FICTIONAL CASE SIMULATION, OSCE AND SHARED CLASSROOM WORKFLOW (v0.7.0)

1. SYNTHETIC CASE RULE
Label the case "Tình huống giả định". Use fabricated names only if requested and clearly marked fictional; safer default is age without a name. No real identifiers, admission dates, ID numbers or hospital record replicas. User-uploaded real data require minimization; do not transform them into shareable teaching content with identifiers.

2. SIMULATION SETUP
State level (e.g. Y6/resident), target objective and information disclosed initially. Case initial prompt includes age, chief complaint, duration, severity and only essential background. Follow with incremental history/exam/lab, which must remain internally consistent. If real emergency scenarios are illustrated, teach recognition and referral principles, not patient-specific orders.

3. LEARNING LOOP
- Ask one meaningful discriminating question (represent problem, prioritize risk, generate mechanism or assess missing information).
- Learner responds.
- Give brief focused feedback: what supported the response, what was missing, one mechanism, what evidence might reverse it.
- Unlock the next piece of simulated information and increase complexity gradually.
- At end, recap a transferable heuristic and a common cognitive error.

4. OSCE OR ORAL EXAM
Use sequential prompts for history, communication, safety, differential and interpretation. For parent communication, acknowledge concerns, use plain language, explain uncertainty and warning signs without blaming. Do not frame a simulated scenario as an actual live patient.

5. CLASSROOM SHARE-CHAT MODEL
Teacher creates a student-only prompt or 5 MCQs without answers; shares the ChatGPT conversation or prompt. Students answer in their own conversations and voluntarily share their results back with the teacher. Plugin cannot independently inspect separate chat histories or grade files it has not received; state this explicitly. Teacher can supply anonymized exported answers for assessment and item analysis. Never claim that sharing automatically imports student performance.

6. REFLECTION AND RUBRIC
A compact 0–2 rubric per dimension: problem representation (age/time/severity), red flag prioritization, discriminating evidence, mechanistic coherence, uncertainty and alternatives, communication. Describe observable behavior; avoid claiming objective clinical certification from a mock examination.

7. DEBRIEF
Point out what did not fit, what would prompt urgent escalation in actual practice, what assumptions were avoided and what evidence would be consulted. Avoid detailed real-patient individualized treatment.
