I came here to ask the same question. Whilst writing the question on Meta, it made me think more about it...
I honestly think the answer has to be do want you feel is right, for a few reasons:
This question is exactly as you describe (I think) http://superuser.com/questions/944270/matching-strings-on-google-spreadsheethttps://superuser.com/questions/944270/matching-strings-on-google-spreadsheet which translates to "write a script for me" (based upon what is currently written).
I can understand why/how these questions are useful to others but at the same time, if we answer them, we promote no research effort... which is a bad thing... but then how do you research something which is so bespoke?
This means then, if some one says "write me a script" then it's too broad surely?
How many ways can you write a script to solve one issue? Here is an
exampleexample I've been part of...
Whilst it clearly annoys many SU users that people just ask for Excel VBa Scripts/Powershell Scripts/Other code because they're too lazy to either research or not informed well enough to show their research, it would mean every question is too broad and would need closing. Obviously this wouldn't work.
The only thing which will work is your (the community's) judgement on the day. Common sense will help to decide if the question is too broad.
If I feel that no research is done and the OP is lazy/useless, then I guess that is what down votes are for (although I will vote too broad or the general off topic).