deploy: critical safety fixes (negation leakage, age-gate bypass, trust boundary)

deploy: Egyptian-Arabic safety hardening + interview qualifier fix

Fabricated-negative guard, stateless interview API, handoff badge design

Adaptive ODIPARA guided interview, dental-specific question bank, age-extraction safety guard

Submission hardening: safety-rule coverage, honest agent dashboard, mobile/bilingual UX

Scrub legacy persona docstrings — the submission model is Qwen3-4B in-Space on ZeroGPU

Single-model ZeroGPU: local Qwen3-4B inference, named-patient demographics, Egyptian Arabic red-flag triggers

Guided history-taking interview: three Gemma 4 26B personas, one model, deterministic safety sentinel