I disagree on both points, for several reasons.
Concerning the direction of the arrow, I’m not going to argue whether the original concept was to represent data flow or not, I’m just going to point out why the original version isn’t useful.
We’re presenting a tutorial for new users of a tool. The illustration is intended to help them understand the relationship between these entities that they know little about, using mechanisms that they might be familiar with. Perhaps the most common graphical notation that someone might be familiar with is UML, which has a “depends” notation that consists of an arrow between two boxes, where the arrow represents the “depends” relationship. That is an exact fit for what we are trying to illustrate. It may be that under the covers, there is some sort of “data flow”, but I would say that would be much less visible to someone trying to learn this tool.
Concerning whether we need an image to convey this or not, outside of any technical arguments for what the image should look like, it’s valuable in a document like this to have more than just textual information. It’s hard for someone to slog through a large document that consists entirely of text. It’s a good idea to give them different ways to see the same information, if only to keep them awake.