Augment frontends are intentionally kept very simple in order to avoid duplicated effort. In order to implement them, you have to do at least three things:

  1. Parse the layer files from JSON for the current file.
  2. Render the layers in the files they apply to. This should be implemented as just changing the background color of a range of characters.
  3. Allow the user to see the message associated with the layer in which the cursor currently resides.
  4. Bonus: send the file name to augment --interactive when a file is saved and parse the response.
  5. Bonus: allow the user to specify which backends they want to use, preferrably on a per-file basis.

That last step will become more important as more backends are implemented.

TODO: explain more stuff!

Be sure to check on the mailing list to coordinate work for your editor. Someone may already be working on it.