It's probable that your battery box will already contain a suitable blocking capacitor, but for the sake of safety, we'll assume it doesn't.
Besides the physical issue of connecting your battery box to the D:vice's microdot connectors, the main electrical issue here is the lack of any kind of firm specification for the D:vice MMA-A; it's all marketing-speak! Anyone who has one of these, and a cheap digital test meter could measure these parameters. But for now, an educated guess would be that it approximates the plug-in -power inputs on radiomics and small recorders.
These standard PIP inputs load the mic with a resistor which works out at typically 1k / volt. So for 5V power, that would be 5k. This defines the input impedance on this kind of mic input.
On this basis, assuming you want the low frequency rolloff to be about -3dB@20Hz
rolloff frequency = 1/ 2*pi*R*C.
So for 20Hz rolloff & 5k input impedance that would be a capacitor of 1.6uF (minimum).
Standard values to choose from would be 2.2, 4.7 or 10uF. I wouldn't go higher than this.
Don't loose sleep over the capacitor type! The small can electrolytics are fine. Pick a reputable brand. Panasonic, Nichicon, ... &, for lowest measurable distortion, select one with a considerably higher voltage rating, say 25V.
The capacitor needs to go between the battery box and the D:vice.
Connect the equipment ground (shield) terminals together.
Put the capacitor in series with the centre (hot) terminal. This will block any DC from getting out of the d:vice.
IMPORTANT >> Ensure that the + end of the capacitor goes to the D:vice!
I hope this generalized reply is of some use. There are a lot of variables here, but this suggestion will at lest prevent anything bad from happening!