Here's an article that talks about this phenomnem which is sometimes referred to as deep vee wander. The article suggests the problem is less noticeable in duo prop configurations but the admiral would disagree with that statement since she is well aware of the wandering tendancy of our boat in slow speed conditions. Perhaps a single prop has a tendancy to wander a particular way while a duo prop wanders either way.
http://www.boatingmag.com/skills/seamanship/controlling-low-speed-wander
In any case, there is nothing wrong with the OP's boat and the principle can be easlily demonstrated by trying to push (with one finger) a kids small toy boat slowly by the stern while trying to keep it straight. It's almost impossible! Try pushing a leaf; same problem. Interestingly, a piece of wood plank that roughly approximates a scale length to beam ratio of most of our boats will exhibit the same behaviour but a long narrow stick can be pushed in a straight line generally without a problem. This problem is about longitudinal stability and in most of our boats it would be reasonable to suggest that they are not designed with slow speed longitudinal stability in mind. Instead, they exhibit divergent stability where side to side oscillations get worse, not better. As an earlier posted pointed out, this is related to the pivot point of the boat which is generally where the most resistance to the water is. This centre of resistance is about the middle of the boat where the hull sides are deeper in the water, and the wetted surface is greatest. As the sternleg or outboard pushes the boat, any minor directional inbalance causes the bow to swing. Water resistance on the now slightly offset bow assists the swinging which occurs about the pivot point. The stern offers little resistance to the swinging since the hull sides (especially in a vee bottom) are often only kissing the water and the outboard or sternleg offer little lateral resistance. Although not related to slow speeds, aircraft are beset by this problem as well with many larger swept wing aircraft having yaw dampers which are sensors sending signals to a processor which in turn signals rudder actuators to counteract divergent longitudinal stabilty.
Overcorrecting for the problem only aggravates the situation. As with aircraft it has to be damped and that can only be done by pre-empting the swing which generally means steering non-intuitively in the opposite direction by trying to stop it reaching the heading you actually want it to reach. Once mastered, you can helm the vessel in a straight line but your passengers might notice alot of steering action going on by the skipper in order to maintain that pre-emptive situation. Any distraction or loss of concentration and she wanders again.
This whole problem goes away as soon as the bow lifts on power application and the centre of resistance forces head aft.
Terry