quizlisted

# /quiz [<course-slug>] [--lessons 1,3] [--weak] [--n <questions>] Interactive retrieval practice — the core of the testing effect. Read `docs/pedagogy.md` (quiz rules) and `docs/vault-schema.md` (mastery update rule, quiz-log schema) first. Vault: `<vault>`. ## Scope selection - `--lessons`: exactly those lessons of the course. - `--weak`: lessons with `mastery < 60`, plus stale ones (`last_quizzed` > 21 days, `mastery < 85`). - Default: interleaved mix across the course's `status: read|quizzed|mastered` lessons, weighted toward low mastery and staleness. No course given → mix across all active courses (max interleaving). - Read the scoped lessons AND their concept notes to ground questions. Default length: 6–10 questions (`--n` overrides). ## Session loop (one question at a time) 1. Before asking, privately write a 10-point rubric for the question. 2. Ask conversationally. Mix formats per pedagogy.md: free recall, explain-why, worked problems (user answers in text/LaTeX-ish), transfer questions. Never multiple choice. 3. Every 2–3 questions, ask for a confidence rating (1–5) before grading — flag high-confidence misses. 4. Grade /10 against the rubric with partial credit. Give elaborative feedback: correct answer, why, misconception note if relevant, link to the concept note. If the missed knowledge is encoded in the user's mind palace (grep `palace/` for a link to the concept), add the location as a retrieval cue ("this lives in your palace — <room>, <spot>") — cue th