Glad to help. Now that I can resume my posting, I'd like to share the said NASA wind tunnel data and the formula behind such opposite rolls.
Apart from the canted vertical stabilizers you explained earlier, a rolling moment is also created due to how the force is being applied at a higher position relative to the aerodynamic center, which acts as a pivot point. This rolling moment (coefficient denoted as Cl-δr) is opposite to the direction of yaw.
But then as sideslip builds-up, the rolling moment by the rudder can be cancelled and overwhelmed by another rolling moment created by sideslip (β) due to lateral static stability (Cl-β), which is in the same direction of yaw.
To simplify a bit, you can calculate the combined rolling moment coefficient (Cl) at a specific AOA, by (Cl-β * sideslip angle + Cl-δr * rudder deflection).
So at low AOA, it's close to -0.1 * sideslip + 0.02 * rudder deflection. As you can see it can be negative, and also can be positive, depending on how many sideslip and rudder deflection is combined.
DummyCatz's post in Q: Why is the Hornet's rudder the only one that works the other way around? (below ~4.7 AoA) was marked as the answer