I don't think it's possible to draw clear distinctions though, because it's not a philosophical line, it's a "look around us and see where we're at". And "where we're at" is going to change as time goes on.
You could make a principled point of saying you're not withdrawing ILR from anyone whose been granted it - but what if Labour tomorrow grants citizenship to 10 million people? Your principled point disappears because the situation has changed. The concept of "being a citizen" has been diluted, so the protections of being a citizen should be diluted accordingly.
Likewise 2nd gen criminals. This is something you can only realise after you have enough people to have a 2nd gen (see also, ethnic enclaves, discrimination by non-white groups against other non-white groups or based on caste back home etc). So you update your concept of how many 1st gen people you let in, from different countries with different 2nd/3rd gen risk levels, even if they're really good on paper.
In practice, how I think it will work is in tranches based on what's realistic. Cancel some visas, start deportation proceedings for criminals from countries where it's possible; then reassess and see where you're at. Maybe there's then a big group of non-criminal net recipients who would voluntarily leave with some changes to social housing or benefits policy, for example. Maybe there's a country that you can put pressure on to accept returnees, but it only comes through after a few years and is treated as too hard before that.
As an immigrant I know nothing is guaranteed, but based on common sense my personal profile is not going to be at the top of any of those priority lists. So let me admit to significant bias: it's therefore in my interest to argue that immigration (and remigration) policy should be targeted based on profiles. Make that of what you will as you read my posts - I know I'm not alone among my immigrant friends!