1. In the Aidome architecture, where does the Control Plane sit?
2. Which deployment model keeps all AI traffic inside the enterprise perimeter with no outbound calls?
3. What are the two identities carried by every AI interaction in Aidome?
4. Which is a valid model routing strategy in Aidome?
5. Why must the AI gateway intercept traffic for policy to be enforced?
6. What does a GenAI firewall inspect that a traditional WAF does not?
7. Indirect prompt injection is best described as:
8. Why is traditional DLP insufficient for AI?
9. Which action is a valid DLP response in Aidome?
10. AI discovery is the prerequisite for every other control because:
11. What does an AI-BOM (AI Bill of Materials) list?
12. Why must every MCP tool be scoped and audited?
13. Tool poisoning refers to:
14. An AI agent with access to tools and data should be treated as:
15. Human-in-the-loop checkpoints in agent security are used to:
16. What is the core RAG security principle?
17. RAG poisoning is an attack against:
18. In policy engineering, policies are best treated as:
19. Which is a valid policy action in Aidome?
20. What does dry-run mode allow administrators to do?
21. AI observability differs from traditional logging because it captures:
22. Why should AI security events land in the same SIEM as other security events?
23. Which frameworks are used to map AI security events in a SIEM?
24. What does Aidome's kill-switch allow an administrator to do?
25. AI incident response adds semantic containment, meaning you block:
26. Which is an AI incident category?
27. Sovereign AI means:
28. In an air-gapped deployment, how are policy and signature updates applied?
29. AI FinOps is security-adjacent because:
30. Token budget enforcement is an example of:
31. Which is a valid model routing optimization for cost?
32. What does the LLM endpoint switchboard normalize?
33. Just-in-time access in Aidome is used to:
34. Provider health monitoring and failover ensures:
35. TLS inspection at the gateway requires:
36. Hallucination and data exfiltration pattern detection is part of:
37. Contextual DLP blocks based on:
38. SaaS AI discovery in Aidome identifies tools such as:
39. AI-BOM component attestation is used to assess:
40. Agent session boundaries and isolation prevent:
41. Citation and provenance for retrieved chunks in RAG helps with:
42. Policy combinators in Aidome include:
43. Forensic replay of interactions relies on:
44. A SOAR playbook for AI security incidents typically automates:
45. Post-incident review for an AI security event should result in:
46. Which best describes the relationship between the data plane and control plane in Aidome?