A genealogy research skill for Codex / Claude Code: enables agents to search church records, civil registries, military service files, exile records, and repression archives from the former Soviet Union and the Russian Empire using an "evidence-first" methodology, ensuring every finding is traceable.
Star count 1058 (+250 / 24h)
Rationale
- Covers the 15 constituent republics of the former Soviet Union, locating records based on the administrative divisions and archival institutions in place at the time of the event, rather than using today's maps
- Conclusions use only interpretable statuses (hypothesis, single source, corroboration, contradiction, unresolved, etc.); copies of the same record are grouped into a single independent set and cannot be treated as multiple independent pieces of evidence
- Includes an audit script to check the citation integrity of the registry and verify that every ID and status in the final report matches; the skill documentation is in Russian, but the agent communicates in your language
Usage
- npx skills add https://github.com/DenisSergeevitch/ex-ussr-kinship-research，或直接 git clone to ~/.claude/skills/ or ~/.codex/skills/
- Use scripts/init_case.py to create case files outside the skill directory, hash_artifact.py to record the SHA-256 for each scanned document, and run audit_evidence.py before writing conclusions
Risks
- Searches on platforms like FamilySearch depend on your own legally logged-in browser session; after authorization, the agent may send emails to archives on your behalf, so you must first confirm the scope of sending
- Research involves sensitive information such as living relatives and repression archives; the author requires case file data to be stored outside the skill repository; license is MIT