GenAI Security (Merged)

Security is now integrated into the architecture pages:

• AI Infrastructure and Evaluation — identity, authorization, tenant isolation, prompt injection, data exfiltration, data protection, tools, networking, and observability.
• AI Agents — least-privilege tools, MCP trust, approval, idempotency, and autonomy limits.
• AI Knowledge Bases — document ACLs, metadata filters, tenant leakage, and authorization-aware retrieval.
• AWS AI Services — IAM, KMS, Secrets Manager, VPC endpoints, CloudTrail, WAF, Macie, and AWS guardrails.