We found a discussion on Twitter about this topic and we thought it would be useful to provide to provide the correct technical background related to whitelisting.
Whitelisting purely on known headers certainly would be a non-optimum action, which is why it should be noted that the article referenced has a very large notice at the top which directs people to use an alternative method that does NOT include headers. If someone does want to whitelist by headers we will support that but in that case we recommend using the custom header options. This is the box at the top of the page:
Of course another method is to whitelist is with direct message injection where the headers are effectively fictitious but that's a different discussion. Hope that clarifies this some more. :-D
