Guides developers in building health-adjacent AI with proper RAG pipelines, hard-coded escalation triggers, and EU AI Act compliance.
Medical Prompt Safety Architect
Design safe AI systems for health domains
Best for
- ▸Design RAG architectures for medical accuracy and hallucination reduction
- ▸Implement hard-coded emergency escalation trigger systems
- ▸Build compliant health AI under EU AI Act, FDA SaMD, and MDR frameworks
- ▸Create uncertainty quantification and disclaimer systems
What you'll get
- ▸Architecture review of a symptom checker app with 12 safety gaps and EU AI Act classification
- ▸RAG pipeline design for a patient education platform with source hierarchy and retrieval targets
- ▸Escalation trigger system specification with emergency pattern matching and bypass logic
- ▸EU AI Act compliance checklist for a Class IIa health information chatbot
A health-adjacent AI system design, architecture proposal, or specific safety question
Architecture assessment with risk classification, safety gaps, recommended design patterns, and compliance checklist
What's inside
“You are a Medical Prompt Safety Architect. You design AI health systems that catch the specific failure modes that kill patients and evade regulators. - You treat RAG as a hard firewall, not decoration. You know that unconstrained LLMs hallucinate contraindicated treatments 88% when prompted adversa...”
Covers
Not designed for ↓
- ×Building actual medical devices (requires specialized regulatory consultants)
- ×Providing medical advice or clinical decision support directly
- ×Replacing regulatory legal counsel for certification
- ×Designing systems for autonomous clinical decision-making
SupaScore
89.5▼
Evidence Policy
Standard: no explicit evidence policy.
Research Foundation: 8 sources (4 paper, 3 official docs, 1 industry frameworks)
This skill was developed through independent research and synthesis. SupaSkills is not affiliated with or endorsed by any cited author or organisation.
Version History
v5.5 final distill
Works well with
© 2026 Kill The Dragon GmbH. This skill and its system prompt are protected by copyright. Unauthorised redistribution is prohibited. Terms of Service · Legal Notice