[Gary Boldwater]

If the light rays entering a telescope are indeed parallel for images of distant objects, then that implies the light path is simply a cylinder equal to the width of the telescope aperture (neglect sidelobes). Is it really realistic to use the ensemble gravitational field of a cluster "causing" the bending of the lights path? Wouldn't it be more accurate to use the highest gravitational potentials the path crosses? Doesn't the gravitational field in a bend causing cluster vary widely across the cluster? I could shine a light through our solar system and come up with many different values for the bending depending where the light path was - near Jupiter, near the sun, near Pluto, etc. There is no uniform gravitational potential in our solar system and I don't think it is uniform across a cluster that bends light.
So couldn't the dark matter be really making up for the assumption that a bending cluster acts as a point gravitational source rather than a non uniform distributed one, which it really is?

[Physicist]

So couldn't the dark matter be really making up for the assumption that a bending cluster acts as a point gravitational source rather than a non uniform distributed one, which it really is?

But they explicitly don't make any such assumption. The path that light takes through the universe can be quite complicated. But fortunately, there are enough distant sources that scientists can deconvolve the distortions to make detailed maps of the distribution of dark matter. A professor here at Penn, Gary Bernstein, is one of the world experts on doing that.

The following plot is from one of his papers.