概要
The Defense-in-Depth skill provides a rigorous architectural framework for Claude to prevent software failures caused by invalid data. By moving beyond single-point checks, this skill guides the AI to implement validation at four critical layers: entry points, business logic, environment guards, and debug instrumentation. This methodology ensures that even if one layer is bypassed via refactoring, mocks, or alternative code paths, the system remains resilient. It is particularly useful for hardening complex data flows and preventing destructive operations in development and testing environments.