none of the Bluebook styles are terribly good - the main reason is that (and I'm not making this up) the Harvard Law Review people claim that an automatic implementation of Bluebook would constitute a copyright violation.
I believe the version called "Bluebook, 19th edition (incomplete)" on the repository is the best and it looks like it would do Id. correctly.

@roypeled,

If you're comfortable running unsupported software (which is basically a matter of being patient of errors if they occur, being rigorous about keeping backups handy, and knowing how your backup system works) you could give MLZ and the Wisconsin Court Style a try. It implements the same rules as the Bluebook (which we're entitled to do because adhering to them is required by law in many jurisdictions), but avoids use of their trademark.

The WCS style will not run correctly (I think) in official Zotero, but it has been pretty well tested under MLZ. If you would like a preview of how it performs, proofsheets are available here. I'm about to fill out coverage of statutes, which should be ready in a day or two -- there will be some significant changes there, which I'll document in the proofsheets when they go up.

The reason MLZ is unsupported is a matter of developer time. It's likely that the extensions it applies (or something like them) will feature in a future upgrade of official Zotero, but in the meantime it's my own third-party side-project, and since development of the main product would be slowed if they provided support, they don't. It's a sensible position, but you should know where you're going if you install the MLZ client. It should install cleanly against an existing Zotero 3.0 database, and reverting from MLZ to Zotero 3.0 should work fine as well. (For what it's worth, I haven't had any reports of database corruption as a result of running MLZ.)

It allows arbitrary field variants (translations, transliterations and sort strings) to be stored in a standards-conformant way for things like title, author name, publisher and place of publication. The variants can then be weaved into citations using an extra Languages configuration tab in the Preferences panel. You could cite a Hebrew resource in the original script, with a supplementary translation, for example. The one thing it can't do is output parsed dates in non-Gregorian formats; for dates in the Hebrew calendar, you would still need to enter those as plain strings.

Try it out-- the "M" in MLZ is really amazing once you start working with it.