Note INIs that are compatible with Skull's co-driver, together with samples, because why not.
There are scripts .mk_template and .mk_codriver, found in Plugins/Pacenote/Scripts/. Go to a codriver callout directory, for example Plugins/Pacenote/sounds/SabrinaInkwell3 and call .mk_template from there. It will generate a list of things a human needs to say with at least 200 msec pauses of silence between callouts to create a voice bank. Given a properly prepared OGG voicebank, (mono, 11025 Hz), run .mk_codriver script in a directory with existing .template directory to get a .codriver directory with properly named callouts, extracted from the aforementioned file.
I used the following command to convert my Audacity-exported OGG into correctly sampled mono OGG: sox ~/winhome/Documents/Raw\ Rawness.ogg -r 11025 -c 1 .template/jonn.ogg.
We're in progress of implementing boost_volume_by and normalise_callout_file scripts.