The term "primary key" is used in the spec in a couple of places:

(1)  3.2.6 Index
"getKey - Gets the primary key of the record from the referenced object store entry."

In this case, the word "primary" can simply be dropped; it's already referring to the referenced object store, and the algorithm clearly defines what should happen.

(2) 3.5.7 Cursor Iteration Operation
"If /found record/ is not defined, set /cursor/'s /key/ and primary key to |undefined|."

This blurs the line between the IDBCursor (the IDL interface) and the abstract cursor mechanism defined in 3.1.10. IDBCursor has |key|, |primaryKey| and (for IDBCursorWithValue) |value| attributes. A cursor has /position/, /key/ and /value/ properties, and defines /effective key/ (if the source is an index).

The algorithm should not refer to the IDL attributes, but the properties of the cursor mechanism. 

Given that:

"IDBCursor.primaryKey Returns the cursor's current effective key."
and:
"If the source of a cursor is an index, the effective object store of the cursor is that index's referenced object store and the effective key is the cursor's object store position."

.. then it should be sufficient to state:

"If /found record/ is not defined, set /cursor/'s /key/ to undefined. If /source/ is an index, set /cursor/'s /object store position/ to undefined."