What problem does it solve? Repeated literature searches waste time and API quota, and keyword-only retrieval routinely misses key papers. This Skill enforces a local-first knowledge base where every paper is searched once, stored once, and reused across research directions. ## Core Features & Use Cases - Local-first retrieval: Checks existing notes and a query log in .kilo/knowledge/papers/ before calling any external API, with a 7-day deduplication window for identical queries. - Citation-graph expansion: Resolves seed papers to stable IDs (DOI, arXiv, Semantic Scholar, OpenAlex) and chases both backward references and forward citing papers to surface competitors that keyword search misses. - Mandatory capture and quality judgment: Every cited paper is stored as a structured note with role, quality, and reading-depth fields, plus implementation-level facts for directly related work. - Use Case: When starting a new research direction, ask for a literature scan; the Skill searches arXiv, expands from seed papers via Semantic Scholar, judges each paper's evidence quality, and writes reusable notes so the next direction never pays for the same search twice. ## Quick Start Search the literature on my research topic, expand from this seed paper through its citing works, and save the relevant papers into the local knowledge base.
