fix(cache): no-cache on frontend assets + cache-bust (Safari stale JS)

feat: add About panel (raw prompt, LLM version, RAG bundle version)

feat: add retrieval toggle (device parity) + reciprocal repo link

fix(deploy): point llama-server at the Gemma 4 GGUF (start scripts)

feat: mirror device update — Gemma 4 E4B + G1 prompt

docs: sync README/comments with current UI (no gate, gated feedback, citations)

feat(ui): click-through citations open source PDF at the cited page

feat(ui): gate feedback feature behind a flag, off in deployment

fix(deploy): copy prebuilt llama-server instead of compiling

fix(deploy): use HF-allowed Space card colors

feat(deploy): add Docker image and Hugging Face Spaces config

feat(ui): remove entry gate, open chat directly

feat(ui): remove demonstration warning box from disclaimer gate

docs: add cloud hosting review (DEPLOYMENT.md)

chore(scripts): add asset-fetch and local run scripts

feat(ui): chat interface with disclaimer gate and feedback form

feat(backend): chat SSE proxy, prompt assembly, feedback store

feat(retrieval): faithful EmbeddingGemma live retrieval over v0.3.0 store

chore: scaffold clinician-feedback demo project