What problem does it solve? AI assistants tend to hand over answers immediately, which leaves learners unable to solve the next problem on their own. This Skill turns the assistant into a tutor that diagnoses where a learner is stuck and builds durable understanding instead of delivering finished answers. ## Core Features & Use Cases - Diagnosis before teaching: Locates the learner's actual confusion (concept, procedure, notation, or question framing) with one calibrating question before choosing a teaching move. - A toolkit of tutoring moves: Guided discovery, direct explanation, worked parallel examples, inline visuals, reflective pauses, and study-resource creation (flashcards, quizzes, study guides). - Pushback handling and academic integrity: Distinguishes impatience from genuine stuckness, holds the line on graded work, and declines to produce submittable answers while offering legitimate help. - Use Case: A student says "I keep mixing up eigenvalues and eigenvectors." The Skill asks one diagnostic question, offers a small scaffold or parallel example, and keeps the learner doing the final step until they can explain it back. ## Quick Start Ask the assistant to teach you a concept you are struggling with, for example "walk me through how gradient descent works, I'm rusty on it."
