Yes, your attempt gives a maximal value of $k$!
But we need $a<b$, which gives:
$$\min\limits_{0<a<\sqrt3}\frac{a^2-3a+3}{3(3-a^2)}=\frac{1}{6}$$ and it's enough to prove that:
$$6(a+b+c-3)\geq(a-b)(b-c)(c-a).$$
Indeed, let $a+b+c=3u$, $ab+ac+bc=3v^2,$ where $v>0$ and $abc=w^3$.
Thus, $v=1$ and since $$a+b+c\geq\sqrt{3(ab+ac+bc)}=3,$$ it's enough to prove that:
$$36(a+b+c-3)^2\geq\prod_{cyc}(a-b)^2$$ or
$$36v^4(3u-3v)^2\geq27(3u^2v^4-4v^6-4u^3w^3+6uv^2w^3-w^6)$$ or
$$w^6+2u(2u^2-3v^2)+(3u^2-4v^2)^2v^4\geq0,$$ which is obvious for $2u^2-3v^2\geq0.$
But for $2u^2-3v^2\leq0$ it's enough to prove that:
$$u^2(2u^2-3v^2)^2-(3u^2-4v^2)^2v^4\leq0$$ or
$$4(u^3-3uv^2+2v^3)(u^3-2v^3)\leq0,$$ which is true because
$$u^3-2v^3\leq\left(\left(\sqrt{\frac{3}{2}}\right)^3-2\right)v^3<0$$ and by AM-GM
$$u^3-3uv^2+2v^3\geq 3\sqrt[3]{u^3\cdot(v^3)^2}-3uv^2=0.$$