Variables: Export Data Sets marks visibility and image columns with # and @

Squashed from #5141 onto main.

docs/mcp.md

@@ -1646,7 +1646,13 @@ already walks Image Trace. Short agent entry points:
   first free names `Data Set 1`, `Data Set 2` and so on. An import error about a row's values or name gives the file
   line the row starts on. A header cell that names no variable defines one, unbound, in the same undo step as the
   rows. `#Shown` defines a visibility variable `Shown`, `@Photo` defines an image variable `Photo`, and a cell without
-  a prefix defines a text variable. A prefixed cell whose name is already a variable of the prefix's kind names it;
+  a prefix defines a text variable. `dataset.export` writes a visibility variable's header as `#Shown` and an image
+  variable's as `@Photo`, and a text or graph variable's as its name alone. Its reply's `warning` names the variables
+  whose header or values make an import into a document without them refuse the file (a `%` header, a `#` header
+  whose values aren't true or false). Otherwise it names the variables that such an import does not read back with
+  their kind: text variables whose names start with `#` or `@`, and graph variables. Export is refused, naming both variables, when a variable's header is
+  another variable's name (a visibility variable `Show` next to a text variable `#Show`). A prefixed cell whose name
+  is already a variable of the prefix's kind names it;
   one whose name is a variable of another kind fails the import. A `%` cell that isn't a variable's exact name (graph
   data files) is refused, and so are two cells that name one variable (`City` twice, or `Shown` and `#Shown`). The
   reply's `defined` lists the variables the import defined. While the variables are