Electrostatics with Conductors

In electrostatic equilibrium the field inside the conducting material is zero, which forces excess charge to the surface, makes the whole conductor one equipotential, and sets the field just outside to $\sigma/\varepsilon_0$, perpendicular to the surface. A cavity is a separate region: a charge $+q$ inside a hollow fills the hollow with field and induces $-q$ on the cavity wall, with $+q$ left on the outer surface if the conductor was neutral. Both results come from the same move, a Gaussian surface walked through the metal where $E = 0$. On a shape without spherical symmetry the surface density is not uniform; it climbs where the curvature is sharp, so points and corners carry the strongest fields.

Three errors dominate. Extending the metal's zero field into a cavity that contains charge, which erases both the cavity field and the induced wall charge. Writing $\sigma = Q/A$ on a pointed or pear-shaped conductor, which then makes the sparking at the tip unexplainable. And treating a metal enclosure as a symmetric shield: it does keep outside fields out of an empty cavity, but a charge inside still announces itself through induced charge on the outer surface, and only grounding removes that.