On Apple devices that wouldn’t be an issue because you you can load the Twitter Web interface into a real Safari and get the markup from there for processing. Twitter can try things like CAPTCHA but this will degrade the experience for all Twitter users and you can simply delegate this CAPTCHA to your user to solve anyway.
While for reading that works, there is an easy why via posting to spot people use that app. They can change at any moment the input names so that when posting a tweet you are exposing that you are using the scrapping client.
Yes, but the app is not a human and will likely interact with the website in a more predictable (and thus, detectable) way than the human itself would.