I'm not too familiar with continuous probability distributions so I'm unsure if my approach is correct. My first thought was to use conditional probability. I thought that the probability of this event is equivalent to solving the following:
$$Pr[\text{first number is } x \in[1,3]] \cdot Pr[\text{other two numbers are }\leq x ~|~ \text{first number is } x \in[1,3]]$$
This would give me $$\int_1^{3}\frac{1}{5}\cdot\frac{x}{5}dx=0.16$$
Does this approach make sense?
I think it's easier to just analyze this directly. A necessary condition is that no numbers be in $(3, 5]$. The probability that any one number is outside $(3, 5]$ (we'll call this a "failure condition") is $\frac 35$, so the probability that all three numbers are outside that range is $\frac {27}{125}$. Thus, the probability that at least one number is in the range $(3, 5]$ is $\frac{98}{125}$.
Another failure condition is that all three numbers are in $[0, 1)$. The probability that all three numbers are in $[0, 1)$ is $\frac {1}{125}$.
These failure conditions are independent. If neither failure condition is met, then the largest of your three numbers will be in $[1, 3]$. The probability of satisfying one of the failure conditions is $\frac{98}{125}+ \frac {1}{125}= \frac {99}{125}$ (we can add probabilities because it's impossible to satisfy both failure conditions simultaneously), so the probability that neither failure condition is satisfied is $\frac{26}{125}=0.208$.