Firstly, it is clear that $f$ has not constant coefficient. I claim that for any $f\in \mathfrak{p}$, $f$ has not terms $X^1, X^2, Y^1, Z^1$. Since, for example,
$$f(X,Y,Z)=Y^1+\textrm{(other terms)}=Y^1+Y^2(\ldots)+X(\ldots)+Z(\ldots)$$
then
$$0=f(T^3,T^4,T^5)=T^4+\underbrace{T^8(\ldots)+T^3(\ldots)+T^4(\ldots)}_{\textrm{has not term $T^4$}}$$
since $3a+4b+5c\neq 1$ for $a,b,c\in \mathbb{Z}_{\geq 0}$. But, $X^3-YZ, Y^2-XZ, Z^2-X^2Y\in \mathfrak{p}$ have terms $X^3, Y^2,Z^2$ respectively. If $\left<f,g\right>=\mathfrak{p}$, assume $a,b,p,q,n,m\in k[X,Y,Z]$ such that
$$af+bg=X^3-YZ\qquad pf+qg=Y^2-XZ\qquad nf+mg=Z^2-X^2Y$$
Assume that
$$\left(\begin{matrix}f\\ g\end{matrix}\right)
=\left(\begin{matrix}x_f& y_f & z_f\\\ x_g& y_g & z_g\end{matrix}\right)
\left(\begin{matrix}X^3\\ Y^2\\ Z^2\end{matrix}\right)+ \left(\begin{matrix}\textrm{other terms}\\ \textrm{other terms}\end{matrix}\right)$$
and that $a_0=a(0,0,0)$ and so on. Then by looking at the term $X^3,Y^2,Z^2$, one has
$$\left(\begin{matrix}a_0 & b_0\\ p_0& q_0\\ n_0&m_0\end{matrix}\right)
\left(\begin{matrix}x_f& y_f & z_f\\\ x_g& y_g & z_g\end{matrix}\right)
=\left(\begin{matrix}1 & 0 &0 \\ 0& 1 & 0\\ 0&0 &1\end{matrix}\right)$$
which is a contradiction by standard linear algebra. $\square$