I like to look at a shape mod for it's own properties, and maybe solve it slightly differently to the original version. My method would be like Bobinogger's, (reduction to Jing's Pyraminx). But I think solving a MS like a Jing's pyraminx sounds visually difficult, and would bring it's own problems.
Because of the format of the Corners + Centres, the MS Corner parity doesn't exist: because it can be removed by a single D twist (face turn), which makes only a 3cycle of the Large Edge Groups, and also orientation of the Centre pieces is unseen, so it's safe to reduce the Large Centres first. So:
Step 1: Reduce the Large Centres.
Step 2: Reduce Large Edges, (like a 555).
Step 3: Solve Large Edges.
Step 4: Permute Corners.
Step 5 : Orientate Corners.
Solving MST.jpg [ 714.36 KiB | Viewed 199 times ]
My drawing diagrams above show how I did the last 2x Large Centres of the Rexamorphix, it's possible to use that, and then do Centres after. I'm actually yet to solve the Master Skewb Tetrahedron, so I will see if the reduction of Large Centres is easy enough on it's own.. EDIT: Yes, it's pretty easy to just incorporate the Centre pieces into the Large Centre reduction.
On my diagram:
The first 2 images in each row represent the patterns on adjacent Large Centres.
The triangular arrows show which way to twist the U vertex.
The rotational arrows show which way to twist the face.
Some of the results lead to previous starting positions.
Sorry for the shorthand nature, but I think the principle is quite simple.. like solving centres on a 555.
It's not much more to do the same thing on the MST.