怎么用（约 30 分钟）：
- 第1站扫 12 篇「为何读」+ 点开有代码的 2–3 篇仓库（约 12 分）
- 第2站按四条方法线各看标题与一条代码仓（约 8 分）
- 第3站收藏 8–10 个 GitHub org（约 5 分）
- 第4站回到主题包导览，按 Cardiology / Dialogue / Agent 深化（约 5 分）
第1站 · 必读12 篇 · Agent / 对话 / 指南RAG / 过程奖励 / 分诊 交叉最高
1CliCARE: Grounding Large Language Models in Clinical Guidelines for Decision Support over Longitudinal Cancer Electronic Health Records
指南接地 + 纵向病历决策，Cardiology-Chat / CDSS 的核心参考。
2MED-COPILOT: A Medical Assistant Powered by GraphRAG and Similar Patient Case Retrieval
GraphRAG + 相似病例检索的医学助手，对话/会诊 UI 可直接对标。
3MedAgent-Pro: Towards Evidence-based Multi-modal Medical Diagnosis via Reasoning Agentic Workflow
证据驱动的多模态诊断 Agent 工作流，多 Agent 会诊骨架。
4MMedAgent-RL: Optimizing Multi-Agent Collaboration for Multimodal Medical Reasoning
多 Agent 协作的 RL 优化，会诊角色分工与奖励设计可复用。
5MDTeamGPT: Mitigating Context Collapse and Enabling Self-Evolution in Medical Multi-Agent Reasoning
医学多 Agent 自演化与上下文坍塌缓解，团队会诊必读。
6Ask Patients with Patience: Enabling LLMs for Human-Centric Medical Dialogue with Grounded Reasoning
以人为本的主动问诊与 grounded 推理，ProMed 式追问直接相关。
7CLARITY: Clinical Assistant for Routing, Inference, and Triage
路由 / 推理 / 分诊一体助手，院前分诊与 Cardiology-Chat 入口。
8Knowing When to Abstain: Medical LLMs Under Clinical Uncertainty
临床不确定时知何时弃答，安全层与转诊策略必备。
9SOAPTriage: SOAP-Guided Multi-View Clinical Text Modeling Framework for Automated ESI Prediction
SOAP 结构 → 急诊 ESI，分诊文本建模可落地。
10From Conflict to Consensus: Boosting Medical Reasoning via Multi-Round Agentic RAG
多轮 Agentic RAG 从冲突到共识，指南检索 × Agent 交叉。
11MIND: Unified Inquiry and Diagnosis RL with Criteria Grounded Clinical Supports for Psychiatric Consultation
问诊 + 诊断统一 RL，过程奖励 × 主动追问方法线。
12LLM-Based Multi-Agent Systems for Clinical Workflows: A Survey of AI Hospitals
AI Hospital / 多 Agent 临床工作流综述，30 分钟扫全局。
第2站 · 方法线Discover 高重叠切片 · 各 3 篇
Abstention / Defer / Triage
Knowing When to Abstain: Medical LLMs Under Clinical Uncertainty
不确定弃答校准
CLARITY: Clinical Assistant for Routing, Inference, and Triage
路由+分诊一体
SOAPTriage: SOAP-Guided Multi-View Clinical Text Modeling Framework for Automated ESI Prediction
SOAP→ESI 分诊
Guideline-RAG
CliCARE: Grounding Large Language Models in Clinical Guidelines for Decision Support over Longitudinal Cancer Electronic Health Records
指南×纵向病历
MED-COPILOT: A Medical Assistant Powered by GraphRAG and Similar Patient Case Retrieval
GraphRAG 助手
Case2Flow: Bridging Patient Cases and Guideline Flowcharts through Multimodal Retrieval
病例↔指南流程图
ProcessReward × Clinical Dialogue
MIND: Unified Inquiry and Diagnosis RL with Criteria Grounded Clinical Supports for Psychiatric Consultation
问诊诊断统一 RL
CoMMa: Contribution-Aware Medical Multi-Agents for Decentralized Oncology Decision Support
贡献感知多 Agent 决策
AI Diagnostic Assistant (AIDA): A Predictive Model for Diagnoses from Health Records in Clinical Decision Support Systems
病历→诊断 CDSS
Agentic Tool × Clinical Workflow
DermAgent: A Multi-Tool Agentic System for Evidence-Grounded and Traceable Reasoning in Dermatology
多工具证据可追溯
From Conflict to Consensus: Boosting Medical Reasoning via Multi-Round Agentic RAG
多轮 Agentic RAG
MedAgentGym: A Scalable Agentic Training Environment for Code-Centric Reasoning in Biomedical Data Science
Agent 训练环境
第3站 · 跟仓follow_labs_shortlist · 10 org
jinlab-imvrMedAgent-Pro
Deep-AgentMMedAgent-RL
KaiChenNJMDTeamGPT
sakurakawa1CliCARE
Quark-MedicalATPO 多轮对话
LucanycOccuTriage / TriageAgent
Lingxi-mental-healthMIND 问诊 RL
NJU-RLMA-RAG
AnonymousSubmission996CLARITY
HovChenAegle 多学科 Intake
第4站 · 主题包链到 THEME_PACKS_zh.html
Cardiology心内科优先包 · slug
CardiologyCardiology-x-Dialogue心内 × 临床对话 · slug
Cardiology-x-DialogueMultiagent-hospital多 Agent / AI Hospital · slug
Multiagent-hospitalGuideline-Adherence-CDSS指南依从 × CDSS · slug
Guideline-Adherence-CDSSHospital-Triage-Ops院内分诊 / 运营 · slug
Hospital-Triage-OpsClinical-Uncertainty临床不确定性 · slug
Clinical-UncertaintyProMed-adjacentProMed 式主动问诊相邻 · slug
ProMed-adjacentMedical-RAG医学 RAG · slug
Medical-RAG