tutor

What problem does it solve? Answering a question once rarely builds real understanding, and unstructured explanations leave learners without practice or feedback. This Skill turns a tutoring request into a focused lesson with one observable objective, grounded in supplied or trusted sources, followed by a practice task and specific feedback on the learner's actual attempt. ## Core Features & Use Cases - Calibrated lessons: Ties each session to a small observable outcome based on the learner's goal, prior knowledge, and supplied materials, without turning a simple question into a course. - Source-grounded practice: Anchors explanations and exercises in primary documentation or supplied material, clearly separating source claims from inference and disclosing when no source is available. - Evidence-based feedback: Waits for the learner's real attempt, gives specific corrective feedback, and checks transfer with a changed example before claiming understanding. - Optional learner-owned records: On explicit request, saves Markdown lesson records separating covered material, self-reported knowledge, and demonstrated performance, with supersession handling for corrections. - Use Case: Ask "Teach me how this parser handles missing fields using this spec, and give me one exercise" to get a grounded mini-lesson, a practice task, and feedback on your answer. ## Quick Start Ask the agent to teach you a specific concept from a document or spec you provide and to give you one practice exercise with feedback on your answer.