← all repositories
DenisSergeevitch/ex-ussr-kinship-research

An AI genealogist that refuses to guess your family tree

This skill turns Codex or Claude Code into a rigorous family-history researcher for the former USSR, with evidence statuses instead of vibes.

★1.1k stars Python Domain AppsCoding Assistants
ex-ussr-kinship-research
Collecting fresh signals — velocity needs a few days of history.
collecting data…
star history

What it does

It’s an Agent Skill (for Codex, Claude Code, and similar runtimes) that walks an agent through genealogy research across all 15 former Soviet republics — church books, civil registries, military, deportation and repression records. The agent keeps CSV registers of persons, artifacts, claims and searches, drafts archive requests, and produces a final report where every relationship is argued edge by edge. Instructions are written in Russian, but the agent converses in your language and quotes sources in their original spelling.

The interesting bit

The evidence protocol is the real star: claims get explainable statuses (hypothesis, single-source, corroborated, contradicted…) rather than confidence percentages, and copies of the same record are grouped into an independence_group so a mirrored scan can’t masquerade as two sources. An audit script enforces this mechanically — corroboration can’t be declared into existence.

Key highlights

  • Historical jurisdiction model resolves events to the parish or district as of the event date, not today’s map
  • Name-variant matrices handle pre-reform Russian, Ukrainian, Polish, Romanian, German and Yiddish forms, patronymics and transliterations
  • Namesakes stay separate person_ids until distinguishing evidence (house number, parents, military unit) appears
  • Helpers for FamilySearch indexed records and WWI GWAR searches feed the existing registers rather than auto-claiming identities
  • Refuses surveillance-adjacent requests: no current addresses of living people, no CAPTCHA-solver services, no paywall or rate-limit evasion

Caveats

  • The scripts are plain Python 3 stdlib, but build_pdf.py needs a Chromium-based browser
  • FamilySearch work rides on your existing lawful browser session, with documented coverage limits
  • Regional guides (Galicia, Bessarabia, Chechnya deportation) are examples, not substitutes for local research — the README says so itself

Verdict

Anyone tracing family through the former USSR — where borders moved and names mutated — gets a genuinely disciplined research assistant. If you want a chatbot to conjure a family tree in one prompt, this skill will disappoint you on principle.

Frequently asked

What is DenisSergeevitch/ex-ussr-kinship-research?
This skill turns Codex or Claude Code into a rigorous family-history researcher for the former USSR, with evidence statuses instead of vibes.
Is ex-ussr-kinship-research open source?
Yes — DenisSergeevitch/ex-ussr-kinship-research is open source, released under the MIT license.
What language is ex-ussr-kinship-research written in?
DenisSergeevitch/ex-ussr-kinship-research is primarily written in Python.
How popular is ex-ussr-kinship-research?
DenisSergeevitch/ex-ussr-kinship-research has 1.1k stars on GitHub.
Where can I find ex-ussr-kinship-research?
DenisSergeevitch/ex-ussr-kinship-research is on GitHub at https://github.com/DenisSergeevitch/ex-ussr-kinship-research.

heatdrop uses Google Analytics to see which pages get read — nothing else. Your call. How we handle data.