There are some six or seven hundred genera of grasses, and God only knows how many species. Of these, something like six or seven are eaten as cereals. Is there some botanical reason why these are the only grains widely eaten, or could any grass be bred into food, so a PoD of 9999 BC ATL may have krikittle, moophes and chisanno instead of rice, corn and wheat?
Yes. Because these were the grasses that were:
(a) easiest to domesticate; and
(b) give the highest yield.
Wheat and barley were easiest to domesticate, as were (in different climates) plants like sorghum in Africa.
Once a plant is domesticated, it tends to inhibit the domestication of closely related plants unless it offers some special advantage. Why bother going to the effort of domesticating a wheat relative when you have an already domesticated wheat right here? So there's a real first-mover advantage for domesticated grasses.
However, there's more to the story than that. Thanks to modern genetics, when can now domesticate just about any grass we choose (with perhaps the occasional specialised exception). But that doesn't mean that the plants will yield any better than existing crops.
The existing domesticates have thousands of years of selective breeding to increase their yield, and in many cases they already were the best-yielding crops to start with (e.g. wheat, barley). So a new domesticate can't really compete with the existing major staple crops, except sometimes in niche roles.
Mind you, it has been possible historically for some new domesticates to replace older domesticates, if their yield is higher. The classic example is rice. In early East Asian farming, millets (small-seeded grains) were the early domesticates. But they were largely replaced by wheat and rice as those crops gave higher yield.
There are also examples of domesticated cereals which were abandoned entirely (in both North and South America) with the arrival of better-yielding domesticates.