It's a joke, but I sometimes honestly wonder why no one has made a game based on OpenMW. Then I remember that making a game engine is no longer the bottleneck, as Unity and Unreal are essentially free.
Multiplayer games require little content relative to play time. Single player do though and thus there are only very few open source ones.
I found that even with the simple set of features I was really pushing some of the limits of the scripting system, (there's a reason I referred to the forge as "the bane of my existence"), and some of the things I really wanted to do like generating new items from scratch weren't entirely possible.
But as an engine, OpenMW is actually one of the nicer experiences I've had.
(Note on building towards - Items and other objects were slowly being replaced, with the intention to eventually be able to start from scratch, but sourcing art was the hardest part, so starting with a total transformation seemed wisest.)
Certainly people have for a long time licensed engines made available as a side effect of the creation of a certain type of game to build further on what's been accomished, or to cash in by making derivative games. For example all the shooters made with the Quake engines, which no one would call general purpose. There's a few examples of reused RPG engines (e.g. Infinity Engine) and RTS engines, too. There's also plenty of toolkits dedicated to adventure games and visual novels.
Arguing growth is possible is not outrageous.