Like most search engines, Kagi only let you search the semantic content of web pages.
It would be useful to have a mode that instead would let you search for raw HTML tags and attributes, even those that don't directly affect the visible content of the web page. Here are a few examples of things you could do with this feature:
- Find all the pages that load a script from a given host
- Find all the pages that load a script from a given path
- Find all the pages that iframe a specific page
This is basically what https://publicwww.com/ is supposed to let you do, but it costs a lot and its trustworthiness is questionable (how fresh and exhaustive are their results?). Hopefully Kagi could do it at no additional cost, and with the quality we've come to expect from the company.
It could be done through a new search operator, like tag:
The syntax could be as simple as tag:/hcdn-cgi/jschallenge, which would search for the /hcdn-cgi/jschallenge string inside the attributes of any tag.
Or it could use a selector-like syntax, like tag:script[src*="/hcdn-cgi/jschallenge"], which would search for any script tag with an src attribute that contains the /hcdn-cgi/jschallenge string.
