intel clearly said "decrement counter and then check if counter is 0" so that means if ecx is 0 already it should loop.
and thats fine, and we all know this.

however in intel's code samples, representing x86 asm in C\CPP intel represented the loop instruction in a wrong way.
saying that it first check if the counter is 0 and if its not it takes the loop, so thats probably a mistake made by intel.

One way to be sure is to test it and see which of the two descriptions actually happens.

First one.

I wouldn’t tell C/C++ examples serious while talking about instructions. After all, they’re not documentation and are rarely written by people who actually spen much time using the instructions directly.

i know its the first one, im just saying there are some mistakes.

Where are you getting the Intel C/C++ equivalent code? It is not in Intel 64 and IA-32 Architectures Software Development Manual Vol 2. There, the behavior of the instruction is well defined in pseudo-code.