With Cyrillic, because of it's proximity to Latin, there may be a way to have a 'standard transfer pattern' set of rules, whereby Cyrillic<->Latin can be done with clarity and consistency. But we'd have to get our act together ...
Aside: The inverse problem is also funny: when you have a Latin name, it's hard to tell how it's supposed to be pronounced. When one wants to transliterate it into Cyryllic, one would wish to preserve pronounciation. This causes some names to be transliterated in many different ways, depending on the transliterator's opinion of the correct pronunciation of the original name.
And of course choosing the standard Cyrillic-alphabet language
For example, the (quite common) name Юрий has been transliterated as Yuriy, Yurij, Yurii, Yuri, Juriy, Jurij or even other options.
Also, you can't transliterate from Cyrillic, you can transliterate from a particular language, since any phonetic transliterations will be slightly different between, for example, Russian and Ukrainian - even ISO 9 accounts for that, so a sequence of letters without context can't be sufficient for transliteration, the exact same sequence of cyrillic letters may have to be transliterated differently depending on its language.
There are situations in which you have to transliterate, rather than transcribe, because you don't know what language it is. For example, it's a name in a list of names of people from different places.