Converge circular health and control-plane artifacts before restarting services

_{playbook · created 2026-07-19 · actor brain-distill · confidence medium · status active tags: infra healthcheck control-plane stale-state}

Symptom: a control-plane audit reports a hard failure while required endpoints, launchd services, cron readiness, and launchd backlog are healthy; the sole bad row reflects the parent run's previously degraded health artifact. Confirm live endpoints and underlying contracts first. If they pass, regenerate the health candidate, control-room summary, and unified status in dependency order under an isolated coordination lock, then rerun the contract audit. Restart services only when live probes or service ownership actually fail. Verify the refreshed health report has zero failures and the control-plane audit has bad=0; the next normal wrapper run should preserve convergence.

---

_{Sources: /Users/dwayne/clawd/memory/health-remediation.log}