That's the thing that makes me scratch my head about the very premise of the OP.
If they really want to write their own thief-less RPG they should absolutely write it however they like. But I find this attempt to find support in rationalizing that design as somehow more logical or superior strikes me as quixotic.
As best I can infer from the OP's various posts about the RPG they are writing (or have written), it seems to use magic-wielding classes to achieve genre-appropriate fiction without relying on metagame mechanics or other action-resolution and framing devices that don't directly correspond to in-fiction processes. In that sort of game, if it otherwise hews to fairly typical fantasy tropes, knights will not be fully mundane, and - almost by definition - no fully mundane character will be suitable to be taken up by a player to occupy a protagonist's role.
If I'm correct, then in repudiating the logic of "thief" as a class, the OP is just reiterating their underlying design principles and drawing out a fairly straightforward consequence of them. But I don't think that that has any broader consequences for RPG design.
In a class-based game that identifies
fighting as a pre-eminent domain of conflict resolution, and that identifies
magic as a distinctive (both in the fiction and in the mechanics) way of acquiring abilities, then there will be conceptual room for a character who does not use magic, and is a second-tier fighter, but in exchange is good at "other stuff". (Of course that sort of character is always in danger of being overshadowed by magic.) In D&D that's the niche of the thief or "skill monkey".
The parameters that create this niche - to reiterate, a class-based game in which
fighting as a pre-eminent domain of conflict resolution, and
magic as a distinctive (both in the fiction and in the mechanics) way of acquiring abilities - are pretty particular. Which means that the "thief" class is also pretty particular. But perfectly sensible within its appropriate domain.