Helicopters use speed of the wing as a proxy for angle of attack in a similar manner to airplanes; in the airplane's case it's forward airspeed and in the helicopter's case it's "rotary airspeed", expressed as rotor rpm. In the airplane you have the bottom of the green arc on the airspeed indicator; in the helicopter it's the bottom of the green arc on the rotor tach. In the airplane there's a stall warning horn; in the helicopter there's a low rpm warning horn (that's the only "safety system" to prevent blade stall).
If you're in a hover, there's nothing stopping you from yanking the collective pitch all the way to the up stop. This shouldn't be sufficient to induce blade stall because the machine will respond by both climbing and increasing the vertical downwash of air, both of which will make the actual blade AOA increase less than the physical pitch change made by pulling collective, as long as rotor rpm stays in the green arc. Whether this action keeps the rpm in the green is another matter, and on turbine machines it's likely to overtorque/overtemp the engine(s) even if rpm is kept in the green.
If you pull pitch like that and the engine can't deal with the load even at maximum power, the rpm will quickly decay and once it goes below the green arc, blade stall could (will) occur at some margin below the green arc. When blade stall occurs from low rotor rpm while in a hover or at very low speed, the rpm decays even faster, the blade drag is too high for the engine to fight against, and the machine starts to settle vertically, causing blade AOA to increase further.
Before you know it the blades are WAY above stalling AOA, and you become While-E-Coyote strapped to an anvil and down you go; there's no recovery once a vertical descent with a stalled rotor starts, engine or not (if this happens in an engine out autorotation at altitude, the rotor will be close to stationary by the time you hit the ground - I recall a Robinson R22 incident from the 90s where the pilot lost engine power due to suspected carb ice, while flying along at 1000 ft, and didn't lower collective in time to get a proper autorotation started, the rotor rpm quickly decayed, the rotor stalled, and the machine just fell out of the sky like a wounded goose).
So the limitations are not so much about collective movement, it's the effect of collective movement on rotor rpm (and engine limits) that is critical, and the pilot has to monitor rotor rpm and respect the bottom of the green arc and the rotor rpm warning horn to stay safe, knowing that if the limits are exceeded and a stall happens, there is probably no recovery.
So, if you are hovering 10 feet in the air and you let it happen, it's a hard landing with maybe the tail boom getting chopped off. If it happens while hovering at, say, 100 ft, the result is fatal.
To answer each question specifically:
- Yes you can pull full collective any time. But the key safety parameter isn't collective position, but rotor rpm.
- The only safety system is the rotor tachometer and the low rpm warning horn.
- Yes, IF the engine has sufficient power to maintain rotor rpm in the green while pulling full collective. Some helicopters may be able to do this some may not. If it can't, the rotor stall won't occur until the rpm decays to some point below the green arc, and once that happens, you are done, even if the engine is still making power.
- Excessive collective input is revealed not by the actual lever movement, but by its effect on rotor rpm, which is the most critical parameter for the pilot to monitor. See here for a Robinson safety letter that talks about low rotor rpm rotor stall to see just how critical this parameter is.