I would do it like this
$\mathbf x^T \begin{bmatrix} a & \frac b2 \\ \frac b2 & c\end{bmatrix} \mathbf x = 0$
I say that there exist vectors such $Av = \lambda v$
If such vectors exist, then:
$Av = \lambda I v\\
(A-\lambda I)v = 0$
$(A-\lambda I)$ is a singular matrix.
$\det(A-\lambda I) = 0$
$\det\begin{bmatrix} a-\lambda & \frac b2 \\ \frac b2 & c-\lambda\end{bmatrix} = \lambda^2 - (a+c)\lambda + ac - \frac {b^2}{4} = 0$
lets call the roots of that polynomial $\lambda_1, \lambda_2$
$A v_1 = \lambda_1 v_1,A v_2 = \lambda_1 v_2\\
A[v_1,v_2] = [v_1,v_2] \begin{bmatrix} \lambda_1 &\\&\lambda_2\end{bmatrix}$
$AP = P D\\
A = PDP^{-1}$
I am not going to prove it here but if $A$ is symmetric,
1) it is diagonalizable
2) it has real eigenvalues
3) there is an ortho-normal set of eigenvectors