There are a lot of answers to this question - the best will depend on the details of your exact situation (single workstation or cluster? Doing it once or doing it routinely?). The simplest that I can think of is to just run the jobs in bash. So for example:

# Run FastQC on all the FastQ files
for f in *.fastq.gz;
do
    fastqc $f
done

# Run MultiQC on the results
multiqc .

This is simple but it will be fairly slow, as it runs FastQC on one file at a time. Depending on the size of your compute setup, it may be better to run a dedicated pipeline tool such as Nextflow or Snakemake. I work with Nextflow and the nf-core community, and the template pipeline that we base new pipelines on does basically what you're asking for, so you could even use that: nf-core create to make a new pipeline.