Orbit Droo with craft with 6,000m/s deltaV. Burn prograde and have a hyperbolic orbit with eccentricity greater than 3. The higher the eccentricity, the easier to see the bug. The attached save has a craft on a trajectory with eccentricity above 7. The Vizzy program is calculating the predicted orbital velocity at the edge of the Droo SOI. Timewarp until you notice the velocity length stops decreasing, but you are far from the SOI edge.
Notice how the actual velocity freezes above a certain altitude(depends on the value of e) This causes the predicted velocity value to change. The velocity vector updates normally if you exit timewarp. I believe this indicates the problem is likely in the timewarp code when dealing with high eccentricity trajectories.