A risk-visualization navigator built around 4 typical AI application scenarios (Enterprise RAG, Agentic AI Support, AI Code Review, AI Fraud Detection). Each scenario stars a specific AI system and draws its applicable risk subset from the ARKO v8 knowledge base (7 top-level RS categories, 208 risk scenarios), laid out as a 5-stage propagation chain. Node content (descriptions, controls, framework mappings) is taken verbatim from the production knowledge base; severity and propagation chains are editorial initial calls in this scenario's context — per the ARKO methodology, a risk's likelihood and severity only crystallize within a specific system context, so the generic knowledge-base layer presets no fixed values.

• Node outer ring: pipeline stage (red/orange/yellow/purple/teal; the 5 stage names differ per scenario)
• Node inner circle: in-scenario severity call (red = critical, orange = major, yellow = moderate, green = minor)
• Edge thickness: propagation strength (1.0–0.3)
• Edge direction: risk flow (arrowheads)

1 2 3 4 5 Highlight that stage · R Reset view · S Export SVG for slides · H Toggle this help · Esc Deselect

1. Open with Enterprise RAG to introduce the risk graph
2. Press 1–5 to walk the risk distribution stage by stage
3. Click the starring risks (S0039 knowledge-base poisoning, S0126 retrieval injection, S0119 indirect injection) for a deep dive
4. Switch tabs to show the other scenarios (Agentic Support, Code Review, Fraud Detection)