An example of an exceptionally good question is What is the rank of the universe of small sets in Feferman set theory?. Let me quote it step by step and explain what makes it good. I will not repeat the general principles of a good question, but rather explain how they are applied.
Let me start with the title.
What is the rank of the universe of small sets in Feferman set theory?
This title directly asks the question, and when reading the post, it becomes clear that this is a good summary of the post. What is also good is that this title is a question. A title such as "The rank of the universe of small sets in Feferman set theory" would be more ambiguous, since this question could also be about how to define the rank, if the rank satisfies an inequality, how this rank compares to others, etc.
Also notice that nothing is missing. The title is not "What is the rank?" (the rank of what?), or "What is the rank of the universe of small sets" (in which theory?), or even "difficult set theory problem" (worst case; no information about the question).
The title also only refers to the mathematical subject and nothing else. "How can I find the rank ..." or "I am struggling with the rank ..." would be worse titles, since they contain irrelevant stuff.
The post starts as follows.
Feferman set theory (also known as "ZFC with smallness", or ZFC/S) adds to the language of set theory a constant ...
The OP starts by giving context before asking the question, which is good, since apparently this topic is not quite standard. Right from the first few words we know what this question is about: it is about a special kind of set theory. The OP mentions its name, but also an alternative name, and its acronym, which is good to know for readers, but it also improves SEO (internal and external).
... adds to the language of set theory a constant symbol $\mathbf{S}$ and adds to ZFC the transitivity axiom $$\forall x \, \forall y \left( x \in y \land y \in \mathbf{S} \to x \in \mathbf{S} \right)$$ and, for each formula $\phi$ in the language of set theory, the reflection axiom $$\forall x_1 \, \cdots \, \forall x_n \left( \left( x_1 \in \mathbf{S} \land \cdots \land x_n \in \mathbf{S} \right) \to \left( \phi \leftrightarrow \phi^\mathbf{S} \right) \right)$$ where $x_1, \ldots, x_n$ are all the free variables of $\phi$.
This is the technical definition of ZFC/S, in a nutshell, and it is good that the OP has included this: doing so, not just experts will be able to answer this question. Also, since the definition is right here, more people are enabled to even think about the problem. Also, presenting the definition is good for educational purposes, and SEO. It will definitely happen that someone googles "Feferman set theory", lands on this question, and finds this summary useful.
Bonus suggestion: add a canonical reference where the reader can read more about Feferman set theory. With other topics (not here) which everybody treats a bit differently, adding the reference one is working with is also beneficial since it will help to formulate more appropriate answers.
Notice that the OP has used MathJax, and has used display style for the important and complex formulas. This make this part easy to parse, even though it may appear technical at first.
Also notice that the OP has used the boldface $\mathbf{S}$ to denote the special constant, even when this might not be the case in classical treatments of this set theory. This is much easier to parse than just using $S$, indicates that $\mathbf{S}$ is of "another level" than normal sets, and shows a great effort of the OP to increase readability.
It is more or less immediate that every set definable with parameters in $\mathbf{S}$ is itself an element of $\mathbf{S}$. Hence, $\emptyset \in \mathbf{S}$, $\omega \in \mathbf{S}$, and $\mathbf{S}$ is closed under pairs, unions, and powersets, so $\mathbf{S} = \mathbf{V}_\alpha$ for some infinite limit ordinal $\alpha$.
This part gives more context, and it also presents the work already done by the OP. The OP demonstrates that they have spent time with this theory before asking a question about it.
Also notice that these sentences are very concise and only about mathematical content just like the title: The OP doesn't tell a personal story like "I wanted a new challenge and looked at the axioms and came to the following conclusions", which is irrelevant to the question.
We say a set is small if it is an element of $\mathbf{S}$.
The OP gives an important definition, and the defined term (small) is written in boldface. This makes it very easy to locate this new term.
Question. What can we say about the ordinal $\alpha$? Is it a cardinal? If so, is it regular?
The OP asks the question, and it is nice that this part starts with the boldface Question, which makes the question easy to find within the rather long post.
The ordinal $\alpha$ has been defined right before, which means this question is clear and well-defined. The question "Is it a cardinal?" is specific, short, has a clear yes/no answer. The same is true for "If so, is it regular?".
Notice that the question "What can we say about the ordinal $\alpha$?" alone would be much less clear, since this doesn't tell exactly what the OP is interested in. Maybe they just want to know if $\alpha=42$, who knows. This is why the OP has clarified what exactly they want to know about this ordinal number. This is good since answerers know exactly what problem they need to address. It also avoids clarification requests in the comments.
In summary, the OP doesn't just ask an open-ended question, and instead asks a specific question.
Since ZFC/S is conservative over ZFC, we cannot prove $\alpha$ is inaccessible. If I understand correctly, $\alpha$ may not even be a worldly cardinal, since $\mathbf{S}$ is not assumed to be a model of internal ZFC. On the other hand, every set definable without parameters is small, so $\alpha$ must be greater than any ordinal definable without parameters. Thus $\alpha$ must be large (in some absolute sense) and yet not too large (because of conservativity), and the only way I know how these facts can be reconciled is to see that $\alpha$ may be singular. Is there such a model of ZFC/S?
The rest of the question explains in detail what the OP has tried to answer the question. It also shows what the OP already knows about the problem. The main question seems to be too hard to answer for the OP, but the OP tries to make progress by answering more easy questions. For example, right the first sentence answers the question "Is $\alpha$ inaccessible?" to the negative (sort of).
In this section, it makes sense that the OP leaves the pure technical language from before and also adds sentences like "If I understand correctly, ...", because this section is exactly about what the OP has already tried. Even sentences like "I cannot progress here because ..." would be perfectly fine.
Most importantly, this section shows that the OP doesn't just use MSE to get an answer and walk off. Instead, the OP clearly wants to learn.
The question is also not a duplicate, i.e. it hasn't appeared before on MSE. Maybe the OP has used site search to verify this before asking, but I don't know. But it is even more than that. I will argue that this question is of general interest for anyone studying set theory (and category theory, since this is the raison d’être for Feferman set theory). Therefore, it brings value to MSE as a whole, also in combination with the already mentioned fact that this question will likely pop up when searching for Feferman set theory. This is something not every question needs to satisfy, but it is worth mentioning.
Finally, the tags
set-theory, large-cardinals, alternative-set-theories
give a good classification of the question. The question belongs to set theory, an alternative set theory to be precise (disclosure: I have added the third tag myself). Specifically, it is about large cardinal numbers. In conjunction with the title this is a good summary of the question as well. Users who are interested in alternative set theories will more likely find this question via the tag page. Users who have fav'ed set theory will more likely see this question.
Summary. This is a prime example how to ask a good question.
