Per https://pypi.org/project/rfclint/, the actions performed are:

• Validate the file is well formed XML and that it conforms to the XML2RFC Version 3 schema as defined in RFC 7991.
• Verify that embedded XML stanzas are well formed.
• Verify that embedded ABNF is complete and well formed.
• Identify misspelled words.
• Detect duplicate words.

The following is a description of how the selection is done for what is spellchecked. The first thing to note is that this list contains v2 (RFC 7749) as well as v3 (RFC 7991) vocabulary.

Elements that are checked:

annotation, area, artwork, blockquote, city, cref, code, country, dd, dt,
email, keyword, li, name, organization, phone, postalLine, refcontent,
region, sourcecode, street, t, td, th, title, uri, workgroup

Attributes that are checked:

   'title': ['ascii', 'abbrev'],
   'seriesInfo': ['name', 'asciiName', 'value', 'asciiValue'],
   'author': ['asciiFullname', 'asciiInitials', 'asciiSurname', 'fullname',
'surname', 'initials'],
   'city': ['ascii'],
   'code': ['ascii'],
   'country': ['ascii'],
   'region': ['ascii'],
   'street': ['ascii'],
   'blockquote': ['quotedFrom'],
   'iref': ['item', 'subitem'],

where the first item is the element name and then the list of attributes that would be checked.