Hey guys! I started writing a little blog-style series about fivem security aimed at server owners, ressource developers ect.
I’d love to have your inputs on it and let me know if that’s something that interests you! I have a few other things in mind for later 
Have a good read! (aprox 10 min read)
1 Like
I’ve made a part 2 where we look at the server; not as a target, but as a vector to hit other players! Would love to know what you guys think of it!
I don’t know how I got baited into reading all of it (because I am very sure that big evidence script that you’re talking about doesn’t create new evidence from the f**king NUI), but hollyyy, that `window.invokeNative` bit gave me a chuckle.
No, it can’t just invoke natives in-game, that’s just your AI hallucinating. Here is the “undocumented officially but widely reverse-engineered by the community“ source code for it: fivem/code/components/nui-core/src/NUICallbacks_Native.cpp at be6fcc506def9db3ccabef0c8519e0310c05f4c4 · citizenfx/fivem · GitHub
1 Like
Thanks for the detailed feedback.
The script doesn’t create evidence — it attaches plaintext metadata to existing evidence items via unsanitized UI fields. Different thing.
Regarding my “AI hallucinating”, I’m sorry to hear that you confuse my research with AI slop. I do use AI… for spelling and rephrasing. While the specific sentences might be tagged and sound AI, the actual research and writing is my own; granted, riddled with spelling mistakes and awful sentencing. So no, no hallucination here.
I had a look at the link you tagged but you seemed to have misunderstood my point (or I may have been unclear, possible). While the right way to do callbacks (via NUICallbacks_Native) is indeed safe(r), I’m talking about a payload firing window.invokeNative outside the script’s scope which is totally plausible as chain of permission is broken with the NUI system (game permissions - which have window.invokeNative in scope - outright the “correct” NUICallbacks_Native permissions).