This a new feature and Tinderbox is, at present, very conservative about the frequency at which Taggers run. It may be too conservative. A concern is that not all Macs have dedicated neural engine cores, and setting up a tagging scan requires a certain amount of prep work. The general idea here is that you want rules and agents to run now, but taggers are a sort of background discovery task.
The original notion for taggers is that they are slowly mining data and marking down relevant things they discover. They are envisioned like good research assistants; they may not get things done right away, but eventually they will get answers that are interesting and usually useful.
As taggers all run on document open, saving and re-opening the document is a reliable way to force the process to update on demand, at least until this feature evolves further control methods. Editing a note's text ($Text) will also update the tagger. However, it will not update the displayed note. So, to see the effect—e.g. in the Displayed Attributes table—it will be necessary to re-select the note (or click the current tab's label to refresh the whole tab display).
Another way to force everything to be retagged is to edit the $Text of the tagger-defining note. However, tagging tasks and indexing tasks might be delayed in large documents when lots of things are going on, e.g. complex actions in rule or edicts using a lot of regex tasks, or very complex find() queries.
Editing a tagger note marks the tagger as needing to be reloaded when the index is next revised.