Totebox orchestration as the development environment
fix(systems): retire fabricated mediakit-os.md, correct wikilinks to os-mediakit
@@ -88,7 +88,7 @@ The workspace is intended to run on two separate compute nodes: **[[os-orchestration|os-orchestration]] node** — the development environment. Hosts the Command Session, all [[totebox-archive|Totebox Archives]], the [[service-slm|`service-slm`]] access-control gateway, source code, and workspace tooling. Intelligence work, code development, and staging are intended to happen here. Not accessible from the public internet. **[[mediakit-os|os-mediakit]] node** — the delivery layer. Hosts the live public websites, nginx, and production service instances. Receives deployments from the [[os-orchestration|os-orchestration]] node via a deliberate human-gated rsync bridge. No source code, no development work, no direct connection to the os-orchestration node. **[[os-mediakit]] node** — the delivery layer. Hosts the live public websites, nginx, and production service instances. Receives deployments from the [[os-orchestration|os-orchestration]] node via a deliberate human-gated rsync bridge. No source code, no development work, no direct connection to the os-orchestration node. The separation is designed so that a development incident — a failing test, a disk-full event, an unfinished feature — cannot affect the live public websites. The rsync bridge is intended as the deliberate human gate between development and production, consistent with the [[architecture-decisions|SYS-ADR-19]] discipline of no automated publishing to live environments.