What problem does it solve? Tracking a research field manually means repeatedly searching arXiv, OpenAlex, Crossref, and Semantic Scholar, deduplicating results, and deciding what is worth reading. This Skill turns that into a recurring automated pipeline that searches, scores, summarizes, delivers, and archives papers on a daily or weekly schedule. ## Core Features & Use Cases - Multi-source retrieval with graceful degradation: Collects up to 30 candidates from arXiv, OpenAlex, Crossref, and Semantic Scholar, automatically switching sources if one is unavailable. - Six-dimension scoring and filtering: Scores candidates on topic match (35), methodology (20), journal quality (15), network relevance (10), applied value (10), and archival value (10), with score validation and triple deduplication by DOI, arXiv ID, and OpenAlex ID. - Digest delivery and archival: Formats top papers into a ranked digest for Feishu, Telegram, or email, then archives structured literature notes with YAML frontmatter without touching curated wikis. - Use Case: A materials science researcher configures keywords for molten-salt corrosion, sets a daily 08:30 cron job, and receives a Top 5 digest each morning with scores, key results, and archived notes. ## Quick Start Tell your agent your research area, keywords, delivery target such as a Feishu group, and archive path, then ask it to set up a daily literature push at 08:30 with 30 candidates and the top 5 delivered.
