The difference between a skill and a request

A request is what you type to your AI in the moment. "Turn these notes into a five-slide presentation" is a request. A skill is a set of directions, installed beforehand, that explains how that kind of job is done. One says what you want, the other knows how to do it.

Picture a restaurant. The request is your order. The skill is the recipe in the kitchen. You place the order again every time you visit, but you never have to explain the recipe to the cook, because it is already in the kitchen.

Without a skill, getting a good result means putting everything into the request: text size, color scheme, page structure, what to avoid. You save these long requests somewhere and paste them in each time. One day you forget a line and the result changes. With a skill installed, those details are already in place. You only describe the job itself.

There are three practical differences. The first is permanence: a request stays in that conversation, while a skill installed in a tool that can write files is ready in every new session. The second is scope: a request is a few sentences, while a skill can carry pages of detail, templates and helper programs. The third is consistency: work done with the same skill looks alike, because it all follows the same recipe.

This does not make the request unimportant. A skill knows how the job is done, but it cannot know your topic, your reader or your goal. You still have to say those. A good result is the two together: an installed skill and a clear request.

Here is a simple way to tell which one you need. If you will do a job once, a well-written request is enough. If you do the same kind of job again and again, a report every week or a presentation every month, installing a skill saves you from retyping the same details and helps the results stay at the same level each time.