All considerations of FOS depend on a structure that starts out intact, and take into account design loads up to a defined maximum level of catastrophic stress (typical example being maximum wind loads up to some defined wind speed, or a maximum seismic activity) - and not beyond.
All such considerations are off the table if you exceed these maximum design forces, OR if you cut out a major portion of the structure in one go - as the plane impacts did.
Similar thing with fire resistance: Fire ratings are typically predicated on reasonable and normal assumptions of maximum fire scenarios, and may well not assume the stripping of fireproofing over major areas, the total failure of sprinklers, or the concurrent starting of intense fires on several floors and over large areas.
So from a design perspective, a plane crash with full fuel load is clearly and very far outside of the parameters that you could possibly design for, and it is an open question if such events should even be factored in. So I don't see the designers of the twin towers very much at fault or liable just because they did not design for every possible way to exceed design limits (note the implied paradox). To every building strength there is a force, however rare, that will overwhelm it. No building can ever be made indestructible. The ancient egyptian pyramids being obvious exceptions - with equally obvious problems regarding their appeal and usability for office work or living.
What if the buildings were secretly CDed with tens of thousand of charges - would anyone ask for future highrises to be built such that they can't be taken down with tens of thousand of charges? Nonsense.
The interesting building is WTC7, because it apparently succumbed to fires only, with the loss of sprinklers being the only thing that came on top of "usual", designed for, office fires. If I understand correctly, NIST identified a particular vulnerability that wasn't taken into account for in the applicable building codes, namely the effect of expansion (and contraction) of long beams under fire on truss connections. And then there are a few peculiarities that SanderO pointed out that seem to be too close to being single point of failure vulnerabilities.
What ozeco wrote: Whether or not NIST identified correctly the actual drop of water that proverbially made the barrel run over, or not, doesn't affect very much those who in the future design highrises: They will be very much aware of the problems with long trusses,open office spaces and naked steel protected by light-weight fireproofing only, and take them into account.
