ex-ussr-kinship-research

What it does

• Explainable evidence protocol — every claim carries a status (hypothesis, single-source, corroborated, strongly-corroborated, contradicted, unresolved), and copies and mirrors of one record share an independence_group so corroboration cannot be declared into existence
• Historical jurisdiction model — each event is resolved to the state, province, district, parish or civil registry as of the event date, with name-variant matrices covering pre-reform Russian, Ukrainian, Polish, Romanian, German and Yiddish forms, patronymics and transliterations
• Case tooling on the Python standard library — init_case.py for intake, hash_artifact.py for a SHA-256 of every saved artifact, audit_evidence.py checking referential integrity and flagging weak corroboration before you write conclusions, and build_pdf.py for a readable PDF report
• Source-specific reference manuals — FamilySearch indexed search and an optional MCP connection (run through your existing lawful browser session), WWI GWAR record search, Habsburg Galicia and Bukovina, the Chechen deportation, forced labourers in Germany, plus institutional correspondence and government-portal routing

When to use it

Family-history questions where provenance is the point — the README's own examples: "My great-grandfather was exiled to Kazakhstan in the 1930s. Help me find his parents in church records."; handing it a marriage certificate and a soldier card to check whether they are the same person and build the evidence chain; or finding the parish and the right FamilySearch film for an 1890s village, then reading the births for one surname.

Getting started

npx skills add https://github.com/DenisSergeevitch/ex-ussr-kinship-research
# or, for Claude Code personal skills:
git clone https://github.com/DenisSergeevitch/ex-ussr-kinship-research ~/.claude/skills/ex-ussr-kinship-research

The optional Skills CLI needs Node.js; the skill's scripts need only Python 3 and the standard library (build_pdf.py additionally needs a Chromium-based browser, or CHROME_PATH). For a single project, put the folder in .claude/skills/ inside the repository.

Things to know

The skill's instructions are written in Russian — the agent answers in your language and quotes sources with a translation, but the workflow's terminology is Russian-flavoured. Responsible-use rules are built in: it refuses to look up current addresses, phones or accounts of living people, attempts a CAPTCHA at most once before handing off, and steers clear of paywalls, geo-blocks, rate-limit evasion and leaked databases. Keep private case data outside the skill's Git checkout (init_case.py refuses to create a case inside it). FamilySearch research runs through your existing lawful browser session, and coverage of unindexed archives varies by region.