Skip classification catches a check that ran and gave up. It cannot catch a capability that had no check at all — which is #602 exactly: test_workbench_login skips under SSO and no other lane verifies a Workbench login, so nothing is even present to classify.
Derive the expected checks from what vip.toml declares. If Connect is configured with SSO, the run must name a check that actually exercised SSO login; if none executed, report the capability as unproven rather than staying silent. This is what turns the report into "31 of your 47 declared capabilities were verified" instead of a pass/fail count over whatever happened to run.
Depends on the unproven outcome from #617, which already supplies the reporting and exit-code machinery.
Skip classification catches a check that ran and gave up. It cannot catch a capability that had no check at all — which is #602 exactly:
test_workbench_loginskips under SSO and no other lane verifies a Workbench login, so nothing is even present to classify.Derive the expected checks from what
vip.tomldeclares. If Connect is configured with SSO, the run must name a check that actually exercised SSO login; if none executed, report the capability asunprovenrather than staying silent. This is what turns the report into "31 of your 47 declared capabilities were verified" instead of a pass/fail count over whatever happened to run.Depends on the
unprovenoutcome from #617, which already supplies the reporting and exit-code machinery.