Maybe it depends how bad is your QC report and you should ask "To trim or not to trim ?".
By he way, I have heard that most of the sequencing company (e.g from China and south Korea) remove the adapters before sending the results to the scoutmasters (I have seen many clean raw data in Tehran University Omics center). So, do you have "adapter contamination"?
if yes maybe some adapter remover/clipper tools is what you need.
If the quality position in not very bad (Red alert) I suggest use your data and proceed to your next steps. (if they are RED you should contact with your sequencing service and ask them why is that)
I have worked on some fish RNA-seq samples and I have used my raw data first for my de novo assembly and then I have used trimmer tools to improve my read quality and then performed another assembly and the quality metrics of assembly with the second set was decreased!
I assume the origin of the question is because those two modules in FastQC "failed"? Neither "failures" may specifically need trimming though. In general you would want to trim and remove any adapter contamination. I recommend bbduk from BBMap suite. Trimmomatic is an alternate option.
Take a look at this blog post for one of them. Other may just be a characteristic of your library. Post additional data (images) if you need us to look at the actual data.