Healthcare AI Team

A fully operational, specialized multi-agent AI system designed to function as a high-performing healthcare analytics and innovation department.

Vision

Build a team of AI agents that combines deep healthcare domain expertise with strong agentic development, automation, and creative problem-solving capabilities.

Team Composition

Agent	Role	Key Strengths
Orchestrator	Team Lead & Quality Gatekeeper	Planning, prioritization, QA, scope control
Healthcare Data Researcher	CMS, Payer & Regulatory Intelligence	Official data sources, compliance, research depth
Social Intelligence Agent	Trends, Algorithms & Content	X, LinkedIn, healthcare trends, posting strategy
Dashboard & BI Specialist	Power BI + Excel Expert	Dashboards, DAX, KPIs, reporting
Python & Automation Engineer	Scripting & Agent Building	Python, pipelines, custom tools
Tool & Integration Architect	Tool Evaluation & Workflows	n8n, GitHub, new tools, architecture
Creative Strategist	Outside-the-Box Innovation	Novel ideas, product thinking, creative solutions

Operational Cycle

flowchart TD
    A[1. Discovery & Planning] --> B[2. Research & Intelligence]
    B --> C[3. Building & Prototyping]
    C --> D[4. Review & QA]
    D --> E[5. Presentation & Delivery]
    E --> F[6. Feedback & Iteration]
    F --> A

    subgraph Core Agents
        G[Orchestrator]
        H[Healthcare Data Researcher]
        I[Social Intelligence]
        J[Dashboard & BI Specialist]
        K[Python & Automation Engineer]
        L[Tool & Integration Architect]
        M[Creative Strategist]
    end

    G -.-> A & D & F
    H -.-> B
    I -.-> B
    J -.-> C
    K -.-> C
    L -.-> C & Tooling
    M -.-> A & Creative Ideas

Loading

Tool Stack

Core Tools

• Hermes (orchestration)
• Claude Max Pro + Cursor + Grok
• n8n (workflow automation)
• GitHub

Microsoft Tooling (High Priority)

• Microsoft 365 CLI (m365)
• Power BI CLI + Power BI REST API
• Microsoft Graph API
• openpyxl + xlsxwriter (Excel automation)
• Power Automate + Power Apps

Repository Structure

healthcare-ai-team/
├── README.md
├── pyproject.toml          # Python 3.12+ standards + ruff/mypy
├── HERMES_INTEGRATION.md   # How to load as Hermes skills
├── evals/                  # Typed evaluation harness (Hermes self-optimizes here)
├── docs/
│   ├── compliance-phi-guardrails.md
│   ├── microsoft-tooling.md
│   ├── operational-cycle.md
├── diagrams/
│   └── operational-workflow.mmd
├── agents/
│   └── prompts/              # 7 specialized agent roles
└── workflows/
    └── n8n/
        └── healthcare-ai-team-starter.json

Engineering Standards (Added 2026-05-18)

• Python 3.12+ with strict type hints (disallow_untyped_defs = true)
• Ruff + mypy enforced on all emitted code
• Evaluation harness (evals/) — Hermes learning loop target
• PHI Guardrails mandatory (see docs/compliance-phi-guardrails.md)
• Full Hermes skill packaging + auto-evolution ready

Next Steps

1. Define detailed role prompts for all 7 agents (already in agents/prompts/)
2. Set up Microsoft tooling access and documentation
3. Build first working n8n + Hermes workflow
4. Run python -m evals.healthcare_agent_eval and feed scores back to Hermes
5. Iterate via NousResearch/hermes-agent-self-evolution

---

This repository is the central source of truth for the Healthcare AI Team project. Agent-reviewed & hardened.