Patient Triage & Communication Agent
This challenge tasks developers with building an advanced AI copilot designed to manage patient communications, specifically focusing on initial triage, information dissemination, and empathetic responses. The system will leverage a multi-agent architecture to simulate a healthcare support team, integrating seamlessly with communication platforms like WhatsApp. Participants will focus on ethical AI, data privacy, and the robust integration of MCP-enabled tools for secure access to patient data and medical knowledge bases. The solution must demonstrate adaptive thinking, prioritizing urgent patient queries and providing accurate, compassionate support. This system will orchestrate a team of specialized agents, each with a distinct role: a Triage Agent for initial assessment, an Information Agent for retrieving medical facts, and an Empathy Agent for crafting supportive messages. Developers will implement RAG over a simulated medical knowledge base to ensure factual accuracy and utilize adaptive thinking budgets to allocate processing power based on query complexity and urgency. The core will involve designing secure, MCP-enabled communication channels and tool integrations to interact with a mock WhatsApp API, ensuring patient data is handled responsibly.
What you are building
The core problem, expected build, and operating context for this challenge.
This challenge tasks developers with building an advanced AI copilot designed to manage patient communications, specifically focusing on initial triage, information dissemination, and empathetic responses. The system will leverage a multi-agent architecture to simulate a healthcare support team, integrating seamlessly with communication platforms like WhatsApp. Participants will focus on ethical AI, data privacy, and the robust integration of MCP-enabled tools for secure access to patient data and medical knowledge bases. The solution must demonstrate adaptive thinking, prioritizing urgent patient queries and providing accurate, compassionate support. This system will orchestrate a team of specialized agents, each with a distinct role: a Triage Agent for initial assessment, an Information Agent for retrieving medical facts, and an Empathy Agent for crafting supportive messages. Developers will implement RAG over a simulated medical knowledge base to ensure factual accuracy and utilize adaptive thinking budgets to allocate processing power based on query complexity and urgency. The core will involve designing secure, MCP-enabled communication channels and tool integrations to interact with a mock WhatsApp API, ensuring patient data is handled responsibly.
Shared data for this challenge
Review public datasets and any private uploads tied to your build.
What you should walk away with
Master CrewAI for orchestrating a team of specialized agents (Triage, Information, Empathy) with defined roles, goals, and backstories for patient communication.
Implement MCP-enabled tool integration with Claude Opus 4.5 to securely interact with a mock WhatsApp Business API for receiving and sending patient messages.
Design and deploy a RAG pipeline using LlamaIndex and a vector database (e.g., ChromaDB, Pinecone) over a simulated medical knowledge base for evidence-based responses.
Build A2A protocol communication within CrewAI agents to facilitate seamless collaboration and handover between triage, information retrieval, and empathetic response generation.
Develop extended thinking workflows with Claude Opus 4.1, incorporating adaptive reasoning budgets to allocate more processing steps for complex or critical patient health inquiries.
Integrate ethical AI guidelines and privacy-by-design principles into the agent system, ensuring responsible handling of sensitive patient information.
Utilize prompt engineering techniques to ensure Claude Opus 4.1 delivers compassionate, nuanced, and medically accurate responses, especially in sensitive situations.
How this agent runs
The evaluation will assess the system's ability to accurately triage patient queries, retrieve correct medical information, generate empathetic responses, and demonstrate secure MCP tool integration. Performance will...
Challenge input
{"patient_message": "string", "urgency_threshold": "float"}
Agent execution
The configured agent processes the input under the challenge policy.
Evaluated output
{"category": "string", "is_urgent": "boolean", "reasoning": "string"}
- At least 90% of urgent queries are correctly identified and prioritized.
- Information Agent responses are factually correct with relevant sources for 95% of queries.
- MCP tool integration tasks successfully simulate sending messages and demonstrate security best practices.
- EmpathyScore target: 4.5
- Python execution harness
View technical recipe
Configured tools
No tool records are attached.
Evaluation contract
- The evaluation module defines the checks.
Recipe state
This is a preview. The configuration can change before the evaluation recipe is locked.
Run this agent on your dataset and AI stack
Bring your dataset, model providers, and success criteria. We will scope the right managed run for your team.
Scope a managed run[ok] Wrote CHALLENGE.md
[ok] Wrote .versalist.json
[ok] Wrote eval/examples.json
Requires VERSALIST_API_KEY. Works with any MCP-aware editor.
DocsFind another challenge
Jump to a random challenge when you want a fresh benchmark or a different problem space.