What problem does it solve? Writing a systematic literature review requires searching multiple academic databases, screening hundreds of papers, scoring relevance, managing citations, and formatting output—an error-prone process that normally takes weeks of manual effort. ## Core Features & Use Cases - End-to-End Review Pipeline: Automates multi-query search (OpenAlex, Semantic Scholar, Crossref), deduplication, AI-based relevance scoring (1-10) with subtopic grouping, high-score-first reference selection, and word-budget planning. - Expert-Style Writing with Hard Validation: Produces a structured review (abstract, introduction, subtopic sections, discussion, outlook, conclusion) with enforced word count and reference count ranges, citation/BibTeX alignment checks, and mandatory PDF and Word export. - Multilingual & Tiered Output: Supports three quality tiers (Premium/Standard/Basic) and translation into English, Chinese, Japanese, German, French, and Spanish with automatic LaTeX compile-error repair. - Use Case: Ask for a Premium-level review on "deep learning for breast ultrasound diagnosis" and receive a validated 10000-15000 word LaTeX review with 80-150 references, exported as both PDF and Word. ## Quick Start Use the research-literature-review skill to write a Standard-tier literature review on "Transformer applications in financial risk control" covering publications from the last five years.
