Fluffy Clouds is an ancient Perl script that i started using back in 2001 or something. I’ve tweaked it a bit to limit searches to parts of the site, normalize Unicode, etc. I’m not surprised the developer abandoned it. I keep thinking about switching to something more capable, but every time I look, i mostly find things that are way more complicated than I need.
As for how it actually works:
One script crawls all the html files within a folder and builds an index of words. I have it on a weekly cron job.
The other reads an html template with header and footer, searches the index, and plugs the results and search form into the template.
IIRC it skips meta robots pages on indexing and reads the file title and description at display time.