Yup. My relay accepts all kinds (for me at least). 

 Definitely. I do the same.  

 No, this is a legitimate problem. The nostr spec NIP-01 specifies the valid range of kind numbers. All kinds in that range should be accepted. If relays block kind numbers because they are not "known", it completely hampers innovation and keeps the network from growing. I'm curious which kinds are being rejected nostr:nprofile1qqsqddupn4l3cl65wggcyehd009g0pwuatsfudh28f90vewx68vrylqprpmhxue69uhhyetvv9ujuumwdae8gtnnda3kjctvqy2hwumn8ghj7un9d3shjtnwd9kk7tnrv9eksqg5waehxw309aex2mrp0yhxgctdw4eju6t0pqpyr2?

How are kinds established as "known"? They get into a NIP. How are NIPs made? They start as a draft. How are drafts approved? BY IMPLEMENTING THE PROPOSED NIP. You can't implement something new unless it is accepted by the network. This is a circular dependency. This is why new kinds that are unknown must be accepted if they are valid as per NIP-01.

Relays that exclude valid kind numbers are violating NIP-01! 📣 PASS IT ON https://github.com/nostr-protocol/nips/blob/master/01.md#kinds 

 Fair points, but I also think that it's also a fair point to have a development relay that you can spin up and nuke as part of the development process. I don't like rate limiting and I think 5 per minute is nuts, since I do that many in 10 seconds when I'm just doing reactions, but I also understand the pain points of people running infrastructure. 

 Yeah I get rate limiting. That's understandable. It's just blocking valid kinds is a bad thing to do 

 Negative kind numbers are blocked from most relays.

Many will block outside of the documented set. 

A few relays only allow specific kinds but at least they document that. 

 n...negative kind numbers? https://media1.tenor.com/m/dCspeNeobQkAAAAC/mcdm-lord-durok.gif 

 I do all my dev testing on my own relay, and then periodically test other random relays to see if it works.