Could be that I might be of help for you. I am writing on a programm called JTranslator. The version that is uploaded currently, won't be of help to you as it won't support hebrew, at least not appropriately. Still, I'm already working on a new version that is meant to support a bunch of languages - hebrew amongst it. The new version is already capable of that.
In my oppinion, it is better to stick to a charset that consists of 256 symbols. Think of charsets like tables saying symbol no. 1 is an 'A' and so on. The key to read hebrew text is, to interpret the text files like iso 8859-1 at first, which is the default charset of civ4. Then convert the letters to their numeric counterpart and interpret these numbers according to a charset that supports hebrew. To write hebrew texts, you would need to do the exact same thing only backwards. To have this simple, it would be the best choice to use a 256-based charset (not even sure if I could support UTF8 and / or 16 this ways). That's AFAIK what the Russians did and it exactly is what the not-yet-released -version of JTranslator does, using
ISO 8859-8 (well, I filled up the empty positions with letters from ISO 8859-1 to have it save to convert between charsets - but that should be no problem for you). If you'd prefer, I could build in
windows 1255 instead or as second option.
So you would have a tool to write hebrew texts in a way that would keep texts in other languages antouched (you should care about this as if you do it carelessly, you'd could destroy mods text files breaking the mod for users from other countries). Unfortunately, I can't see how I could help you with anything else. I can second things said about the font and the Gamesfont.tga, but I'm not able to help you with that - can't even read hebrew.