Timeout for subkernels using SLURM

I had previously been using a cluster with SLURM to run parallel computations with ParallelTable. However, it has randomly stopped working and I now receive the error:

LinkOpen::string: String expected at position 1 in 
LinkOpen[StringRiffle[{/apps/mathematica/13.2.1/Executables/wolfram, -noinit, -subkernel, -wstp}], 
LinkProtocol -> Automatic, LinkHost -> , 
LinkMode -> Launch, LinkOptions -> 256].

Followed by a LinkObject::linkv error, then finally

Timeout for subkernels. Received only 0 of 8 connections.

I am submitting with the following:

#!/bin/bash
#
#SBATCH --job-name=submit
#SBATCH -p Partition   # partition 
#SBATCH --nodes=1
#SBATCH --ntasks=1
#SBATCH --cpus-per-task 8
#SBATCH -o slurm.%N.%j.out # STDOUT 
#SBATCH -e slurm.%N.%j.err # STDERR 

wolframscript -file code.wl

and code.wl contains:

CloseKernels[];
LaunchKernels[];

result = Table[i^2, {i, 1, 10}];

Print["Results: ", result];

CloseKernels[];
Quit[];

Oddly enough, the system admin has no trouble when running this code so it seems linked with my login.

Any help would be appreciated.

Edit I:

So, if I use math -script code.wl then the problem disappears.

However, I was using wolframscript because this was the only way I could load in command line arguments. In other words if I try

math -script code.wl 4

then Mathematica returns {} if I include $ScriptCommandLine in the code. I am not sure how I reconcile these two issues.

Edit II:

I've discovered that if I use

MYARG=4 math -script code.wl

then I can read in this value using

envVar = Environment["MYARG"];
Print["MYARG = ", envVar]; 

While this gets things functionally working again, it feels more like a work around than a solution.