Based on a script by @jrickmd this can be done with Youtube-Videos.
There has to be some way to accomplish it with audio podcasts. But the RSS feeds I typically see will download a short abstract of the audio content, not the actual audio file which is linked somewhere in the text-based contents.
I wonder if a smart rule could check the content of a html or pdf article for links containing “.mp3”, extract them through a RegEx pattern and then add them to downloads.