A “successful” deployment is not sufficient health evidence for one-shot or cron runners

A “successful” deployment is not sufficient health evidence for one-shot or cron runners. Production readiness must separately verify the scheduler is configured, a recent invocation ran, the downstream HTTP result was successful, and the process exits non-zero when its job fails; otherwise a platform can show SUCCESS for a runner that performed no successful work. Pair this with effective-contract validation for any environment-gated feature: prove target identity and deployed revision, then validate exact enabling literals and cross-variable invariants. This applies to scheduled runtime-gated features; it does not by itself prove vendor-account configuration or end-to-end delivery.