Operations
Operational playbooks for recurring Open-FDD verification, OpenClaw context durability, and operator-style review loops.
| Page | Description |
|---|---|
| Open-FDD integrity sweep | 20-minute operator-style health and model integrity checks. |
| Overnight review | 6 PM-6 AM deeper review pattern for logs, behavior drift, and durable context updates. |
| OpenClaw context bootstrap | What durable context should be stored in-repo vs kept local/private. |
| Operator framework | Canonical operator reasoning contract and source-of-truth order. |
| MCP RAG service | Derived AI retrieval service from canonical docs + generated docs text. |
| Mode-aware runbooks | Collector/model/engine/full orchestration flows for humans and OpenClaw. |
| Testing plan | Engineering plan for bench testing, auth/SPARQL checks, and overnight evidence. |
| Continuous context backup loop | What to commit to the repo vs keep local (no secrets). |
| Usage budget and throttling | Throttle ladder when model/API budget is tight. |