Inspiration

Living in Lagos, I see every day how common symptoms like fever, chills, headache, and body pain create chaos. People often wonder: “Should I rest at home, visit a clinic, or rush to emergency?” Malaria and typhoid are very common here, and long hospital queues waste precious time and money. I wanted to build a fast, friendly AI agent that gives safe, instant guidance using real healthcare standards — so patients get the right care faster and clinics aren’t overloaded.

What it does

LagosSmartTriageOrchestrator is a friendly A2A healthcare triage orchestrator built on Prompt Opinion.

Patients describe symptoms in plain English (e.g. “I have high fever, chills, and headache for 3 days”). The agent:

  • Asks smart follow-up questions
  • Determines urgency level (Emergency / Urgent Clinic Visit / Home Care with practical advice)
  • Gives safe next-step guidance
  • Hands off complex cases via A2A to a DoctorAgent or PharmacyAgent

Everything uses only synthetic/de-identified data. It’s designed specifically for Nigerian healthcare realities while staying 100% compliant with SHARP Extension Specs and FHIR standards.

How we built it

I built the entire solution solo on the Prompt Opinion platform using the Orchestrator Agent (Path B — A2A route).

  • Powered by Google Gemini 3 Flash Preview (free tier)
  • Created two key skills: PerformSymptomTriage and HandoffToDoctorAgent
  • Enabled full A2A communication between agents
  • Added SHARP Extension Specs + FHIR Context Extension for seamless healthcare data propagation and interoperability
  • No heavy local setup — everything ran in the browser/cloud on my low-spec laptop (i5-2450M + 8GB RAM)

The anti-gravity cloud approach made it possible with zero local GPU or heavy installs.

Challenges we ran into

The biggest challenge was moving past the limited built-in templates and properly configuring a clean A2A workflow with MCP-style tools. It took some trial-and-error to get the SHARP + FHIR context flowing correctly between agents, but once it clicked, the platform handled the heavy lifting beautifully.

What's next

I plan to keep iterating after the hackathon — adding more skills (e.g. medication interaction checker) and testing with even more Lagos-specific scenarios. This agent has real potential to reduce unnecessary hospital visits and help people get care faster across Nigeria.

Try it yourself on Prompt Opinion Marketplace!

Built With

  • a2a-communication
  • fhir
  • fhir-context
  • gemini-3-flash-preview
  • prompt-opinion-platform-(orchestrator-agent)
  • sharp-extension-specs
Share this project:

Updates