Depending on what kind of nanopore device this is it may be simpler for the lab to re-do the basecalling and give you the correct results.

However, this time something failed on that step so I was left with all the reads in a single folder.

So you actually have reads in one of the "fastq_pass"/"fastq_fail" folders or no reads at all. With just POD5 files you can re-do the basecalling (using dorado caller) and then you could use chopper (LINK) to filter the reads to avg quality you need.