What problem does it solve? Genealogical research requires converting raw records (census sheets, certificates, PDFs, images) into structured, evidence-classified assertions that meet the Genealogical Proof Standard. Doing this manually is error-prone, and misclassifying sources or informants corrupts every downstream conclusion. ## Core Features & Use Cases - Record triage and delegation: Acquires records from search results, ARKs, uploaded PDFs, or images, logs them to the research audit trail, and delegates each to a record-extractor agent for atomic assertion extraction. - Three-layer evidence classification: Applies source (original/derivative/authored), information (primary/secondary/undetermined), and informant-bias analysis per fact, following BCG standards. - Calendar and transcription safeguards: Routes pre-Gregorian dates through convert-dates and flags suspect index transcriptions for original-image confirmation. - Use Case: After a FamilySearch census search returns 12 results, the skill logs the search, delegates each record for extraction, and persists sources and assertions into research.json without duplicating existing entries. ## Quick Start Extract assertions from the 1880 census record for the Schuylkill County household and classify the evidence for each fact.
