Research Audit Protocols: Duel + MAD

This release expands the repository from the original ChatGPT--Gemini Duel protocol to a broader set of research-audit workflows. Included in v2.0: original Duel v1.7 protocol retained unchanged new 4-model MAD protocol added Markdown prompt file: protocol/ai_mad_protocol_v2.0.md PDF handout: protocol/ai_mad_protocol_v2.0.pdf README updated to present both workflows side by side Use Duel v1.7 for a lighter two-model adversarial workflow. Use MAD v2.0 for higher-stakes document audit using ChatGPT, Claude, Gemini, and Grok.