Application of `validation-no-executable` label is not intelligent enough

Brief description of your issue

The label validation-no-executable is flagging perfectly valid apps that are webapps and have an entry point either of a JavaScript file or an html file (usually index.html). This is wasting technician time, since someone has to add a "waiver" to each of these manifests.

Steps to reproduce

See https://github.com/microsoft/winget-pkgs/pull/27338 for an example of a package that is/was flagged. Another one with some discussion: https://github.com/microsoft/winget-pkgs/pull/25453#issuecomment-903003116

Expected behavior

As these packages are fully validated and install correctly, I would expect the bot to recognize them as web apps. In my case, they are WinRT-based web apps, though there seem to be many other classes of app with this label applied. (In the future I will transition to running these as Store-based PWAs, but for now this gives the widest compatibility -- I suspect PWAs will have similar validation issues as they have no executable.)

Actual behavior

The validation bot cannot find an executable, so flags an error. The definition of "executable" should be widened to include HTML/JS-based entry points, WinRT-based HTML apps, and PWAs (if not already recognized).

The issue names no implementation files or tests; start by reviewing pull request 27338 and the discussion on pull request 25453 to trace how validation-no-executable is applied. Compare the validator's behavior for the linked web-app examples, and consider the work complete when valid HTML/JavaScript, WinRT-based HTML, and applicable PWA packages are no longer incorrectly flagged.