Essentially: Create lore/power systems that are vague and inconsistent. Its literally why historians exist as a job. Don't stop developing ideas and finding the consistencies. Get someone to infodump to.
Well, I'm still in the middle of making Ethra, a world building series, and how I decide what's best for my story is essentially what you do. I literally figure it out as I go. It's less pressure and it keeps my ideas fresh. Not to say creating an outline isn't helpful, but it doesn't really make room for creativity and I like to use it more as a rough guideline to direct my story. Outlines, in my opinion, aren't best for aha moments.
So, going off of that, I create my power systems and lore in a vague way and only reveal a element when I know for sure that it will be in the novel.
How do I know to include it? Well, I essentially compile all of the ideas I've had over the past (however long) and compare them. If there is an element that is universal, I know that it's the right way to go, as all my ideas have them.
As an example I've probably changed my power system hundreds of times, but what stayed consistent in the power system of Ethra weren't stat numbers or profiles, but magic power, paths, exclusive skills, path skills, and skills with a foil type power system, The wind's shadow. It's probably all gibberish to you, because it is gibberish to me as well. I just name the things that are consistent in my ideas and develop them as I go knowing that those have been with me throughout my numerous revisions.
For Lore, I like to build the lore then base the story off of that. I like to call it the Tributary method where I lay down my smaller world building stories before creating the larger narrative. ( I like to create new/uniqe methods, like question based poetry, which uses the sound of a question instead of rhyming. So fair warning: this is not a official method to my knowledge.)
However, in your case, you probably have the entire story drafted out, so my recommendation is to have characters with different views or to create hints about all of your possible histories, and depending on the main narrative, eventually find one that best fits. This simulates history where there are so many differing views that the main view is blurred, but in the end, the real history is what we know and believe, and have consisten evidence in. So I would recommend you doing something like that in your story. Ask yourself: "What elements point to what kind of history?" For example, if you have elite race elements, maybe stick with your ancient dragon's history. If you have a theme of change or evolution, maybe stick with your variations of ancient dragons' history. It will all depend on where you want your ending to be.
Also, if you haven't had already, go find someone to infodump everything and get their opinion. "The voice" in my story became something like a robotic "system" in manhua or a video game, to a part of a suppressed subconscious that is controlled by a higher being, which created paths for all sorts of other lore, like the creation of different beings in Ethra. This was all possible because of my friend's suggestion when I info dumped the story to them, and it helps a lot to have a different opinion as you can gain more confidence that it is the right way.