After just reading most of the first half of the docs (skimming some parts, because in occasions it delves too much into fine details, I feel it lacks a bit of conciseness), I'm still not sure I understood well why this tool needs to use specifically built images, instead of just grabbing whatever is the current `ubuntu:20.04` image and install stuff on top of it.

The page didn't achieve conveying this clearly to me. I might be a bit obtuse today but as I went on and on reading, the question remained about why exactly this tool is not just starting from well known (and trusted) Docker images. The "trusted" part is important: I kind of trust that official Ubuntu images are not tampered with anything fishy.

Could you expand on this?