fix: Sage timeout — gunicorn 300s, max_tokens 150, ZeroGPU pattern, correct torch_dtype

Switch to HF Space as default backend, make Ollama optional

Replace cloud inference with local Ollama

feat: add anti-gaming comprehension pipeline and harden HF space integration

Fix HF Space backend contract and production health diagnostics

Update app defaults and database to use US Curriculum Guide, fix HF requirements