head -20

Me: ... yeah go for it dude

Liewe f0k.

This is how I built my Proxmox Homelab cluster...
List a bunch of wants, needs and wishes, ask for bash commands, copy-paste the snow out of them until something starts to work... for all I know this thing has its own bot running on my bank account by now.

The thing is though, it works... and it works well. The argument about how it's always better to fully understand what you did is fine, the issue is time is precious and for me to spend days or weeks to learn all the ins-and outs of setting up a simple ZFS mirror with some samba sharing on an lxc contaient isn't worth it.

 

[OutlawJoseyWales]

Speed for one thing.
And as much as people like to s**t on vibe coding, these LLM's have instant access to metadata, documentation and samples of millions of already written code libraries out there that can abstract functionality without having to write a lot of functionality from scratch.

I've seen it myself multipe times, I start working on helper classes etc. but then the LLM comes along and sh*ts on my parade by knowing about some insanely abstract library somewhere out there that you can import and drop one-liners for.

Mike Adams built a book creation tool by himself by vibe coding.
It would have cost him hundreds of thousands of dollars if using human coders but he just used AI, and the time in which it was completed was astounding.

It has taken off like wildfire.

You can create one now here https://brightlearn.ai/