quiz-wdyk

Generate concept questions dynamically from your arsenal and index, present them in a persona's voice, score answers on learning dimensions, and track progress via spaced repetition. Load Configuration Resolve the user's home directory. echo $HOME Never pass `~` to the Read tool. Tell the user: "Run `/things:setup-things` first." Then stop. Tell the user: "Run `/setup-wdyk` first." Then stop. Extract `building_skills`, `aspirational_skills` from the professional profile, and `default_depth`, `default_persona`, `session_length` from preferences. Load Arsenal and Index Read all files in `/.things/i-did-a-thing/arsenal/` for skill evidence. You haven't logged any entries yet. Run `/thing-i-did` first -- quiz questions are generated from your actual projects and experiences. Load Session History bash /scripts/search-sessions.sh --recent 20 Use session history for spaced repetition: weight toward topics with lower scores, longer time since last quiz, and gaps identified in explore sessions. Select Topic Use the provided topic. Select a topic using spaced repetition weighting: 1. Topics with gap or partial scores from recent explore/quiz sessions 2. Topics from `building_skills` and `aspirational_skills` not recently quizzed 3. Topics with declining scores (tested before but getting worse) 4. Random selection from arsenal skill areas not yet explored I'm thinking we should quiz [topic] -- it's been [reason]. Sound good? Yes -- let's do it Different topic -- I have something in mind Select Persona Use the specified persona. Use `default_persona` from config. Load the selected persona from `/.things/shared/roles/.md`. Adopt their voice for question delivery and feedback. Determine Question Count Use the provided count. Base on `session_length`: - short --> 3 questions - medium --> 5 questions - deep --> 10 questions Generate Questions Dynamically Generate questions from `/.things/i-did-a-thing/index.json` -- NOT from a static question bank. Each question should reference the user's actual projects and experiences. Questions must be dynamically generated from index entries, never from a pre-built question bank.

Question patterns:
- Mechanism: "Explain how [mechanism from their project] works under the hood." (e.g., "Explain how the PostToolUse hook mechanism works in your screenshotr plugin")
- Tradeoff: "What are the tradeoffs of [approach they chose] vs. [alternative]?" (e.g., "What are the tradeoffs of atomic file writes vs. direct writes?")
- Decision rationale: "You chose [decision from log]. What alternatives exist and why did you pick this?" (e.g., "You chose Homebrew for distribution. What alternatives exist and why?")
- Failure mode: "What would go wrong if [component from their project] failed? How would you detect and recover?"
- Generalization: "You applied [pattern] in [project]. Where else would this pattern be useful? Where would it break down?"
- Connection: "How does [concept A from project 1] relate to [concept B from project 2]?"
- Teaching: "Explain [concept from their work] to someone who's never seen it before."

Selection algorithm:
1. Filter index entries by topic relevance (tags, skills_used, skills_developed)
2. For each entry, identify 2-3 potential question angles
3. Score potential questions:
   - `gap_weight = 5 - avg_score_for_topic` (from session history)
   - `recency_bonus = days_since_last_asked / 14` (capped at 2.0)
   - `variety_bonus = 1.5 if question_type not recently used`
4. Select from top-scored questions with weighted randomness

Present Questions Present questions one at a time in the persona's voice.

<template name="question-presentation">
  <output>
  [Persona Name]: question

  Take your time. Think through it carefully before answering.
  </output>
</template>

<constraint>After each answer, briefly acknowledge but don't give detailed feedback yet.</constraint>
<if condition="session-short">
  <action>Give per-question feedback instead of batching.</action>
</if>

Score Each Answer Score each answer on the five learning dimensions (from `references/quiz-rubric.md`): 1. Depth (1-5) -- Surface-level recall vs. deep understanding 2. Accuracy (1-5) -- Technically correct mental models 3. Connections (1-5) -- Links to related concepts and experiences 4. Application (1-5) -- Can apply to new scenarios 5. Articulation (1-5) -- Clear, structured explanation

<substep name="classify-concept">
  <action>Classify each answer's concept as: strong | partial | gap</action>
</substep>

Deliver Feedback After all questions, deliver batched feedback.

<template name="quiz-results">
  <output>
  Quiz Results: [Topic]

  | # | Concept | Score | Strength |
  |---|---------|-------|----------|
  | 1 | concept | x/5 | strong/partial/gap |
  | 2 | concept | x/5 | strong/partial/gap |
  | ... | | | |
  </output>
</template>

<template name="dimension-scores">
  <output>
  Overall Dimensions:

  | Dimension | Score | Notes |
  |-----------|-------|-------|
  | Depth | x/5 | brief note |
  | Accuracy | x/5 | brief note |
  | Connections | x/5 | brief note |
  | Application | x/5 | brief note |
  | Articulation | x/5 | brief note |
  </output>
</template>

<template name="feedback-summary">
  <output>
  What you nailed: strongest answers

  Where to dig deeper: weakest areas with specific suggestions

  From your arsenal: entries that could strengthen weak answers
  </output>
</template>

Log the Session

  <schema name="session-frontmatter">
  ```markdown
  ---
  type: quiz
  date: <date>
  topic: <topic>
  persona: <persona>
  question_count: <N>
  scores:
    depth: <1-5>
    accuracy: <1-5>
    connections: <1-5>
    application: <1-5>
    articulation: <1-5>
    overall: <average>
  per_question:
    - concept: "<concept>"
      score: <1-5>
      strength: strong|partial|gap
    - concept: "<concept>"
      score: <1-5>
      strength: strong|partial|gap
  arsenal_references:
    - file: "<log filename>"
  ---
  ```
  </schema>

  <section name="body">
  ```markdown
  # Quiz Session: <topic> -- <date>

  ## Questions and Answers
  <per-question summary>

  ## Feedback
  <summary of feedback>

  ## Spaced Repetition Notes
  <concepts to re-quiz, suggested interval>
  ```
  </section>

</write>

Update Progress Read and update `/.things/what-do-you-know/progress.json` and `/.things/what-do-you-know/knowledge-map.json` with new scores and concept classifications. Handle Git Workflow Before committing, pull latest changes from the remote (if one exists) to avoid conflicts. git -C /.things pull --rebase 2>/dev/null || true Based on the `git_workflow` setting in `config.json`: Would you like to commit and push this quiz session? Yes No Automatically `git add` the session file, progress.json, and knowledge-map.json, `git commit -m "quiz: "`, and `git push`. Tell the user the session has been saved and they can commit when ready. Offer Next Steps What next? Quiz again (same topic, different questions) Quiz a different topic Explore a weak area from this quiz Bridge a gap to build a learning plan Done for now

  <if condition="quiz-same-topic">
    <action>Go back to Step 7.</action>
  </if>
  <if condition="quiz-different-topic">
    <action>Go to Step 4.</action>
  </if>
</completion-message>