What problem does it solve? Conducting a literature review requires manually searching many disconnected sources — Zotero libraries, Obsidian notes, local PDF folders, arXiv, Semantic Scholar, and the broader web — then de-duplicating and verifying results. This Skill automates that multi-source aggregation and produces a verified, ranked set of relevant papers. ## Core Features & Use Cases - Multi-source search: Queries Zotero (via MCP), Obsidian vaults, local PDF libraries, web search, arXiv API, and optional sources like Semantic Scholar, DeepXiv, Exa, Gemini, and OpenAlex, with graceful degradation when a source is unavailable. - De-duplication and verification: Merges results across sources by arXiv ID, DOI, and title, then runs an anti-hallucination verification step to filter out fabricated paper IDs before analysis. - Optional PDF download: Can download the top-ranked arXiv PDFs into the local paper library when explicitly enabled. - Use Case: A researcher starting a related-work section on diffusion models runs the skill to scan their Zotero annotations and local PDFs, fill gaps with recent arXiv and Semantic Scholar papers, and receive a verified, de-duplicated reading list. ## Quick Start Ask the assistant to run a literature review on your topic, for example: find papers and related work on diffusion models for image generation.
