v0.79.0

Changed

• A scanner that can only audit a whole cluster is skipped for a component that claimed part of
  one, instead of the descriptor being rejected. Declaring a cluster twice — once narrowed to the
  namespaces a component owns, once whole — is a reasonable thing to write, and 0.78.0 refused it,
  asking you to turn the scanner off by hand for the narrowed component. Draugr already knows which
  scanners cannot be narrowed, so it works it out: kubeBench and kubeBenchJob run against the
  component that claims the whole cluster, draugrK8sPolicies runs against both, and nothing needs
  disabling. draugr validate accepts the descriptor again.

Added

• A Not measured section in the report, naming any scanner that was planned and then not run,
  the component it could not answer for, and why:

  Not measured:
    infrastructure  kube-bench-job on team-a — audits the whole cluster and cannot be narrowed to namespace team-a
  

  A scanner that quietly does not run reads exactly like one that ran and found nothing, and the
  difference decides what a PASS is worth. Nothing is printed when every scanner could answer.