In the "whoops" scenario (i.e. non-malicious), an engine misalignment would be apparent with enough time to correct. You're a year out.
We watch asteroids zing by regularly. The ones we know about have their chances of hitting calculated years or decades out.
Sideways motion is harder to determine than motion towards us; the former needs angular resolution, the latter doppler radar.
And the sort of rock that could shift 67P from a safe to a dangerous path (LEO to 45 degrees one year ahead) would be about 13 metres radius. The Wikipedia citation is 5 years old, so I don’t know what we know today, but in 2016 the estimate was we only knew somewhere around 1.3% of asteroids between 40m and 3.5m diameter (and 0.003% of smaller than that, because things get harder to spot as you get smaller).
Wiki: https://en.wikipedia.org/wiki/Near-Earth_object#Size_distrib...
ArXiv link because the citation in Wikipedia was giving me a server error: https://arxiv.org/pdf/1604.06328.pdf
I don’t think so? If you’re bringing in a big one you need more engines to push it into the right orbit to start with, so I think the risk is always things that scale up directly with the comet’s mass, e.g. “the engines are misaligned by $foo degrees”.
Also, a pebble in space terms is still a potential city-killer, and no nation in their right mind is going to want to let a private company — not one of their own, let alone a private company in another nation if they have even the slightest say in the matter — wield that kind of potential: https://en.wikipedia.org/wiki/Meteor_Crater