At a minimum you should set up the "no clip" compressor preset on every output to help prevent blowing HF drivers.
That may or may not work depending on available amp power vs. the rms rating of the driver(s). If you set a compressor as a clip limiter and get solid feedback, you might still risk of cooking drivers. Unless the amp is sized so that it doesn't exceed the rms rating of the driver running at full power without clipping, you can still blow the driver. Most folks size their amps so that they're 1.5x to 2x the rms rating of the drivers (so-called "program power rating"). Feedback is essentially pure sine wave. Full power sign wave without clipping means your hammering the driver with maximum heating power the amp has to offer. If your amp is larger than the rms rating of the speaker, a few seconds of solid feedback could let out the magic smoke. If you set your compressor so that it's attenuating the signal to a level below the rms rating of the speaker, sure, you can save the speaker. But you have to measure and calculate that stuff to set it up correctly.
Also, most compressors won't work as brick wall limiters to kill transients even at fastest attack like the kind you get hot-plugging an unmuted condenser mic (a common cause of blown HF drivers). They usually let the initial attack through
unless they have a "look-ahead" side chain setup. You only find those type of limiters on high end digital speaker processors (like XTA) or high end amps with built-in processing. As far as rms level protection goes, that also requires properly calibrated limiting to keep drivers safe. Again, the high end processors and amps will have that type of limiting along with the look-ahead peak limiters.
Bottom line, you can't trust regular compressors to protect speakers unless you set them up to the point where you're also waisting headroom/output for the speakers. and you can't rely on them to protect speakers from high power transients.
