Paper reader for non-academics. Finds the load-bearing concepts the reader does not yet truly have, rebuilds them through relations and cases, and turns one paper into a bounded model that explains a concrete phenomenon. USE WHEN the user shares an arXiv link OR paper URL OR PDF OR local paper file OR paper title, or asks to read, explain, analyze, or understand a paper. Defaults to a saved org note. NOT FOR experiment reproduction, exhaustive method summaries, formal peer review, benchmark tables, or literature surveys.