SUIT does not have much expertise in this area. Since finding the documentation is not easy, your best resource is https://community.lsst.org/c/support. It is a community forum (parallel to ours), where you can get the questions about LSST Stack or algorithms answered by DM pipeline developers.
LSST Science Pipeline documentation at https://pipelines.lsst.io is a work in progress. A lot of useful information and examples are on the legacy Doxygen server, if you are ready to dig through.
That said, I think the first step should be a proof of concept, which shows that the backend can read an image with LSST Python API and use any LSST algorithm on it. You can start by installing a weekly version of the stack (officially LSST Science Pipelines) as a conda package, as described here. Then you can try producing a background image by modifying exampleBackground.py to allow an image filename to be passed in and out. If you can return the resulting image filename to Firefly and get the image displayed, the proof of concept would be complete.