What problem does it solve? Building a study course from scratch risks invented grading standards, duplicate courses, and silent conflicts with a learner's existing roster. This Skill discovers real, sourceable curricula and compiles them into a canonical course folder only when a genuine rubric or mark scheme exists, while enforcing roster capacity and level-lock rules. ## Core Features & Use Cases - Sourced course compilation: Searches for real qualifications or cited course instances, shortlists up to 5 options, and transcribes the source's rubric into rubric.json, course.json, and an itemised curriculum_map.json — never inventing grading standards. - Roster and level gating: Runs roster_check.py before building to enforce the incomplete-course cap, detect duplicate canonical courses, resume dropped enrollments, and warn about level-lock consequences. - Whole-syllabus coverage: Itemises every specification item, maps each to a stage, and lets coverage_check.py derive coverage_status (full, partial, or unverified) with honest gap reporting. - Use Case: A learner runs /add-course for GCSE Maths; the Skill checks roster capacity, finds no existing canonical course, shortlists AQA and Edexcel specs, and after the learner picks one, builds the full course folder with lessons, practice, tests, and an exam. ## Quick Start Ask the AI to run /add-course for the subject you want to study and follow the shortlist and confirmation prompts.