It is designed for family-history research across the former USSR, the Soviet Union, and the Russian Empire, using church and civil registers, military records, deportation and repression files, and unindexed archive scans.
The skill defines a research intake, registers for people, artifacts, claims and relationships, a CSV search log, and a final-report template. It assigns claims statuses such as hypothesis, single-source, corroborated, strongly-corroborated, contradicted, and unresolved. Its audit script checks referential integrity and flags weak corroboration. The repository is MIT licensed and shows 242 stars and 14 forks.
Version 1.1.0 adds field lessons drawn from a multi-week case without its personal data, covering FamilySearch image-only records, historical jurisdiction mapping, name variants, military sources, and archive requests. It also says CAPTCHAs must be handled by a human and that agents should stop when public sources are exhausted.
The repository does not establish accuracy rates, successful completion rates, or comprehensive access to the archives it describes.