First of all, thanks for trying it out and for your feedback.
The mod is basically an early beta that still needs tons of work (and time that at the moment I haven't) in terms of content and balance, that's why it's still hidden in the dark alleys of the VP discord.
You are right that there is not much to do, and this highlights an evident trade-off between historical accuracy and engaging gameplay.
Prehistoric life is mostly survival and exploration to some extent, and very very very slow progress in knowledge acquisition and cultural progress (that's why you start at 3 million BC and between turns tens of thousands years pass, at least at epic that's the only game speed calendar I have more or less managed to balance so far).
I tried to represent survival struggle with an unfair battle against a stronger and vicious lion unit (the only animal unit I added btw). I can add a reward like yields bonus or a grant "hero" promotion to the killer, there are many possibilities for this.
On the other hand, I had to think of some gimmick to represent hunting-gathering, so we have the forager who basically constructs an invisible dummy improvement that removes itself on completion on visible resources in exchange for food and production/gold per turn.
I cannot represent one crucial element that is nomadic lifestyle because without founding a city you can't even open the tech tree. Implementing a nomadic mechanics would be neat but unfortunately it requires heavy DLL coding I don't have skill for.
Then there's the whole matter of balancing civ unique abilities.
For now I addressed only India, whereby they don't get the pantheon at start but at Burial tech (that is when the first Faith building unlocks).
In a polished version, you should not be able to "bug" the game using a certain civ because their gimmick would be postponed until the Upper Paleolithic tech tier (the column before Agriculture) or Agriculture at latest, in order to avoid unbalance and snowballing (and also to be historically accurate as human cultures started branching out around 80-40 thousand years ago).
Then there's the mystery of civs getting 2-3 policies around the time the human player has researched half Upper Paleolithic tech tier, which let AI to snowball like crazy.
One of the reasons I think is the opener of Progress that gives Culture from researching techs. Given there are double number of techs you get double yields.
In an erlier version of the mod it was even worse: once you clicked the progress opener, the culture bonus was so much you got another policy for free (so you go from 0 to 2 policies in one turn).
In the version you played I already tried to sandbag the problem by nerfing the culture bonus to 1/3 (Policy_YieldFromTechRetroactive is 5 instead of 15).
Other reasons may be the Dolni Vestonice wonder that has a filled Great Work slot (i.e. +3 culture) on construction, or the opener of Tradition, or a friend/allied cultural CS.
All in all, I realized that the way I would like to represent Prehistory is inevitably constrained by what the game allows me to add and bend to my will.
It's the whole "slinger compromise" all over again. It's not historical accurate but it's necessary for a better gameplay.
That's the main reason I posted the early build: to have feedback from players so we can go together towards a playing experience that is enjoyable.