Partial derivative notation is ambiguous. But you've correctly identified the crucial missing piece of information that was intended in your calculation — all of the derivatives are meant to be computed subject to the restriction that $y$ is constant.
It's not $F_y$ that's zero, however; $F_y$ still equals $ze^{yz}$. It's $\frac{\partial y}{\partial x}$ that's zero — it's zero because the intent of the calculation is to take derivatives while holding $y$ constant.
Note that if the intent was to take derivatives while holding $z$ constant, then $\frac{\partial y}{\partial x}$ would not be zero.
One notation I've seen for adding in the missing information is
$$0=F_x \left.\frac{\partial x}{\partial x}\right|_{y \text{ const}} + F_y \left.\frac{\partial y}{\partial x}\right|_{y \text{ const}} + F_z \left.\frac{\partial z}{\partial x}\right|_{y \text{ const}} $$
or some abbreviation thereof.
For the sake of precision, I should point out what $F_y$ means too. If we write down the function $F(a,b,c) = a c^2 + e^{bc}$ (so that the given equation is the value of $F(x,y,z)$), $F_y$ is shorthand for $F_2(x, y, z)$, where $F_2$ means to take the derivative of $F_2$ in its second place.