To find the matrix of the transformation, you need to find the images of the three basis vector $e_1,e_2,e_3$.
Recall that a vector normal to the plane $ax+by+cz+d=0$ is given by $v_1=(a,b,c)$. Of course, $a,b,c$ are not all zero, so we suppose without loss of generality that $a\not=0$.
Note that $v_2=(-b,a,0)$ is a non-zero vector normal to $v_1$ (compute the dot product), so it is "parallel to the plane". Moreover, $v_3=v_1\wedge v_2$ is a third non-zero vector normal to both $v_1$ and $v_2$.
Therefore, you get a basis $(v_1,v_2,v_3)$ of $\mathbb{R}^3$ constituted of
One vector normal to the plane
Two vectors parallel to the plane
Therefore, in this basis, the matrix of the transformation is very simple: $v_1$ is sent to $-v_1$ and $v_2,v_3$ are fixed!
Now, two possbilities:
You know about change of basis matrices: then you compute the change of basis matrix from $(e_1,e_2,e_3)$ to $(v_1,v_2,v_3)$ and you will get the matrix in the canonical basis after a short calculation.
You don't know about that: you give $(e_1,e_2,e_3)$ as linear combinations of $(v_1,v_2,v_3)$, apply the transformation (now it is very simple) and then go back as linear combinations of $(e_1,e_2,e_3)$. Actually that's what we do in the previous item.
To find the matrix of the transformation, you need to find the images of the three basis vector $e_1,e_2,e_3$.
Recall that a vector normal to the plane $ax+by+cz+d=0$ is given by $v_1=(a,b,c)$. Of course, $a,b,c$ are not all zero, so we suppose without loss of generality that $a\not=0$.
Note that $v_2=(-b,a,0)$ is a non-zero vector normal to $v_1$ (compute the dot product), so it is "parallel to the plane". Moreover, $v_3=v_1\wedge v_2$ is a third non-zero vector normal to both $v_1$ and $v_2$.
Therefore, you get a basis $(v_1,v_2,v_3)$ of $\mathbb{R}^3$ constituted of
Therefore, in this basis, the matrix of the transformation is very simple: $v_1$ is sent to $-v_1$ and $v_2,v_3$ are fixed!
Now, two possbilities: