A Skill has a short summary up front — what it is and when to use it. Thanks to that, the AI doesn't have to read the whole thing to tell what a Skill does or when it applies; it can pick the right one out of many just from the opening.
This "you can tell the gist from the top" structure isn't limited to Skills. Put the same kind of overview-and-when-to-use block at the head of a context file you feed to the AI, and it's more likely to get by reading just the opening instead of loading the whole file — which keeps context consumption down.
Input context windows in today's major LLMs keep getting larger. Even so, you want to keep as much room as possible free for the main job. The habit of not reading what you don't need to read pays off even as capacity grows.
One caveat: to make this work, the AI has to recognize that "this file is structured with a summary at the top." The structure alone gets skipped over, so you also tell the AI that reading the opening is enough — and that's the other half of the pair.