oc-fleet-ops

原文涵盖的主题

• Fleet Ops
• /oc-fleet — Command Reference
• How This Skill Fits the Build Pipeline
• Operating model — gated ops phases
• Phase 0 — Topology (/oc-fleet topology)
• Phase 1 — Provision (/oc-fleet provision) — IaC, Terraform when it fits
• Phase 2 — Deploy (/oc-fleet deploy)
• Phase 3 — Verify (/oc-fleet verify)

On first invocation, read references/orchestrator.md and follow its welcome protocol. Deploy and operate one-or-more containers across arbitrary, self-managed environments — the territory oc-deploy-ops routes elsewhere today. Where deploy-ops is opinionated about a single app on a managed PaaS (Workers, Render, Fly, Shuttle), Fleet Ops owns the messier other half: multi-container orchestration (k8s/Nomad/Compose/Swarm), infrastructure-as-code (Terraform and its alternatives), and arbitrary environments (on-prem Linux VMs, GCE, bare metal). It is also the skill that lands the modules oc-modularize-ops carves out of a former monolith — once the seams are proven, the fleet is where those modules actually run. This is not a tri-agent forge and not a managed-PaaS deployer. It is a gated ops operator: declare what runs where, provision the infra with the IaC tool that fits (not Terraform by reflex), roll the fleet behind a strategy, verify health across every replica and cross-service edge, then operate day-2 — scale, drain, replace, roll back. The IaC apply step is the highest-blast-radius operation in the entire opchain catalog, so it is gated behind a mandatory dry-run/plan that never

这里只展示节选，请以完整原文和当前要求为准。