What problem does it solve? Writing effective multiple-choice questions is hard because weak distractors (obviously wrong alternatives) make assessments trivially easy to game. This Skill produces and reviews question drafts grounded in real course materials, with mandatory manual curation focused on distractor quality. ## Core Features & Use Cases - Grounded Question Generation: Creates questions with exactly 4 alternatives (A-D) and one correct answer from local course files, lesson plans, transcripts, or NotebookLM notebooks. - Distractor Quality Review: Applies a per-question checklist to detect absurd alternatives, category mismatches, and correct answers that stand out by length or precision. - NLM-Assisted Drafting: Uses NotebookLM as a generation engine with explicit prompts, then treats output as raw material requiring manual rewriting. - Use Case: A professor needs a midterm quiz on distributed systems. The Skill pulls from the course transcript, drafts 15 questions via NotebookLM, rewrites weak distractors to fail on subtle conceptual nuances, and delivers a reviewable .md draft with an answer key. ## Quick Start Generate a draft of 10 multiple-choice questions from my distributed systems lecture notes and save it as a markdown file for review.