You point out that there are acts easily agreed to be evil and acts easily agreed to be good, but that doesn't imply a definable boundary between good and evil. First postulate a boundary between good and evil. Now, what is necessary to refute that boundary? A clearly defined boundary would require actions that fall near the boundary to always fall to on side or the other without fail. Easily, that is not the case. Stealing food is clearly evil if you have no need but the victim has need for the food. If the needs are opposite, then it is not clearly evil. So there is no clear boundary, but what would a vague boundary require? A think a vague boundary requires that actions can be ranked in a vague progression from "certainly good" through "overall good, slightly evil" and descend through progressively less good zones as they approach from one side, then crossing a "evil=~good" area, into a progressively more evil side. I do not see that is necessarily the case.