No. The implication of 'in' for AAs is that it is a fast, O(lgn) or better, operation. Your assertion requires O(n) performance.
What about making it work for (builtin) tuples? Then it could be make use of a variadic std.algorithm: find() or some lazily defined hash-table.
Maybe it could be part (if not already there) of Kenjis great tuple improvement PR which is awaiting review and merge.
