does pants have any means of informing subsequent runs of dependenckies discovered after running a build? (like analyzing some outputs post build)

Do you mean dependency inference within your repository of code, or the transitive graph of 3rd party depdencies?