Подготовка лекции из научной статьи

learning

teach-from-paper is a Claude Code skill that converts a single research paper into a ready-to-build teaching package: a lecture outline, 3–5 teachable results with intuition spelled out, a slide skeleton shaped for `/create-lecture`, discussion questions, and an exercise brief for `/scaffold-exercises`. It reads the paper end-to-end — supporting `.tex`, `.qmd`, `.md`, and `.pdf` via pdftotext — then pitches depth to a stated audience level: undergrad drops proofs and keeps intuition, phd retains identifying assumptions and one key derivation, seminar foregrounds contribution-vs-literature framing. Before extracting anything, it emits a Pre-Flight Report covering the one-line thesis, audience level, time budget, and prerequisites, and waits for confirmation to avoid building a deck pitched at the wrong level. The lecture arc follows motivation → setup → key result → method → takeaways, with roughly one slide per two minutes of target runtime. Output is written to `quality_reports/teach_from_paper_[title].md` and explicitly names both hand-offs. Designed for instructors, PhD students, and researchers turning a specific paper into a lecture, reading group session, or job-market practi