Not a mistake, per se, but adding an unnecessary assumption towards a contradiction.
Consider the following argument that any number divisible by $4$ is divisible by $2$:
Towards a contradiction, assume that $n \in \N$ is divisible by $4$ but not divisible by $2$.
Since $n$ is divisible by $4$, we may write $n = 4k$ for some $k \in \N$.
But then $n = 2(2k)$ and $2k \in \N$, so $n$ is divisible by $2$.
This contradicts our original assumption, and so the claim is proved.
It is cleaner to argue directly:
Suppose $n$ is divisible by $4$.
We may therefore write $n = 4k$ for some $k \in \N$.
But then $n = 2(2k)$ and $2k \in \N$, so $n$ is divisible by $2$, and the claim is proved.
There is a similar temptation when arguing the contrapositive.
Consider the following argument that if $x^2$ is even, then $x$ is even:
Towards a contradiction, suppose that $x^2$ is even while $x$ is odd.
The product of two odd numbers is odd, so $x\cdot x = x^2$ is odd.
This contradicts our assumption that $x^2$ is even, and the claim is established.
Again, it's cleaner to just argue the contrapositive itself:
To establish the claim, we will argue its contrapositive.
Suppose therefore that $x$ is odd.
The product of two odd numbers is odd, so $x\cdot x = x^2$ is odd, and the claim holds.