Just for your curiosity
It is clear that Riemann sum make the problem quite simple.
But we can go beyond the limit and approximate the term which is
$$a_n=\left(n^{-\frac{1}{2} n (n+1)} H(n)\right)^{\frac{1}{n^2}}\implies \log(a_n)=\frac 1{n^2}\log\left(n^{-\frac{1}{2} n (n+1)} H(n)\right)$$ where $H(n)$ is the hyperfactorial function.
Expanding the logarithm and using the asymptotics of $\log(H(n))$ we quickly end with
$$\log(a_n)=\frac{1}{ n^2}\left(-\frac{n^2}{4}+\log (A)+\frac{1}{12} \log
\left({n}\right)+\frac{1}{720 n^2}+O\left(\frac{1}{n^4}\right)\right)$$ where $A$ is Glaisher constant that is to say
$$\log(a_n)=-\frac{1}{4}+\frac{12\log (A)+ \log
\left({n}\right)}{n^2}+O\left(\frac{1}{n^4}\right)$$ which shows the limit and how it ia approached.
Moreover, this gives you a shortcut method for computing $a_n$. If you are patient, compute $a_{10}$; the exact value is $0.78224014$ while the above (very) truncated series would give $0.78224004$.