promptfoo-framework-dod-ai-ethicslisted

# DoD AI Ethical Principles The U.S. Department of Defense adopted five AI ethical principles in 2020: **Responsible, Equitable, Traceable, Reliable, Governable**. Promptfoo maps each principle to concrete red team plugins via `dod:ai:ethics`. ## Use this skill when - Testing DoD AI ethical principles compliance - Testing responsible AI (human accountability, excessive agency) - Testing equitable AI (bias across age, disability, gender, race) - Testing traceable AI (hallucination, source attribution, unverifiable claims) - Testing reliable AI (injection, SSRF, unsafe practices) - Testing governable AI (control boundaries, hijacking, RBAC, tool discovery) ## Do not use this skill when - Testing NIST AI RMF → use `promptfoo-framework-nist-ai-rmf` - Testing OWASP LLM Top 10 → use `promptfoo-framework-owasp-llm` - Testing ISO 42001 → use `promptfoo-framework-iso-42001` - Testing general LLM app security → use `promptfoo-redteam-llm` ## 🛡️ Edge cases (mandatory handling) - **DoD principles are ethical, not just security** — bias and fairness testing (Equitable) requires demographic-specific plugins. - **Traceable includes hallucination** — not just logging; test for fabricated claims and unverifiable statements. - **Governable requires control boundary testing** — `indirect-prompt-injection`, `system-prompt-override`, `rbac`, `bfla`, `bola`, `tool-discovery`. - **Combine with security frameworks** — DoD ethics testing pairs with NIST AI RMF, OWASP Agentic, ISO 42001. ## 🎯 C