yes, ideally you add an nf-core module for your tool and then try to add it to the assemble workflow. Whatever inputs you need to you need to prepare and pass to that module accordingly,

reads.fof

should be trivial to create from reads input channel (

echo ${reads} > reads.fof

) for the config we have to see what the best way is, but I assume that this would also come up when you add the module to nf-core. I guess you could simply make use of the existing

params.assembler

to switch inside the

assemble

subworkflow (as e.g. here https://github.com/nf-core/genomeassembler/blob/dev/subworkflows/local/assemble/main.nf#L47)