I probably won't be able to review this for a few weeks.

Alright. I'm back and starting to give this a look.

I think these are distinct work items that should be considered separately. Trying to review them together muddies the picture of what's going on with either

I see motivations for skipping skip messages in #453.
Could you explain the perceived fix to FsCheck replay?

Yeah my bad should have made two PRs. I updated the description with links to the diffs for the two changes

Could you explain the perceived fix to FsCheck replay?

In FsCheck3 they seem to have added (or maybe factored out) a size parameter for replay in addition to the seed. Without the size, replay was playing all test inputs leading up to the failure; this change allows skipping directly to the failing test input

That makes sense. Skipping right to the broken case would be good for debugging.

I noticed that you also changed from printing the original seed to the final seed.
I'm unsure how that impacts users focused more on deterministically repeating a specific run.
Expecto itself has some tests that use the seed that way

Ah didn't think about that. I added it to the output