We could leverage schema level triggers to avoid time-consuming parsing, when scanning schema for suites.
If we would use triggers after create/alter/drop on schema, we could initiate parsing for a particular object and update annotation cache.
That way we minimize the impact on users when using context-menu in SQLDeveloper on large schema.

That feature could be installed/enabled optionally (decission at install time).
Alternatively, utPLSQL could try to create a trigger on user-behalf when running a scan on a schema.

Regardless of apprach, we should make sure that when utPLSQL gets removed, the leftover triggers will not cause any issues to schema where they were created.

There are few more things to consider however before we happily start implementing it.
Some resources:
https://docs.oracle.com/cloud/latest/db112/LNPLS/create_trigger.htm#LNPLS01374
https://asktom.oracle.com/pls/asktom/f?p=100:11:0::::p11_question_id:267415465220
http://psoug.org/reference/ddl_trigger.html