It is better to evaluate the engine yourself for the project you are planning to do with the requirements it has (not making the entire project, of course, but if you - say - want to create a huge detailed world you can procedurally place entities in the world and see how the engine copes with that and if it provides tools to address any issues).
but if you see a similar game already made, then you can clearly claim that it's possible in the engine. Saves time and effort.
Really? What's the killer app for Dark Basic? I've never hated an environment more than that one. I would be shocked if anyone could put up with it long enough to make something good.
Most game engines have no games made with them - including my own garbage engine that I'm hacking on right now. I think you absolutely should and must judge an engine by, at the very least, whether or not games can be made with it.
And seeing what kind of games are made with it can show you what the engine is capable of, how well it performs, what sort of games it's optimized for, what the structure of its output is and how restrictive it is in terms of proprietary formats, what the development workflow looks like, etc. All of that requires example implementations to study.
Just to clarify (being one of the Godot devs), Deponia was not originally developed in Godot, but in the Visionaire engine.
It's only the iOS and PS4 ports of Deponia which were made using Godot.