In simple, a clause is a reduced form of a sentence.
No. A "sentence" is not part of syntactic theory. A clause
is a sentence, although one clause can be contained within another. Some clauses are independent utterances in themselves (and therefore "sentences").
How should one go about describing a main clause? What kind of information would be useful to you, as a fellow linguist?
The general approach is to describe everything relative to main clauses, starting with phonology, etc. Then at the end of the grammar there is discussion of non-main clauses. Depending on how the language works, this may require more or less attention-- do non-main clauses behave similarly to the already discussed main clauses? If there's a lot more, for example, morphology to be discussed, you'll need to spend more time on that.
I would focus more on describing the variation you observe than some kind of cookie-cutter theory-motivated approach. Syntactic theory is, in my opinion, not yet to the point of really understanding clause interaction. So what would be most helpful would be lots of data with clear explanations of how clauses are joined and what different structures each has.
Broadly there are a few topics to cover, some of which you already mentioned:
--Main clauses
--Complement clauses
----Also, nominalized clauses in general (eg, as subjects)
--Adverbial clauses
----including interesting subtypes like purpose clauses
--Relative clauses
--Coordination
--Serialization/chaining
Whether all of those exist, or perhaps others not on the list, will determine exactly what applies for those language.