Commander review: direction is correct. This PR records DEV M4 agent-loop evidence as a bounded preflight, keeps M4 subordinate to the M3 trusted-loop mainline, and preserves blockers for DB live readiness and skill version injection. No PROD/service restart is included.
Commander review: direction is correct. This PR separates runner kubeconfig/read-only observability gaps from D601 public endpoint/k3s availability and keeps M3 blocked until real operation/trace/audit/evidence are observed. It does not touch PROD or restart services.
Refs HWLAB#37, HWLAB#64, HWLAB#78.
Reviewed by commander. This keeps M4 as blocked at DB live readiness, does not promote dry-run/source evidence to DEV-LIVE, and preserves the patch-panel boundary.
Merge PR #71 after commander review. Rebaselines DEV live evidence to 16666/16667, removes stale edge/frp blocker classification, keeps DB live and M3 patch-panel runtime as real blockers. Complies with HWLAB #78 / DC-DCSN-P0-2026-003 by preserving DEV-LIVE evidence boundaries and not treating Gate/UI as MVP success.
Refs #66. Merged by commander after reviewing Code Queue task codex_1779422778882_1. Separates process HTTP, Docker daemon push, and k3s pull registry capabilities so runner loopback HTTP degradation does not block artifact publish when Docker/k3s paths are proven.
Refs #49
Merged by commander after reviewing Code Queue task codex_1779422762669_1. Adds redacted DB readiness layering and refreshes DEV endpoint/report evidence. Follow-up: deploy/verify the new readiness probe in DEV via standard image CI/CD.
Code Queue Review