What problem does it solve? Conducting literature-grounded research and paper quality audits requires deep reasoning over many sources, but local models lack that capacity and cloud APIs are costly. This Skill turns NotebookLM into a free Gemini-grade knowledge brain, driven entirely through its CLI, with battle-tested protocols that avoid the tool's many pitfalls (upload failures, timeouts, security-scan blocks, project pollution). ## Core Features & Use Cases - Source & Notebook Management: Create notebooks, upload PDFs/Markdown/Drive files, rename sources to BibTeX keys via UUID prefixes, and clean error or duplicate sources. - Iterative Q&A Protocol: One-question-per-round method (Q1 domain map through Q6 experiment design) with 90s+ timeouts, pure-ASCII prompts, and three-step answer verification. - Content Generation: Generate reports, mind maps, infographics, quizzes, flashcards, audio, slide decks, and video, with per-type download commands and timing expectations. - Layer B Paper Quality Audit: Full workflow for auditing papers against a five-dimension rubric with T1/T2 thresholds, plus a manual fallback when Google services are unreachable. - Use Case: Upload a manuscript PDF and its references.bib to a fresh notebook, send a pure-English audit prompt, and save a scored quality report to 07-quality/layer-b-report.md. ## Quick Start Ask the agent to create a fresh NotebookLM notebook, upload your paper PDF after checking its text layer with pdftotext, then run an English Layer B quality audit question with a 90-second timeout.
