Personalized Health AI Agent
This challenge tasks you with creating a personalized, proactive health and wellness AI agent. This agent will leverage Gemini 2.5 Pro (especially its Deep Think mode for complex medical reasoning and multimodal capabilities for analyzing health data) to provide tailored advice. It will use an A2A (Agent-to-Agent) protocol for secure, privacy-preserving communication with specialized medical knowledge agents or hypothetical external health services, demonstrating advanced multi-agent collaboration and data exchange.
What you are building
The core problem, expected build, and operating context for this challenge.
Develop a personalized health agent using Gemini 2.5 Pro with A2A Protocol for secure communication and Deep Think mode for complex reasoning.
How work is evaluated
Evaluation will assess the personalized agent's ability to accurately interpret health data, provide relevant and safe advice, and correctly utilize the A2A protocol for specialist consultation. Emphasis on reasoning quality and secure communication.
Challenge datasets
Public datasets, plus the private datasets you have access to.
What you should walk away with
Employ Deep Think mode for nuanced medical reasoning and personalized recommendations.
Integrate multimodal input (simulated sensor data, health charts) for comprehensive analysis.
Design and build an A2A protocol for secure, asynchronous agent communication.
Implement a core personalized health agent utilizing Gemini 2.5 Pro's capabilities.
Reference links and supporting material
Sample data for 1 tasks
Challenge setup
Evaluation will assess the personalized agent's ability to accurately interpret health data, provide relevant and safe advice, and correctly utilize the A2A protocol for specialist consultation. Emphasis on reasoning...
Challenge input
{ "user_profile": { "age": int, "gender": "string", "conditions": ["string"], "preferences": ["string"] }, "daily_data": { "steps": int, "calories_...
Agent execution
No agent tool is listed for this challenge.
Evaluated output
{ "recommendation_text": "string", "flags": ["string"], "specialist_consult_request": { "agent_type": "string", "query": "string" } }
- Are the recommendations safe, non-harmful, and appropriate for the user's profile?
- Was the specialist agent correctly identified and queried via A2A when appropriate?
- Relevance Score target: 85
- Python execution harness
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
No tool records are attached.
Evaluation contract
- Are the recommendations safe, non-harmful, and appropriate for the user's profile?
- Was the specialist agent correctly identified and queried via A2A when appropriate?
Metrics
- Relevance Score · Target 85
- Reasoning Depth · Target 80
Workflow tasks
- Task 1
Input: { "user_profile": { "age": int, "gender": "string", "conditions": ["string"], "preferences": ["string"] }, "daily_data": { "steps": int, "calories_intake": int, "sleep_hours": float, "blood_pressure_sys": int, "blood_pressure_dia": int, "mood_score": int, "glucose_chart_base64": "string" } }
Output: { "recommendation_text": "string", "flags": ["string"], "specialist_consult_request": { "agent_type": "string", "query": "string" } }
Execution harness: Python
Recipe freeze marker
No recipe freeze time is recorded.