Two sided marketplaces are a very common type of MVP to build. Using no code (shopify, wordpress, bubble) can get you some traction and gets you some amount of painted doors (someone knocks on the door, but it's not real, at least you know how many people knocked).

This can work for a month or two, but it is not technically defensible, and if your idea succeeds, most of this work will have to be discarded. You want to validate cheaply, but you can do this with a MVP that is hyper focused.

Check out some more of my thoughts on no code platforms:
https://mvpengineer.com/no-code-platforms-worth-it/

and minimizing your MVP:
https://mvpengineer.com/building-tech-demos-for-micro-components-not-a-full-user-experience/

You definitely want a contract in place. Depending on whether you're a full partner (with 33 or 50% equity) vs a contributing member, you may want to use a SAFE or simple agreement for future equity.

Splits in work and responsibility are definitely tough, especially when you're offering a very specific expertise and working part time.

If you're doing this for huge scale with VCs involved, you want to make sure all your ts are crossed (documents, contracts etc). But if you're exploring an idea with friends, the first thing is to validate your hypothesis with an MVP.

I've written a bit about arrangements in early companies for equity and cash here.

https://mvpengineer.com/pros-and-cons-of-equity-based-technical-cofounders/