This uses the GNU wget to spider the U.S. Navy's public instructions website ("Department of Navy Issuances") for PDF files, and downloads them (but no other files) into a dedicated directory.
So the Navy's instructions website uses Javascript and wget isn't a big fan of that. The quickest way I found was just to use Chrome DevTools and scrape out the list of URLs after paginating manually, by running this in the console:
x = $$("tr a.ms-listlink[href$='.pdf']").flatMap(x => x.href).join("\n");
Once run Chrome will prompt if you want to copy to clipboard, which is what I did (to then paste into the file). At the time I did this, that meant 3 separate page loads to get the ~1100 instructions in their table, 500 at a time.
The [href$='.pdf'] part is to ensure that the non-canceled instructions are
chosen only (canceled instructions still have a hyperlink but it seems to
normally go to a .docx file instead explaining that the instruction is
unavailable).
Of course, there are also non-cancelled but still not-public instructions. These still link to a PDF, but it's a PDF saying that you can't have this instruction. Luckily the Navy has adopted a sane naming convention for instruction filenames, so filtering these out is relatively easy as well.
There's probably an easy way to do this using sed(1) but I just ended up
using vim to find and delete lines matching this regex: \/[CSF][0-9][^/]\+$.
That is, searching within the last path component of the URL after all other
/ characters, find filenames starting with C, S or F and then a digit. These
are the instructions that are Confidential (C), Secret (S) or For Official Use
Only (F) respectively.