Pre-existing economic models of, and networks for, value/currency... are unlikely to be usable, since their motivations, valuations, and participation are not centric to or aligned with "code". ie: Though it maybe / is possible, to directly build and use a repo upon existing native ETH/BTC network, it would be futile due to high per tx fees and tx/day limits. Global democratization of coding, code, access, even grand socialized determination of dividend-for-accepted-work, etc... out to third world accessibility, requires repo operation fees to be trivial, say $0.001 per op, or even "free" depending on the network model. Either clone and tune an existing network that can work, or create one optimized as a global coding repo from scratch. A "blockchain on a git" is not a "git on a blockchain". A global repo that depends on DNS/dotcom services, that isn't replicated, under consensus, etc... isn't "permanent" and won't survive censorship's assault against critical application code and its metadata. PGP, DeCSS, MP3, and PopcornTime were examples of app code whose survival was under duress and thus forced to go mobile to survive. A global distributed repo running at a level of 1000+ repos, 500GiB, 1000tx/day, or larger will be quite hard to shutdown than Bitcoin since its primary purpose is not money but codespeech. More apps that need code censorship protection will surely be coming in the future. Github, and even git on an onion/i2p, neither being distributed stores, will hardly provide enough resistance.