"Antistatic" bags are dissipative, not conductive.
This suggests the very reason for the materials, and why some components are used with them.
Diodes, BJTs, SCRs and so on, have relatively high leakage currents. Thus, they dissipate any static charge by themselves.
A very small diode might not need static-dissipative packaging, but might still be susceptible to quite low ESD (under 1kV).
ESD is an extremely destructive event. An 8kV (IEC 61000-4-2) discharge releases a peak of almost 30A, ramping up in under five nanoseconds! The discharge lingers on for over 50ns, delivering 5 millijoules of energy. That's a peak power of ~100kW!
A very small semiconductor junction requires very little energy to completely vaporize. Even a (1um)^2 diode is unimaginably small. (Go on, try imagining it. Hair? Nope, human hairs are like gain silos! Pollen grains? Closer, but those are still like basketballs, compared to an about softball sized diode junction.)
So why static-dissipative materials? Some semiconductors are such good insulators, that they can accumulate static charge just from ambient fields. It doesn't leak away. MOS gates are this way. (But hey, small PN diodes are nearly as good!)
As for conductive materials (like antistatic foam, from the days of DIP ICs), it is conductive enough to short out any accumulated charge, plus short out ambient electric fields (thus avoiding electric induction as well). But it's not so conductive that the surge current would be destructively high: if the foam is exposed to a spark, the relatively high resistance limits current flow (effectively exposing any components, plugged in nearby, to a greatly attenuated ESD pulse, though still with a high voltage capability behind it), and the planar or voluminous nature of the foam causes the voltage to spread out and dissipate quickly with distance (as 1/r to 1/r^2).
You could use something like metal foil, to short out ICs and whatnot -- but the peak voltage isn't eliminated (the foil has equivalent inductance!), and the available current is still huge.
What about components that are *really* sensitive? Foam isn't good enough; they're stored with a metal spring wrapped around the leads (e.g., the classic 3N106 or whatever it was), or inside a cocoon of heavy foam.
The vast majority of components sold today are rated for 2kV or more, which is enough to cover most nuisance cases while handling, and suffices to protect against the spike that's leftover (usually 10s of V) from using external ESD protection on the outside connections where it's necessary.
Tim