Skritter has a lot of recordings in the system that I think could be put to better use than simply playing alongside the flash card prompts.
I am thinking of something like the pinyin prompt on the webapp, where a sound clip would be played, and the user would have to transcribe the sound into the text box. I think it could be a great help for people without day-to-day exposure to Chinese and Japanese, although I have never met anyone who found Japanese pronunciation too difficult, but I digress.
Personally I would really like a way to train my tones without getting hints from characters or translations. Just plain sound-to-text/tones transcriptions.
I have no desire to have this integrated into the already established SRS-system, but if user data were to be collected, it could give users a clear view of how where they had problems listening wise, and I am sure Jake would agree that this sort of data would be valuable from a scientific perspective. I think Skritter is in a great position to collect this data, and perhaps make money off of it.

I realize that this will probably take a lot of work to stitch together, but I sure am not going to do it myself, so I just wanted to suggest it, and if the team sees it as a business opportunity, then that's awesome.

Would anybody else be interested in this sort of thing?

Yes, I'd like a voice prompt in order to write a character set or phrase. I don't see how this could work with just one character because so many sound the same.

+1 for me

There are 4 modes available (iOS - chineese for me)in advanced study panel : Writing, reading, Tone, Definition

I suggest to add 3 new modes :

Writing from Audio,
Definition from Audio,
Tone from Audio

"Gameplay" : you push the button to listen the input as many times you need and try to write, find definition or tone according to validated modes. You self evaluate your result as usual.

It's already possible to study like this right now, using the listen button. But your progress in "audio understanding" is not tracked. Would be a nice improvment ;)

Just to clarify: That is not the same thing I was asking for. I just want to practice Listening > Pinyin + tones, regardless of whether I know the word or not.

Laspimon's idea sounds more doable, simply because Chinese has so many homonyms. A given syllable can correctly be written a number of different ways depending on exact meaning.

A way around this might be to hear a sentence, and then be prompted on one of the words in the sentence.

I also think an exercise where you are prompted to identify tones would be great.