[art] How to do protocol-level research effectively?

["Soni L." <[e-mail]>]

Hello!

Most protocol implementations (HTTP libraries, IRC libraries, etc) tend 
to be fairly strict in how they expose protocol concepts to library 
consumers. This is fine, if all you're doing is using the protocol. 
However, it'd be nice to have libraries - or even a framework for 
building them - that focus on protocol-level research.

Effectively these would be built to allow all sorts of cursed 
protocol-level things. For example, let's say we wonder what it would be 
like if you could embed an HTTP session inside an IRC session; doing 
that with a "normal" IRC library is a pain, and even then we'd have to 
write our own HTTP library to handle it because a "normal" HTTP library 
wouldn't know what to do with it (while personally we'd use an HTTP 
upgrade to switch back to IRC, hooking that up with "normal" HTTP and 
IRC libraries is effectively impossible).

Does something like this already exist? Honestly as far as we can tell 
it does vaguely resemble monkeypatching but the goal is to be able to do 
protocol-level research with it and we feel like monkeypatching would be 
frowned upon even for that, so something built for the purpose seems 
more appropriate.

Thanks!

-- 
1st person: we/us
2nd person: it/its
3rd person: she/they