Clean up task3 and task8 simulations — only imperfect where reasoning warrants it

Fix task3 simulation: replace generic lab lookup with clinically relevant HR/conditions check

Make simulations realistic with redundancy penalties visible in trace

Fix patient banner to show real FHIR names instead of hash-generated fakes

Update reward.py for 5 in-scope task types and rebuild fhir_cache

Remove redundant reward score section from sidebar

Deduplicate FHIR API entries in tab renderer

Load FHIR APIs tab on page open

Load FHIR APIs tab on page open

Fix baseline Qwen row: add lb-baseline-tasks id, fix chip label

Add task5 CT oncology follow-up

Add task5 CT oncology follow-up

Add task5 CT oncology follow-up

UI: sidebar scroll fix, FHIR APIs tab, aligned reward breakdown

UI + data updates

UI + data updates

Add task7 QTc entries to benchmark

Add task7 QTc interval review UI

Rename quality score to reward score with penalty range note

Remove system architecture card from dashboard

Dashboard: add SOTA leaderboard with per-task success rates

UI: collapse action panel to compact status bar

UI: white theme, PCP Clinic branding, auto-agent execution (correct path)

UI: white theme, PCP Clinic branding, auto-agent execution

Fix port: use 8000 to match app_port in README

Remove HEALTHCHECK so HF Spaces flips to Running immediately

Remove ENABLE_WEB_INTERFACE: stops env from being instantiated at startup

Slim Dockerfile: drop build-essential and git, only install curl

Fix Dockerfile for HF Space: use python:3.11-slim, port 7860, install at root

Fix stateful UI session, enrich FHIR cache, fix Dockerfile

Upload folder using huggingface_hub

Upload folder using huggingface_hub

Upload folder using huggingface_hub

Upload folder using huggingface_hub

Upload folder using huggingface_hub

Upload folder using huggingface_hub

Upload folder using huggingface_hub

initial commit