The key point is knowledge of the limits, the reasons they were set and the consequences of exceeding the limit. A really good captain will know the above and hopefully make a reasoned decision. Just because a limit is exceeded doesn't necessarily make him a bad pilot.
On a helicopter, maximum IAS limits are set to avoid excessive vibration and a reduction of fatigue life and not for reasons of "aerodynamic flutter". In extremis, depending on type of aircraft, retreating blade stall may also be a limiting factor, subject to all up mass. It isn't really possible to directly compare helicopter limits with those of a glider or any other fixed wing though.
Transmission torque limits are set for similar reasons. In some cases, helicopters transmissions are able to carry quite a bit more torque than the published limit, but at the expense of possible excessive wear. It really depends on the weakest part in the transmission.
If the situation justified it, I would possibly risk the life of a machine to save human life, provided I was sure there was no danger to persons on board. One vital thing is that if a normal limit is exceeded, it must be documented so that corrective engineering maintenance can be carried out if required.
A few years ago I overtorqued a fully laden S-76 because on approach to a newly refurbished and supposedly FOD-plodded and cleared helipad, a hidden piece of wood suddenly flew up in the downwash towards the tail rotor. The crewman screamed that it was going to hit and full power didn't give sufficient avoiding action so adrenalin made me keep pulling until he stopped screaming. The overtorque (fairly large but short lived) was properly written up and no damage was found by the engineers. I was not criticised but then we (7 of us) would possibly have been dead had I NOT exceeded the normal transmision limits.