In case the Apollo Lunar module had approached the landing site at way too high velocity, could the ascent stage have separated and returned to the command module before a crash? Thus performing a task similar to a launch abort tower. One challenge would be to synch with the orbit of the command module.
Yes. There is a lot of good information in this presentation from the June 1966 Apollo Lunar Landing Mission Symposium relative to landing flight design including abort planning. The crucial figure is this one:
which shows the capability of the ascent engine to abort all the way down to landing. It assumes a 4 second delay to separate the landing stage.
As far as returning to the orbit of the command module, the Apollo Experience Report - Abort Planning document explains on page 24 (33 of the pdf) that the Lunar Module would attempt to attain a safe orbit and the Command Module would rendezvous with it.
That maneuver was called "fire in the hole". See the wikipedia article. According to this NASA paper there was an extra "fire-in-the-hole shield". During Apollo 5, 9 and 10 there was a test of "fire-in-the-hole", see 1 , 2 and 3. The LM test of Apollo 5 was done unmanned.