What problem does it solve? Generic AI study plans are one-off, stateless documents that forget your progress. This Skill turns any skill you want to learn into a persistent private tutor that remembers where you are, adapts exercises to your level, and schedules spaced-repetition reviews. ## Core Features & Use Cases - 8-Route Research Pipeline: Runs parallel research on knowledge maps, canonical resources, expert learning paths, deliberate practice, common pitfalls, assessment rubrics, feedback communities, and motivation before building anything. - Three-Layer Tutor Generation: Produces a self-contained {skill}-learn directory with a Dreyfus-staged curriculum (curriculum.json/md), a tutor SKILL.md, and a private learner-state.json with SM-2-lite spaced repetition. - Quality Gates: Validates prerequisite DAGs for cycles, requires assessable exercises per module, enforces behavioral competence checks, and honestly caps the ceiling at "competent". - Use Case: Say "learn Rust" and receive a rust-learn tutor that loads your saved state each session, quizzes due review items, targets your known weak spots, and grades your exercises. ## Quick Start Tell your agent "learn rust" and confirm your current level, goal, weekly hours, and locale so it can research and generate a personalized tutor skill.