The pyslim example in its tutorial that is titled "Generating initial diversity with msprime" loads a saved .trees in and simulates starting from there. It gets neutral diversity from msprime, so it defines a neutral mutation type. This results in the usual warning about using a neutral mutation type with tree-seq; but SLiM should be smart enough not to emit that warning if a load from a saved file occurs in the first tick of the simulation, at least. The warning is a bit confusing in this case.
The pyslim example in its tutorial that is titled "Generating initial diversity with msprime" loads a saved .trees in and simulates starting from there. It gets neutral diversity from msprime, so it defines a neutral mutation type. This results in the usual warning about using a neutral mutation type with tree-seq; but SLiM should be smart enough not to emit that warning if a load from a saved file occurs in the first tick of the simulation, at least. The warning is a bit confusing in this case.