###It is in the very weak side.
- There is no reason to use this coin if the player has the advantage, therefore, at most they would use it on a normal attack or, at best, they would save it when they are at disadvantage (since it cannot get worse).
1.1. In the long run, statistically speaking, if they use it on a normal attack they would have 0 benefit from it (50% of the time is good, 50% is bad) without sneak attacks.
1.2. The damage increase on a sneak attack attack would be negligible*. - Having advantage does not mean that the attack would succeed, just that it is more likely to succeed.
- Because of 2, If the PC/NPC have other things to do with its bonus action, it would be better to use those, particular at will.
- It is only once per short rest. It is too low to be useful in non disadvantageous situations, thus, waiting for one is not an exploit, it would be probably its best use.
*- If we consider a 50% of hitting a target and a 10d6 sneak attack, the damage once per short rest would be, statistically speaking, a 3.75d6 (~4d6 avg ~14, an off-hand bonus attack with a rapier at will will yield on avg 4.5 + 50% chance increase to apply sneak attacks). The damage increase would be too low for a Rogue, even if you math in the critical chance (~10% vs 5%).
###Balance issue
Statistically speaking, the advantages of the coin are extremely low, thus making the coin usable at least 3 per short rest, or 5 per long rest would definitely increase its uses.
###Discouraging uses on attacks with disadvantage
A way to discourage is to apply a double punishment (probably rewards too) on usages in determined situations. For example, adding a 1d10 damage (radiant, double on critical fail) on the user if it lands on Beshaba on an attack with disadvantage is a very good way to dissuade this "abuse". On the other hand, risking an advantage for an extra 1d10 damage (radiant) on the enemy would be kind a cool if you ask me. If you consider this, I would balance the usage at 3 per long rest.