Sorry it's not a Kyma solution, but if you know how to use Python, you will also need to install numpy and librosa.
In the mean time, it would be more convenient if your sample files are mono. Multichannel sound files are stored differently based on the type of files (wav, aiff, etc..).
Here is the code you can use:
import numpy, librosa
x, sr = librosa.load('soundfile.wav')
numpy.savetxt("soundfile.csv", x, delimiter=",")