defense-in-depth

SKILL.md

Defense in depth applies multiple security layers so that if one fails, others still protect the system. For AI applications, this means validating at every boundary: edge, gateway, input, authorization, data, LLM, output, and observability.

Core Principle: No single security control should be the only thing protecting sensitive operations.

| Defense in Depth | NIST | Multiple validation layers | | Zero Trust | Google BeyondCorp | Every request verified | | Least Privilege | AWS IAM | Minimal permissions | | Complete Mediation | Saltzer & Schroeder | Every access checked |

Verwenden Sie es beim Aufbau sicherer KI-Pipelines oder beim Härten von LLM-Integrationen. Defense-in-Depth implementiert 8 Validierungsebenen vom Edge bis zum Speicher ohne Single Point of Failure. Quelle: yonatangross/orchestkit.

Original anzeigen