Let $R$ be some (possibly non-commutative) ring, and let $X$ be an $R$-module with dual $X^\ast=\mathrm{Hom}_R(X,R)$. I've shown both $X,X^\ast$ exist in an exact sequence of the form
$$0\to B\to X \to A\to 0 \text{ and } 0\to B\to X^\ast\to A\to 0.$$
For ease, I'll denote this by $X,X^\ast\in\mathrm{Ext}^1(A,B)$ even though I actually mean their respective classes. In addition, I have shown $\mathrm{Ext}^1(A,B)\cong\mathbb{Z}/3$ so, up to isomorphism, we have three classes. Now, if $X$ is indecomposable then does it follow that $X\cong X^\ast$? Here is my argument:
If $X,X^\ast$ are in the same class, then clearly they are isomorphic by the Five Lemma, so suppose not. As neither is indecomposable then one belongs to the class $c_1$ (say) and the other to $c_2$. In particular, $c_1=-c_2$ so I have
$$0\to B\to X \stackrel{f}{\to} A\to 0 \text{ and } 0\to B\to X^\ast\stackrel{-f}{\to} A\to 0.$$
Let $g:X\to X^\ast$ be the map sending an element to its dual, then surely this gives us a commutative diagram from which we can once more use the Five Lemma? Or am I missing something?