Consider the following ODE system \begin{align*} \dot{x} &= x^5 + y^3 = f(x\,,y) \\ \dot{y} &= x^3 - y^5 = g(x\,,y) \end{align*} Clearly, $(x\,,y) = (0\,,0)$ is the only fixed point. Linearizing $$ J = \begin{bmatrix} 5x^4 & 3y^2 \\ 3x^2 & -5y^4 \end{bmatrix} $$ evaluating at the fixed point will just give a zero matrix. Thus, need to use some other approach to analyze the stability, e.g., Lyapunov function.
However, I can't find any Lyapunov function related to this problem. Since this ODE system is at least $C^3$ in both $x$ and $y$. So I am wondering, for the linearization, is it possible to go all the way to 3rd-order derivative? Maybe something like $$ \begin{bmatrix} \frac{\partial^3f}{\partial x^3} & \frac{\partial^3f}{\partial y^3} \\ \frac{\partial^3g}{\partial x^3} & \frac{\partial^3g}{\partial y^3} \end{bmatrix} $$ This way the matrix will not be a zero matrix and can find corresponding eigenvalues and eigenvectors.

From
\begin{align*} \dot{x} &= x^5 + y^3 \\ \dot{y} &= x^3 - y^5 \end{align*}
we have
\begin{align*} x^3\dot{x} &= x^8 + x^3 y^3 \\ y^3\dot{y} &= x^3y^3 - y^8 \end{align*}
and then
$$ \frac 14\frac{d}{dt}(x^4-y^4) = x^8+y^8 $$
analyzing the behavior along a line $y(t) = \alpha x(t)$ we conclude
$$ \frac 14(1-\alpha^4)\frac{d}{dt}x^4 = (1+\alpha^8)x^8 $$
so depending on $\alpha$ the dynamics along $y(t) = \alpha x(t)$ changes concluding that the origin is unstable.