With file:// one can supply a file name which contains a URL per line.
--single-threaded disables parallelizm.

Most of all, abstract HTTP Client.
Test reporters.
Test using mock http client.
Test recursive and non-recursive.
Test Axios client separately.

Add verbosity (with dots and x).
Add GET vs HEAD (retrying HEAD).
Add more prefixes to ignore.
Add JUnit reporter (test on GitLab).
Add option to ignore skipped URLs.
Check URL before running inspector.
Code up proper timeout.
Add <iframe src> match.
Suppress STDOUT in tests.

Download HTML from URL.
Extract URLs.
Remeber parent-child relationships.
Recursive go withing same host.
Skip checked.
Catch errors.
Compile report.
Run asynchronously.

Store coverage HTML artifacts.

Parse HTML and extract a few types of links.
Has tests, coverage and debugging scripts.