RTMB might finally solve the basic issue of tuning marlin, or at least help, since allows for TMB style fitting but can pass R functions directly. So, I think could have a function to say alter total effort and q to match catches and get a depletion to reach target values. Would also give you uncertainty in parameters that you could then run simulations over.
https://kaskr.r-universe.dev/articles/RTMB/RTMB-introduction.html