ex-USSR Kinship Research

# ex-USSR Kinship Research

Turns an AI agent into a careful genealogist for the former USSR and the Russian Empire, with registers, evidence statuses and an HTML report. Made for family historians who use Codex or Claude Code.

Made by DenisSergeevitch on GitHub. These notes are Mindket's, written from its README; the project's own files and license are in its repository.

## About

What it is: ex-USSR Kinship Research is a genealogy skill for AI agents such as Codex and Claude Code. It makes the agent read church books, civil records, military files and unindexed archive scans for the former Soviet Union and the Russian Empire. Every find stays traceable, and the agent says what is proven and what is not.

Who it suits: Family historians with roots in the former USSR suit it best. Researchers who need an audit trail for each claim will also find it useful.

```table
Industry | Who uses it | How they use it | Example
Family history | Hobby genealogists | Trace ancestors through church books, civil records and archive scans | A person asks the agent who a great-grandmother's parents were and which archive holds the record
Professional genealogy | Researchers who take on cases for others | Keep registers and a search log so each conclusion can be checked later | A researcher hands over an HTML report that lays out every relationship and every failed search
Historical research | Historians and archive readers | Follow military, deportation and repression files and keep sources apart from interpretation | A historian marks which claims rest on one source and which have several independent ones
```

If your family has no roots in the former USSR or the Russian Empire, you can skip it.

Where to use it:

- Codex and Claude Code
- Other tools that run Agent Skills
- A lawful browser session signed in to FamilySearch
- A connected FamilySearch MCP server
- Email or official web tools, for requests to archives and libraries

```table
Use case | Who | What they get back
Research one family line | Family historians | Registers of people, sources, claims and relationships, plus a search log in CSV
Check the evidence | Researchers | An audit that flags weak corroboration before conclusions are written
Ask an archive for a record | Anyone chasing a vital record | A drafted request or a ready-to-send letter, and an editable request form template
Read the result | Relatives or clients | A complete HTML report with contradictions, failed searches and next steps
```

How to set it up: You need an agent that runs Agent Skills, such as Codex or Claude Code. Node.js is only needed for the optional Skills CLI. FamilySearch work also needs a lawful browser session or a connected FamilySearch server.

1. Run the optional Skills CLI, pick your agent, and the skill goes into the project you are in.
2. Add the -g flag for a personal installation that works across projects.
3. Or clone the repository straight into your agent's skills folder.

How to use it: Tell the agent your research question, the facts you already know, who may still be alive and what is off limits. It then works through the parts below.

```table
Tool | What it does
Case intake | Fixes the question, the known facts and the limits before any search
Registers | Hold persons, artifacts, claims and relationships, with a search log and a report template
Audit script | Checks that references hold together and flags weak corroboration
Institutional requests | Drafts requests to archives, libraries, universities and museums, and logs each outcome
FamilySearch helpers | Search indexed records, and read a connected server for profiles, relatives and sources
WWI source search | Looks up World War I reports, hospital stays, awards and prisoner-of-war files for a person or a place
```

Example: You ask the agent to research the parents of a great-grandmother born in a village in Lithuania, and you say which relatives may still be alive. The agent logs each search, labels each claim with an evidence status and sends back an HTML report. The report shows each relationship, the contradictions it found, the searches that failed and what to try next.

For newcomers and experts: A plain AI chat tends to guess about old family records. This skill makes the agent work like a careful researcher instead. It writes down every record, says where it came from and labels each claim as proven or not. A first small thing to try is giving it one question about one person.

Underneath, each event is placed in the state, province, district, parish or registry that held it on that date, not on today's map. Claims carry a status such as hypothesis, corroborated or contradicted, and a record and its copies sit in one independence group, so repeating a source cannot pass for confirmation. Namesakes stay separate people until distinguishing evidence appears. The search helpers do not establish identity or kinship by themselves, so the final judgment stays with you.

```flow
You state one research question
The agent searches and logs every find
Each claim gets an evidence status
The audit script flags weak support
An HTML report lists next steps
```

Good to know:

- Coverage: 15 former Soviet republics. Each archive is picked by where the event belonged at the time and who holds the records now.
- Language: the skill's own instructions are in Russian, but the agent replies in your language. It keeps each source's original spelling and adds a translation.
- Output: an HTML report is the default, and a PDF is made only when you ask for one.
- Requests to archives are sent only when you authorize it.
- FamilySearch work goes through a lawful browser session and keeps only the evidence fragments the source permits.
- The archive request template is anonymized, so check the archive's current form and submission rules before you use it.

## On GitHub, 6 November 2026

- Stars: 2,005
- Forks: 203
- Watching: 13
- Open issues and pull requests: 3
- Language: Python
- License: MIT
- Topics: agent-skills, archives, claude-code, codex, family-history, familysearch, genealogy, ussr

## Links

- Code: https://github.com/DenisSergeevitch/ex-ussr-kinship-research
- Releases: https://github.com/DenisSergeevitch/ex-ussr-kinship-research/releases
- Issues: https://github.com/DenisSergeevitch/ex-ussr-kinship-research/issues
- Star history: https://www.star-history.com/#DenisSergeevitch/ex-ussr-kinship-research&Date

## Get the code

```
git clone https://github.com/DenisSergeevitch/ex-ussr-kinship-research.git
```

Then follow the setup in its README. Nothing in this kit runs by itself.

README.md  6 KB