Interesting thread. I take the rather prosaic view that fish recognise some things as food, which might be a combination of smells/odours/colour/appearance and these can be divided into:
‘completely natural foods’, that is, insects, crustaceans, worms etc.
‘food fish have learned is food’, so bread, sweetcorn etc.
‘things fish readily accept as food’ as they smell/taste a bit like food, due to common odours or scents – so for example, cockles are enough like natural food they are ‘instant’ baits, despite coarse fish never seeing a cockle in the wild.
Once a fish has been trained to eat a new food in any category it’ll tuck in until it’s been trained not to (by being caught). It may forget that danger association in time.
If, say, the bait is plain bread, adding custard powder isn’t magic, it’s just a slightly new sort of food, retaining enough of the signals to be recognised as food but without a previously learned association with danger.
It’s possible there is some ‘magic additive’, but it seems more likely that ‘additivities’ simply change a food associated with danger to one that is not so far associated with danger, although it may also be the ‘additive’ might simply dissolve well in water and spread the scent around a bit, making it more likely the fish will find the bait.
Any additive, if used by everyone, would probably become ineffective eventually.