Quiz

Focus on:

• Central thesis or main argument
• Supporting evidence and examples
• Key definitions and terminology
• Logical structure and progression
• Conclusions and implications

Question types: "What is the main argument?", "What evidence supports X?", "How does the author address counterargument Y?"

--technical

Focus on:

• Core concepts and definitions
• Procedures and processes (steps, order)
• Relationships between concepts
• Practical applications
• Prerequisites and dependencies

Question types: "Define X", "What are the steps for Y?", "How does A relate to B?", "When would you use X vs Y?"

--story

Focus on:

• Plot events and sequence
• Character motivations and development
• Themes and symbolism
• Narrative techniques
• Setting and atmosphere

Question types: "What motivates character X?", "What does Y symbolize?", "How does the ending connect to the opening?"

--novel

Same as --story, but:

• Ask about broader character arcs
• Track themes across the work
• May focus on specific chapters if requested

--math

Focus on:

• Definitions (precise statement)
• Theorem statements (hypotheses and conclusions)
• Proof techniques and key steps
• Problem-solving approaches
• Connections between results

Question types: "State the definition of X", "What are the hypotheses of theorem Y?", "Outline the proof strategy for Z", "Solve this problem"

Quiz Behavior

1. Start: Confirm you've loaded the material and state the mode. Ask if user wants to focus on any particular section.

2. Question pacing: Ask ONE question at a time. Wait for response.

3. Format selection (per question):

   • Use AskUserQuestion tool for:
     • Factual recall (definitions, specific facts, dates, names)
     • Sequence questions (what comes first/next)
     • Identification (which character, which concept)
     • Multiple choice is natural fit
   • Writing good multiple choice options:
     • Labels should be short and self-contained (the answer itself)
     • Descriptions are optional tiebreakers, not explanations
     • BAD: label "ERM" with description "Minimizing average loss on training data" (gives away the answer)
     • GOOD: label "Empirical risk minimization" with description "Section 1.2.4" (just a locator)
     • GOOD: label "Ice-albedo feedback" with no description needed
     • For term questions ("What term...", "What does X call..."): show only the term, no description
     • Never let descriptions do the cognitive work — if someone could answer correctly just by reading descriptions, the question tests reading, not recall
   • Use conversational format for:
     • Analytical questions (why, how, explain)
     • Interpretive questions (what does X mean, symbolize)
     • Synthesis questions (connect A to B)
     • Open-ended discussion

4. Feedback: After each answer, provide brief feedback:

   • If correct: Acknowledge, optionally add context
   • If incorrect: Give the correct answer with explanation
   • If partially correct: Acknowledge what's right, clarify what's missing

5. Scoring: Track informally (e.g., "4/5 so far"). Don't be pedantic about scoring.

6. Pacing: After 5-7 questions, ask if user wants to continue, shift focus, or wrap up.

7. Wrap-up: Summarize performance, highlight areas of strength and areas to revisit.

Example Session

User: /quiz --article (Article about climate feedback loops already in context)

Claude: "I'll quiz you on this article about climate feedback loops. It covers ice-albedo feedback, water vapor feedback, and carbon cycle feedbacks. Want to focus on a specific section, or cover all three?"

User: "All three"

Claude uses AskUserQuestion: "Which feedback mechanism is described as a 'positive feedback' that amplifies initial warming?"

• Ice-albedo feedback
• Ocean heat absorption
• Increased cloud cover
• Volcanic aerosol effects

User selects: Ice-albedo feedback

Claude: "Correct. As ice melts, less sunlight is reflected back to space, causing more absorption and further warming.

Next question: Why does the author argue that water vapor feedback is difficult to model accurately?"

(Conversational format for analytical question)

...continues...