We cannot protect against large once-off bodies like long period comets. There is simply not enough time to detect and track them in order to intercept early enough for a diversion.

Imagine if Varuna were on a collision course. How does one steer a 900km body with puny human technology?

Until you admit the danger can never be totally eliminated, you aren't being scientific.

Instead, you need to work on risk reduction by a variety of methods, adopting those strategies that give best bang-per-buck (or is that anti-bang?).

Somewhere in that spectrum fits dispersion of habitats to the moon, Mars, LEO and the asteroid belt. I don't know where exactly. I know it comes way behind a better Earth-based sky survey, but it doesn't neccesarily come ahead of developing massive space tugs capable of intercepting high speed arrivals of huge individual masses - the "Chicxulub's" of the distribution.