Joanna, for details you can see my webinar for ASI on using diacritics in Sky 7 and 8, or the latest issue of The Indexer, in which I have an article on diacritics in Sky 8.
Briefly, you have to tell Sky where to sort these characters.
Go to Options > Index Style > Define alphabet groups. That will open a table in which you have to enter the lowercase version of each character you have used in the right letter group. So S with dot under S, etc. I find it best to take these characters from Sky's character map so that you don't have problems with different fonts.
Also, if your author uses half circles for 'ayn and hamza -- rather than single quotation marks -- you have to put these into the Ignored Characters. Go to Options > Sorting > Sort Method > Ignored characters. Paste in the half circles from the Character Map. But note that there is a bug in Sky, and when you do this, the slash mark (/) changes to &slash; and causes major weirdness in sorting the subs. Simply delete &slash; and replace it with an actual / .
Some other tips: If you author uses single quotation marks for 'ayn and hamza, turn off Smart Quotes in Sky and in Word. Otherwise the program will take over and change the direction of many of them. I also find it best to sort word-by-word when there are a lot of Arabic names, especially when both classical and modern names are in the text. That way the variations of spelling sort together.
I also have a webinar on ASI and an article in The Indexer on indexing Arabic names, if you need such help.
Happy indexing!