There's no downside to it being part of the API. Not making it part of the API is to basically implement a much slower pathfinder on the script side, which doesn't make much sense, in either cost or memory. A flood fill is just a clone of the existing pathfinding api(A*), without the heuristic(making it Dijkstra), and with different handling of the completion criteria, and of course returning a room sized array instead of an individual path.