HongtaoMa1/grand-to-ground

Adversarial, Lakatos-inspired research loop for AI-assisted social science: converge grand frameworks into identifiable, falsifiable questions — now with a severity dial, declared-assumption protection, parallel deduction, and tick-to-approve research plans.