Knowledge Base admin: graph and articles

## Overview The Knowledge Base admin panel is at **Admin → Knowledge Base**. It has two surfaces: 1. **Knowledge graph** — the structured issue-symptom-cause-safe-check network that powers AI triage 2. **Article editorial workflow** — drafts, review, approval, and publication ## Knowledge graph The graph maps maintenance issues to: - **Symptoms** — observable signs reported by tenants - **Causes** — underlying reasons for the issue - **Safe checks** — steps tenants or property managers can safely carry out themselves - **Escalation triggers** — conditions that require immediate action - **Tradesperson types** — which trade should be dispatched - **Audiences** — who the article is intended for - **Intents** — the purpose of a given article (explain, guide, assess urgency, etc.) ## Article lifecycle ``` draft → in_review → needs_changes → approved → published → archived ``` Topic ideas are generated from operational signals (recurring ticket types, compliance gaps) or added manually. AI generation runs propose article drafts from the knowledge graph. A human editor reviews each draft, requests changes if needed, or approves it for publication. ## Featured articles Published articles can be marked **featured** to surface them prominently on the Knowledge Base landing page. ## Feedback loop End users can flag articles with feedback types including `inaccurate`, `unclear_audience`, and `safety_concern`. Flagged articles are routed back to the editorial queue for review. ## Related - Help: Knowledge Base admin — graph and articles - Tour: `kb-admin` - KB: [connected-learning-system](kb://connected-learning-system)