you can model this scenario with a binomial probability since you have a fixed number of trials (n = 2*12 = 24), a set probability (p = 0.03), and then an exact measure you are looking for (x <= 2). i would technology like Excel or a binomial distribution calculator in order to solve ... when i do that i get
P(x <= 2) = 0.9659
meaning that there is about a 96.59% chance that no more than 2 of your eggs are cracked