Okay a few things to unpack:
Adding code direct vs informing design.
Consent to have code released from mod authors and into the hands of the developers.
Focusing on the second point; I agree with Kulze, that an opt-in option would rule out issues of consent:
An Upstreaming Repository
Anego could maintain a public repo where mod authors submit features they think belong in core. Only mods intended for upstreaming would go there. Contributors would agree to hand over code and waive their existing rights in favour of eventual integration. Anego would decide what fits the game’s future vision, so even popular mods could be rejected if they don’t align with the team’s plans. It's important to note that a seperate repository is required, and not just a badge in vsmoddb, since we'd want automated testing that would be hard to ahieve from within a php website.
A brilliant point to make, I think that in this case Anego would set the terms in the licencing agreement for contributing to the upstreaming repository, it's really upto the mod author to agree/disagree. It would be unfair to force a mod developer to contribute, and I couldn't agree more that some would expect compensation and that Anego wouldn't be prepared to pay. I just hope people here are modding for the love of the game, and not personal gain.
I would expect Anego to not compensate monitarily at all, instead they might offer some devs a job. I don't personally like the idea of monetised mods, but feel free to disagree.
On the first point; Adding code direct vs informing design. As to whether the mod is deployed "as-is" or "informs design" should be left to the desision of the development team. It's their game, and unless they design the CI tests and a method to acually fold-in code, then it's unlikely they'd ever use code "as-is". On making the trasisition from Mod Author led to Game developer lead:
Proving out Mods as Modules with CI Tests
Instead of merging upstream mods immediately, selected community mods could ship bundled with official releases using the existing mod API. This keeps features isolated, easy to remove, and low risk. Lightweight CI could automatically build mods and run basic load tests to ensure they start and register systems correctly. Hosting these tests on GitHub would prevent mods that fail on updates from reaching release. Only a small number of mods should be included this way, since CI tests would need to check compatibility between them. If a bundled mod works well for a release or two and is popular with players, it could be gradually folded into core systems. If it doesn’t, it can be removed without extra maintenance. Upstreamed mods would retain visible credit in the game’s internal credits, with ownership transfer agreements already in place via the upstream repository consent form.