I understand and respect your opinion. However, all the examples you’ve given are more on the “passive/listening” part. I was referring more to new use cases when you actively create/update/remove channel point rewards, and update their statuses (fulfilled/cancelled) based on some logic. Using mock data just for triggering whatever event is okay, but in this scenario you can’t simply test creating/updating new custom rewards, for example.
That said, I’m sure you’ve worked on this longer than me, and better than me (no irony at all). And for sure there are workarounds and all that. Personally, I haven’t had this issue until now either, but this time it just feels more blocking than in the past. As I said, even though obviously possible, testing pre-production shouldn’t rely entirely on mock data imho.
But it is what it is 
For reference: feature request in UserVoice for those interested in upvoting: https://twitch.uservoice.com/forums/310213-developers/suggestions/41270905-provide-a-way-for-developers-to-test-features-whic