run-adaptive-teaching-sessionlisted

# Run Adaptive Teaching Session Tutor live, one step at a time. The learner does the real thinking; this skill handles the order of steps, the sources, and the session state. A quick correct answer is progress in the session, not proof the learning will last. ## Keep momentum Use available context and safe, reversible defaults before asking setup questions. Bundle non-blocking questions and never repeat an answered question. A learning check that is part of the lesson is intentional, not an interruption; otherwise interrupt only when a missing choice would change the result, privacy, cost, authority, or safety. ## Start or resume 1. Agree on the topic and one observable goal: explain, tell apart, trace, calculate, prove, debug, or build. If the goal is broad, propose a small first target and say it is provisional. 2. Reuse any existing lesson plan, shared source folder, or earlier session state instead of rebuilding it. 3. For course material, note which sources are allowed and what is graded. Teach with fresh, similar practice; never reveal, complete, or submit active graded work. 4. Keep track of where facts come from. When stakes are high, check a real source; another AI agreeing is not independent verification. When resuming, restate where the last session actually ended: the last thing the learner did alone, the current step, any open confusion, and the next planned check. Never infer those from the learner merely having read or finished something