If $c=0$ then the series obviously diverges (as it is then identical to harmonic series). I think it then follows by direct comparison test that the series also diverges for $c>0$.
I therefore guess that the series is divergent for any $c$ but can't figure out how to prove it precisely.
The series does not necessarily diverge for all $c$. Consider any $c=n-n^2$ for some natural $n$ and you will find that $a_k=0$ for all $k\ge n$.
For $c$ not of the above form, however, the series does in fact, diverge. Note that we have
\begin{align}a_k&=a_2\prod_{n=2}^k\frac{n^2-n+c}{n^2}\\&=\frac{a_2}k\prod_{n=2}^k\frac{n^2-n+c}{n^2-n}\\&=\frac{a_1}k\exp\sum_{n=2}^k\ln\left(1+\frac c{n^2-n}\right)\\&\sim\frac{a_1}k\exp\sum_{n=2}^k\frac c{n^2-n}\\&=\frac{a_1}k\exp\left(c-\frac ck\right)\end{align}
which diverges by the limit comparison test.