Hi Rob, good catch, and yes, that's really how we run: seven scheduled agents every night, each with its own persona, its own model and its own memory of the previous run.
We're going to write it all up in a dedicated blog post: how the night is organised, which model does which job (and why we don't use the same one everywhere), how the agents hand work to each other, the rules we had to learn the hard way so that run N+1 doesn't redo run N, and the actual prompts, so you can copy them.
We'll notify you on this ticket as soon as it's published.