Let $\mathfrak{g}$ be a Lie algebra, we say that a subvector space $\mathfrak{h}\subset \mathfrak{g}$ is a ideal of $\mathfrak{g}$ if for any $v\in \mathfrak{h}$ and $w\in \mathfrak{g}$ we have $[v,w]\in \mathfrak{h}$.
I'm having troubles to prove that the only ideals of $\mathfrak{sl}_2(\mathbb{C})=\{M\in \mathfrak{gl}_2(\mathbb{C}); \text{Trace}(M)=0\}$ are $\{0\}$ and $\mathfrak{sl}_2(\mathbb{C})$ itself.
I picked $A\in \mathfrak{gl}_2(\mathbb{C})$, an ideal $\{0\}\neq J\subset \mathfrak{sl}_2(\mathbb{C})$ and $M\in J\setminus\{0\}$, then I was trying to find some smart choice of $N\in \mathfrak{sl}_2(\mathbb{C})$ such that $[M,N]=A$, where the Lie brackets here is $[M,N]=MN-NM$. But I'm not finding this $N$.
I think maybe that the "subvector space" hypothesis has to be used, maybe I can suppose that $J$ has dimension $1$ or $2$ and try to get an contradiction, but I'm not sure how to do this.