question-factory

name: question-factory description: Generates high-quality, curriculum-aligned multiple choice questions from source texts using AI.

Question Factory Skill

This skill defines the methodology for generating educational content from raw text.

🧠 Pedagogical Principles

Bloom's Taxonomy Levels

1. Recall (Level 1): Basic facts, definitions, recall.
   • Keywords: What, Who, Define, List.
2. Comprehension (Level 2): Understanding concepts, explaining ideas.
   • Keywords: Explain, Summarize, Interpret.
3. Application (Level 3): Using knowledge in new situations (Clinical Cases).
   • Keywords: Apply, Solve, Demonstrate, Use.
4. Analysis (Level 4): Drawing connections, diagnosing.
   • Keywords: Analyze, Differentiate, Compare.
5. Evaluation (Level 5): Justifying a stand or decision (Treatment plans).
   • Keywords: Evaluate, Argue, Select/Justify.
6. Creation (Level 6): Producing new or original work. (Less applicable to MCQ).

Question Quality Standards

• Distractors: Wrong answers must be plausible. No "Mickey Mouse" answers.
• Clarity: Stems should be clear and concise.
• Explanation: Every question must have an explanation that clarifies the correct answer AND addresses why distractors might be chosen.

🛠️ Operational Workflow

1. Ingest: Read text from PDF/Chapter.
2. Chunk: Split text into semantic sections (approx 1000-2000 tokens).
3. Prompt: Send chunk + Level Constraint + "Act as Medical Examiner" prompt to LLM.
4. Validate: Ensure JSON structure matches the app's schema.
5. Insert: Append to the correct topic file.