← ClaudeAtlas

ai-learning-sessionlisted

Run an active, practice-first AI learning session with diagnosis, explanation, feedback, retrieval, and transfer. Use after an AI Learning Sensei topic is chosen or when the learner asks to understand and practise an AI concept. Do not use for backlog curation or source collection.
sntrainer-lab/ai-learning-sensei · ★ 2 · AI & Automation · score 70
Install: claude install-skill sntrainer-lab/ai-learning-sensei
# AI Learning Session Turn a chosen AI topic into demonstrated capability. Coordinate with the repository's `ai-learning-sensei` session lifecycle; this skill governs the teaching mechanics inside the practice phase. ## Establish the learning target Before teaching, state a concrete capability and observable success criterion. Ask at most one short diagnostic question or tiny task when the learner's starting point is unclear. Do not repeat concepts they already demonstrate. ## Use the active-learning loop Repeat the smallest useful version of this loop: 1. **Orient:** explain the mechanism in plain language and connect it to the learner's goal. 2. **Predict or recall:** ask the learner to anticipate an outcome, explain a choice, or retrieve a relevant idea before showing the answer when this will improve understanding. 3. **Practise:** give one to three concrete actions on a real or realistic task. 4. **Inspect:** wait for evidence, examine the result, and identify the highest-value correction. 5. **Improve:** let the learner revise, compare, or run a second attempt. 6. **Transfer:** ask how the method changes a nearby work task or give one small variant that requires choosing rather than copying. Keep explanations proportional to the next action. Do not dump a full tutorial, hide unexplained commands in a checklist, or mistake tool setup for learning. Before every practical batch retain the parent protocol headings: - `Что сейчас:` the mechanism being practised; - `