This is a collection of ideas and Howtos which were either very surprising to me, took me a while to find or even needed quite some work to find out.
IntelliJ allows you to define external tools which you can call after selecting e.g. a file and pass that file as command line argument "$FilePath$" . Interestingly, it also allows prompting you with "$Prompt$" for an additional text to be included into the command line. (There are many interesting other macros as well).
This is particularily interesting with createJavaScript, which you could give the arguments
"$FilePath$" "$Prompt$". But doing the same thing with chatgpt itself, using the arguments -f "$FilePath$" "$Prompt$" gives you a quick way to present ChatGPT with the contents of a file and have it do something, like
creating another file with examples taken from the given file.
ChatGPT has good CSS knowledge, but doesn't have (yet) the ability to view and understand a webpage. And if you are developing a big application, the CSS and HTML are usually too large to present it to ChatGPT. So if you need help with a detail: you can copy the HTML of an inspected element in Chrome devtools using "Edit as HTML" from the context menu in the elements inspector, and use the "CSS Used" Chrome extension to copy the CSS that apply to it and the subelements. That seems small enough to give to ChatGPT, though it's some work to translate the advice back into the original CSS files.
(There is also SnappySnippet extension, whose CSS extration is not exactly helpful - it gives every element an ID and collects the effective styles of the element, massively repeating everything.)
https://stackoverflow.com/questions/9196451/how-to-export-all-relevant-html-css-for-one-element https://stackoverflow.com/questions/8407550/how-to-get-and-extract-matched-css-rules-on-dom-node