But them getting a definition in the glossary, just like "saving throw", presupposes that ability checks are a standard and accepted mechanic. This is the real distinction between how they're treated in 2E as opposed to Moldvay. In 1981 they were an idea, probably one with some currency, but not universally known. By 1989 they were extremely common and part of the expected toolset for an AD&D player.
This is pretty hairsplitty, but I do see an evolution of their usage and expansion of their ubiquity.
Maybe this is worth a longer historical writeup, but while Moldvay put this down in writing, it's hinted at in OD&D (the LBBs) but never explained, and was somewhat common in the 1970s.
You can trace the "roll under" as far back as pre-D&D Blackmoor, when it was used for abilities, skill rolls, and even attacks.
In short, Moldvay helpfully included something that was assumed in play by a lot of OD&D players, but simply wasn't written as a rule. Because ... reasons? Writing is hard?