It's funny because stable diffusion actually has the tools to do all of this-it's called inpainting and img2img generation--but it requires the user to have an artist's gaze and tastes and often requires nearly as much work in refinement as painting it in properly would have.

People who think only about "prompts" can only create nearly random "cool images", the processes needed for refinement and consistency are all there but will require patience and dedication to learn that almost no one who wants to do this but refused to pick up a pencil will actually have.