(A) When using Gauss' law for spherically symmetric systems, $r$ is evaluated at the radius of the Gaussian surface. I understand where the confusion comes from, because in the "standard" equation for the electric field, the distance $r$ is defined as the distance between the point where you're measuring the field and the point where the charge is located, and in the case of a spherical surface, you might think that measuring the field on the surface that since you are infinitesimally close to the surface, $r$ should be zero. In a real system, if you were able to get that close to the surface, you would actually have to start worrying about where the quantized charges are located, so things would get complicated close to the surface. For these types of problems, it is assumed that the charges are "smoothed" out over the surface, and using Gauss' law comes in handy. Gauss' Law is used in the following way.
$\oint \vec{E}\cdot d\vec{a} = \frac{Q_{enc}}{\epsilon_0}$
Since you chose a spherical surface $\vec{E}$ and $d\vec{a}$ are parallel. Also, $\vec{E}$ is constant over the surface, so the integral is just the following:
$E\oint da = \frac{Q_{enc}}{\epsilon_0}$
where $\oint da= 4\pi r^2$, and $r$ is the radius of the Gaussian surface. So,
$\vec{E} = \frac{Q_{enc}}{4\pi r^2} \hat{r}$
Now, you can see that $r$ is the radius of the sphere, and the only place that the size of the charged sphere would matter is when determining whether the charge on the surface of the sphere is enclosed or not. Since on a conductor, all of the charge is on the outer surface, at the surface itself, the charge is generally not considered to be enclosed, but just above the surface (infinitesimally), the charge is considered to be enclosed (all of the charge on the sphere).
(B) In this case, you have to remember that conductors do not like to have electric fields on the inside. So when connecting the two sphere's by a conduction wire, any net charge will want to move to the outer surface. So you are right in this case, the charge on the inner sphere is zero and the net charge is on the outer sphere.
(C) The work done by an electric field on a particle is $q\Delta V$. The way to think of it is that the work done on something is equal to the change in potential energy of that object. In this case, the change in the potential energy of a charge is given by $\Delta U = q\Delta V$. Now, the sign of $q$ and $\Delta V$ should be determined by the system.
For instance lets consider free charges at rest in an electric field. A proton has $q=+e$ and an electron has $q=-e$. Protons move from high potential to low potential, and electrons move from low potential to high potential. So, with $\Delta V = V_f - V_i$, in both cases you see a negative work done by an electric field, because the charges will move toward a position that lowers their potential energy. In order to do (positive) work on a charge, you need to raise the potential energy. This is something that batteries do.
Generally, you'll want to calculate the electric field in each region of space (both inside and outside the shell) using Gauss' Law. When point charges are involved, since the potential and electric field tend to diverge at the location of the point charge, you'll want set the potential equal to zero at infinity, and integrate the electric field as you've suggested in part (C) from infinity to the point $r$ that you want to know the potential at. If you know the electric field in each region of space, then you can integrate all the way in from infinity.
As for part B, think about Gauss' Law. You should not expect the electric field to be zero inside the shell if there is a point charge in there.
Best Answer
A solution to this problem is provided on the okphysics website. It does not use integration but assumes that the potential at the surface of an isolated sphere of charge $Q$ and radius $R$ is $kQ/R$. It also uses the superposition theorem and the fact that the charge on each shell can be replaced with a point charge at the centre.
The potential at the surface of the outer shell is that at radius $b$ from charge $Q_a+Q_b$ : ie
$V_b = k\frac{Q_a+Q_b}{b}$.
In going from the outer to the inner shell we are inside charge $Q_b$ so it does not contribute to the potential difference. We move from radius $b$ to radius $a$ from charge $Q_a$. So the potential difference between the 2 shells is
$V_{ab}=kQ_a(\frac{1}{a}-\frac{1}{b})$.
The potential at the surface of the inner shell is
$V_a=V_b+V_{ab}=k(\frac{Q_a}{b}+\frac{Q_b}{b}+\frac{Q_a}{a}-\frac{Q_a}{b})=k(\frac{Q_a}{a}+\frac{Q_b}{b})$
Another way of deriving the last result is by superposition.
First consider the outer shell in isolation. Since there is no electric field inside this shell, the potential is the same at all points inside as it is on the surface : ie $k\frac{Q_b}{b}$. In particular this applies at all points at radius $a<b$ from the centre.
Second consider the inner shell in isolation. The potential at the surface is $k\frac{Q_a}{a}$.
Now superpose the 2 potentials to get the potential at the surface of the inner shell when both charged shells are present :
$V_a=k(\frac{Q_a}{a}+\frac{Q_b}{b})$.