cargo kani assess: experimental test case runner

[tedinski]

#1497 will give a basic report on features Kani does not support in a crate and its dependencies.

We should be able to try running a crate's existing test suite to measure how many of those features are actually hit.

Further, we should be able to track "failure reasons" for other "test case fails to run under Kani" situations. For instance, if it clearly fails because of the stdlib linking issue.