A title becomes a slug. An editor sends a selection to an agent. A route becomes HTML. These sound like separate projects until something goes wrong and the same question turns up: can I see what crossed the boundary?
The Auto Gen extension makes a field rule visible where the editor is working. Wayline has a way to inspect the context before a model sees it. Zenith makes a route a concrete output I can open. I keep returning to these seams because the ability to inspect a result changes how I can work on it.
Build a small thing you can check
The first useful demo of a field formula can be one title and one slug. A project memory example can be a stopping point and a next action. There is time for the longer story after the small example holds up. If I cannot explain what went in and what came out, adding a more polished presentation usually makes the problem harder to find.
Leave the next person an entry point
Writing my own tools also means owning the places people get stuck. A precise error, a runnable example and a route that still opens matter alongside the larger architecture. Documentation has to explain the path the tool actually takes.
I try to bring the same habit to this portfolio. The case study is more useful with an image of the running thing, the image is more useful with a caption that says what to look for, and the caption is more useful with a link to try it. That is how a piece of work becomes something someone else can inspect.