Thank you for your input regarding fix GCMC. I hope that you find it useful. I'll start off by noting that the molecule capability in fix GCMC is still fairly new and can use more testing to work out final bugs, as you have found. As far as I know, you're the first person to try it with a molecule as big as hexane.
You're correct about issue #1. I've checked in that change as you suggest into the repository.
The second issue you point out is a known limitation that I hope to remedy at some point in the not-too-distant future. The model molecule does need to be present whenever fix GCMC is initialized, since initialization of the fix does a rebuild of the model molecule. In the meanwhile, I've updated the docs in the repository to reflect this more accurately.
Any further feedback you might have on fix GCMC is welcome. Thanks for your input thus far.