MCP-Server (Model Context Protocol — ein offenes Protokoll zur Anbindung von KI-Assistenten an externe Tools) für die Arbeit mit EGRUL (Einheitliches Staatliches Register juristischer Personen der RF) und EGRIP (Einheitliches Staatliches Register von Einzelunternehmern). Quelle — offizielle Open-Data-Dumps des FTS (Föderaler Steuerdienst).
Status: v0.1.2 — Open-Version (Self-Host über SQLite) vollständig fertiggestellt + Client-Teil hosted Pro (HTTP-Client HostedClient für api.atomno.ru). Veröffentlicht auf PyPI, indexiert in Glama und Smithery. Die hosted Pro-Infrastruktur befindet sich in aktiver Entwicklung. Coverage 100.00% (345 Tests, ruff clean, fastmcp 3.2.4, erzwungen durch --cov-fail-under=100).
Partnerprojekt: mcp-fns-check (Risiko-Check-Ebene über EGRUL).
Sieben MCP-Tools, die für den KI-Assistenten (Cursor, Claude Desktop, Cline, jeder MCP-Client) sichtbar sind:
Tool | Beschreibung | Argumente |
| Suche nach INN (10 Ziffern — jur. Person, 12 — ИП) |
|
| Suche nach OGRN (13) oder OGRNIP (15) |
|
| Fuzzy-Suche nach Name (FTS5) |
|
| Vollständige Karteikarte mit allen Sektionen |
|
| Nur Gründer mit Anteilen |
|
| Nur aktueller Geschäftsführer |
|
| Massenprüfung (bis zu 100 INN) |
|
Plus ein diagnostischer ping, um zu prüfen, ob der Server läuft.
Die vollständige Spezifikation der Payloads befindet sich in src/mcp_egrul/schemas.py (Pydantic-Modelle CompanyCard, IECard, SearchResult, BulkResult).
Related MCP server: onec-meta-mcp
# Без локального clone — работает «из коробки»
uvx atomno-mcp-egrul
# Или установка глобально
pipx install atomno-mcp-egrul
atomno-mcp-egrul
# Или классический pip в venv
pip install atomno-mcp-egrul
atomno-mcp-egrulErfordert Python 3.11+ und uv (schneller Ersatz für pip, optional).
git clone https://github.com/atomno-labs/mcp-egrul
cd mcp-egrul
uv venv
uv pip install -e ".[dev]"Alternativ über pip:
python -m venv .venv
.venv/Scripts/activate # Windows
# source .venv/bin/activate # Linux/macOS
pip install -e ".[dev]"atomno-mcp-egrulDer Standardtransport ist stdio (Standard-Ein-/Ausgabe JSON-RPC). Geeignet für die Verbindung mit Cursor / Claude Desktop / Claude Code.
claude_desktop_config.json){
"mcpServers": {
"egrul": {
"command": "uvx",
"args": ["atomno-mcp-egrul"]
}
}
}.cursor/mcp.json im Projekt oder ~/.cursor/mcp.json global){
"mcpServers": {
"egrul": {
"command": "uvx",
"args": ["atomno-mcp-egrul"]
}
}
}Wenn Sie kein
uvverwenden, ersetzen Sie"command": "uvx", "args": ["atomno-mcp-egrul"]durch"command": "atomno-mcp-egrul"(erfordertpip install atomno-mcp-egruloderpipx install atomno-mcp-egrul).
# 1. Скачайте дампы ФНС (acceptance на сайте ФНС — раз в жизни).
# Источники:
# ЕГРЮЛ — https://www.nalog.gov.ru/opendata/7707329152-egrul/
# ЕГРИП — https://www.nalog.gov.ru/opendata/7707329152-egrip/
# Положите их в структуру:
mkdir -p dumps/egrul/2026-04-24 dumps/egrip/2026-04-24
cp ~/Downloads/EGRUL_*.zip dumps/egrul/2026-04-24/
cp ~/Downloads/EGRIP_*.zip dumps/egrip/2026-04-24/
# 2. Первоначальный полный импорт (однократно, ~30-60 минут):
docker compose --profile import run --rm \
mcp-egrul-import atomno-mcp-egrul-import --registry egrul --full
docker compose --profile import run --rm \
mcp-egrul-import atomno-mcp-egrul-import --registry egrip --full
# 3. Запустите сервер + фоновый cron-демон:
docker compose up -d
docker compose logs -f mcp-egrul-schedulerEtwa 10 Minuten nach dem Import antworten alle Tools (search_by_inn, search_by_name etc.) bereits mit Daten aus dem lokalen FTS-Abbild.
Schema des Volumes /data innerhalb des Containers:
/data/
├── mcp_egrul_data.sqlite # SQLite + FTS5
└── dumps/ # read-only монтируется из ./dumps
├── egrul/
│ └── YYYY-MM-DD/*.zip
└── egrip/
└── YYYY-MM-DD/*.zipDer Cron-Daemon (atomno-mcp-egrul-scheduler) holt sich selbstständig den neuesten Export, nachdem Sie ihn unter dumps/<registry>/<YYYY-MM-DD>/ abgelegt haben — nachts um 03:00 Uhr Europe/Moscow. Wenn nichts Neues vorhanden ist, beendet sich der Job mit nothing_to_import und nimmt keine unnötigen Einträge in import_log vor.
Quellen:
EGRUL open-data: https://www.nalog.gov.ru/opendata/7707329152-egrul/
EGRIP open-data: https://www.nalog.gov.ru/opendata/7707329152-egrip/
Format: Tägliche XML-Archive in ZIP, ca. 15 GB für ein vollständiges Abbild. Rechtlich müssen diese von der FTS-Website nach Akzeptanz der Lizenz heruntergeladen werden — der Server lädt die Archive nicht selbst herunter (strikt).
CLI:
# Полный первоначальный импорт (однократно):
atomno-mcp-egrul-import --registry egrul --full
atomno-mcp-egrul-import --registry egrip --full
# Инкремент (cron / ручной): загружается только если появилась более
# свежая YYYY-MM-DD-папка, чем последний успешный `import_log.source_dump_date`.
# Если новее нет — exit-code 5 и сообщение `nothing_to_import`.
atomno-mcp-egrul-import --registry egrul --incremental
# Фоновой cron-демон с ежедневным 03:00 MSK (вызывать вручную редко;
# обычно запускается сервисом mcp-egrul-scheduler в docker-compose).
atomno-mcp-egrul-scheduler --run-nowExit-Codes von atomno-mcp-egrul-import:
Code | Bedeutung |
0 | Import erfolgreich abgeschlossen |
2 | Ungültige Konfiguration / CLI-Argument |
4 | Ingest-Fehler (beschädigtes XML, kein Dump-Verzeichnis, DB-Fehler) |
5 |
|
api.atomno.ru)Wenn der Benutzer ATOMNO_API_KEY angibt, werden alle sieben Tools automatisch auf die hosted Pro API (SPEC §5.4, §5.4.1) weitergeleitet. Das lokale SQLite wird in diesem Modus nicht verwendet — die hosted Pro bietet:
Aktuelle Daten für heute (ohne die tägliche Verzögerung des Open-Data-Dumps): direkter Scrape von egrul.nalog.ru + Dadata-Fallback auf Serverseite.
Bulk-Endpoint ohne Rate-Limit (POST /companies/bulk) — eine Anfrage statt N lokaler Gather-Operationen.
AI-Summary der Karteikarte, Änderungshistorie, Suche nach dem Namen des Geschäftsführers (Pro-only Tools — kommen zusammen mit dem hosted-Server in Phase 2, siehe §5.4.1).
Preis: Pro — $10/Monat einzeln oder $15/Monat im Paket mit mcp-fns-check (Bundle-Key). Free Tier: 30 Anfragen/Tag/IP ohne Registrierung (SPEC §1).
Konfiguration in Cursor (.cursor/mcp.json):
{
"mcpServers": {
"egrul": {
"command": "uvx",
"args": ["atomno-mcp-egrul"],
"env": {
"ATOMNO_API_KEY": "your-pro-key-here"
}
}
}
}Verhalten und Fehler — kein Silent Fallback: Wenn die hosted API nicht verfügbar ist, löst der Client eine typisierte Exception aus, anstatt stillschweigend Daten aus einem veralteten lokalen Dump auszugeben. Die Zuordnung HTTP ↔ MCP-Fehlercode finden Sie in SPEC §5.4.1:
HTTP-Antwort hosted API | Client-Exception |
|
200 | — | — |
400 |
|
|
401 |
|
|
403 |
|
|
404 (code=not_found) |
|
|
404 (wrong route) |
|
|
413 |
|
|
429 |
|
|
5xx |
|
|
timeout / DNS fail |
|
|
Die Validierung von INN/OGRN bleibt Client-seitig (Prüfziffern werden vor der HTTP-Anfrage geprüft — Einsparung von Round-Trips bei fehlerhaften Identifikatoren).
Variable | Beschreibung | Standard |
| Pfad zur SQLite-Datei mit dem EGRUL/EGRIP-Abbild |
|
| User-Agent des HTTP-Clients |
|
| HTTP-Timeout in Sekunden |
|
| Verzeichnis mit FTS-Dumps, Struktur |
|
| Logging-Level |
|
| Zeitzone für Scheduler (cron 03:00) |
|
| (Pro) Hosted-Abonnement-Key — aktiviert Proxying auf | nicht gesetzt |
| (Pro) Basis-URL der hosted-API |
|
Beispiel — siehe .env.example.
apps/mcp-egrul/
├── pyproject.toml
├── LICENSE # MIT
├── README.md # ЭТОТ ФАЙЛ
├── Dockerfile
├── docker-compose.yml
├── .env.example
├── .gitignore
├── src/mcp_egrul/
│ ├── __init__.py
│ ├── server.py # FastMCP entrypoint, регистрация 7 тулзов + ping
│ ├── context.py # ServiceContext (DI: SQLiteStore + HTTP-клиент)
│ ├── config.py # Чтение env-vars в типизированные поля
│ ├── constants.py # Все магические числа и enum'ы
│ ├── validators.py # Контрольные цифры ИНН (10/12) и ОГРН (13/15)
│ ├── schemas.py # Pydantic-модели CompanyCard/IECard/SearchResult/...
│ ├── errors.py # McpEgrulError и подклассы
│ ├── db/
│ │ ├── __init__.py
│ │ └── sqlite.py # Async-клиент (aiosqlite), init/query/upsert/search + import_log
│ ├── sources/
│ │ ├── __init__.py
│ │ ├── base.py # Абстрактный интерфейс Source
│ │ ├── opendata.py # ФНС open-data адаптер (read-local → SQLite upsert)
│ │ ├── opendata_parser.py # Потоковый lxml.iterparse парсер ЕГРЮЛ/ЕГРИП XML
│ │ └── hosted_adapter.py # HTTP-клиент hosted Pro API (SPEC §5.4.1)
│ ├── tools/
│ │ ├── __init__.py
│ │ ├── search_by_inn.py
│ │ ├── search_by_ogrn.py
│ │ ├── search_by_name.py
│ │ ├── get_full_card.py
│ │ ├── get_founders.py
│ │ ├── get_director.py
│ │ └── bulk_cards.py
│ └── scripts/
│ ├── __init__.py
│ ├── import_opendata.py # CLI `atomno-mcp-egrul-import` (ручной / одноразовый)
│ └── scheduler.py # CLI `atomno-mcp-egrul-scheduler` (apscheduler cron 03:00 MSK)
└── tests/
├── __init__.py
├── conftest.py
├── fixtures/
│ ├── egrul_sample.xml # Мини-ЕГРЮЛ (2 валидных + 1 skip на неизвестный статус)
│ └── egrip_sample.xml # Мини-ЕГРИП (active + closed)
├── test_validators.py
├── test_schemas.py
├── test_config.py # Config.from_env + _parse_float_env (валидация env)
├── test_sqlite_store.py
├── test_cards.py # _cards.py: parse_iso_date/datetime + build_*card
├── test_server_ping.py # FastMCP tool-layer + server.main()
├── test_tools.py # 7 тулзов: happy-path + validation + not_found
├── test_opendata_parser.py # XML-парсер (zip, xml, skip-на-неизвестный-статус)
├── test_opendata_source.py # OpenDataSource.run_ingest (full/incremental)
├── test_integration_import.py # Полный цикл import → search → get_card
├── test_import_cli.py # CLI `atomno-mcp-egrul-import`
├── test_scheduler_cli.py # CLI `atomno-mcp-egrul-scheduler` + _run_scheduler
└── test_hosted_adapter.py # HostedClient + маршрутизация тулзов (respx-моки)pytest -v --cov=src/mcp_egrulAktuelle Coverage: 100.00% (345 tests passed, ruff clean, 1529 statements + 382 branches, 0 misses). Erzwungen durch die Richtlinie --cov-fail-under=100 — jede Regression bricht das CI. Die Tests decken ab:
Validatoren für INN/OGRN/OGRNIP (Prüfziffern);
Config.from_env + Parser für Float-Umgebungsvariablen (Validierung, kein Silent Fallback);
alle 7 MCP-Tools (Happy-Path + Validierung + not_found + bulk partial);
SQLite Store + FTS5 + import_log;
XML-Parser für EGRUL/EGRIP (zip, xml, skip-Eintrag mit unbekanntem Status);
OpenDataSource.run_ingest (full/incremental/nothing_to_import);
vollständiger Integrationszyklus import fixture → search → get_card → bulk;
beide CLI (atomno-mcp-egrul-import, atomno-mcp-egrul-scheduler) — Registrierung von Cron-Jobs, Argument-Parsing, _run_daily_ingest bei all-happy/nothing_to_import/McpEgrulError, vollständiger Zyklus _run_scheduler mit mock-ed asyncio.Event;
FastMCP Tool-Layer über mcp.call_tool() — Serialisierung von Fehlern in strukturierte Dicts, server.main() mit gültigem und ungültigem Env;
HostedClient (hosted Pro API Proxy) — Happy-Path aller 7 Methoden, alle HTTP-Fehler aus SPEC §5.4.1 (401/403/404/413/429/5xx), timeout/ConnectError, ungültiges JSON/Payload vom Server, Client-seitige Bulk-Validierung, async with-Kontext; plus Routing von Tools im hosted-Modus (bei gesetztem ATOMNO_API_KEY — Anfrage geht an api.atomno.ru, nicht an SQLite, INN-Validierung vor HTTP);
Edge-Cases des XML-Parsers (75 separate Unit-Tests für _parse_company/_parse_ie/_parse_share/_parse_director/_parse_founders/address fallbacks/legacy-Attribute/ungültige Längen von INN/OGRN/KPP);
private Helper des SQLite-Stores (_wrap, _prepare_row, _row_to_dict, _normalize_bm25, auto-init über _ensure, Ablehnung ungültiger finish_import-Status);
ServiceContext Reentry-Idempotenz, atexit-Cleanup, Config.from_env ValidationError → Exit-Code 2 aus atomno-mcp-egrul-import CLI.
Externe APIs werden niemals direkt aus den Tests aufgerufen — nur über respx (HTTP-Mocking) und lokale XML-Fixtures (tests/fixtures/).
Alle Quellen sind öffentlich zugängliche Daten des FTS (EGRUL / EGRIP Open-Datasets), deren Verbreitung durch das FZ „Über Information…“ und EGRUL-spezifische Normen erlaubt ist (siehe SPEC §8).
Juristische Personen fallen nicht unter das 152-FZ (Über personenbezogene Daten).
Die Namen von Geschäftsführern und Gründern werden vom FTS selbst im offenen Register veröffentlicht — die Übermittlung dieser Daten ist legal.
Keine Schreiboperationen an irgendeine externe API.
Geheimnisse — nur über Umgebungsvariablen, im Repository — .env.example ohne Werte.
Der Dienst ist ein Aggregator und eine komfortable Schnittstelle für öffentliche Daten des FTS. Nicht mit dem FTS verbunden. Nutzung auf eigene Gefahr. Die Informationen in den Antworten des Dienstes sind kein Ersatz für eine vollständige rechtliche oder finanzielle Bewertung.
MIT. Datei LICENSE im Stammverzeichnis des Ordners.
Массовая выгрузка до 100 карточек за один вызов.
Вернёт объект с полями cards (успешные) и errors (точечные ошибки
по отдельным ИНН) — один плохой ИНН не ломает весь bulk.
| Name | Required | Description | Default |
|---|---|---|---|
| inns | Yes |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries full burden. It discloses important behavioral aspects like returning both successful cards and per-TIN errors, and that one bad TIN doesn't break the whole call. However, it omits whether the operation is read-only or has side effects, and no mention of authorization or rate limits.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is two short sentences with no fluff. The first sentence front-loads the core purpose and capacity, and the second explains the return structure and error handling. Every word contributes value.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the simple input schema (one parameter) and the existence of an output schema (not provided but implied), the description sufficiently covers maximum batch size, return format (cards/errors), and partial failure behavior. It is complete for its complexity.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Although the schema has no description for the 'inns' parameter, the description mentions 'отдельным ИНН' (individual TINs), clarifying that the array contains Russian tax identifiers. With 0% schema coverage, the description adds meaning beyond the raw schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool performs a massive upload of up to 100 cards per call, specifying the action (выгрузка) and resource (карточек). It distinguishes itself from siblings like search_by_inn by handling multiple INNs and returning partial errors.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description indicates usage for batch card retrieval with a capacity limit of 100, but does not explicitly state when not to use it or list alternatives. However, the context of sibling tools and the capacity hint provide implicit guidance.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Текущий руководитель юр.лица по ИНН (только 10-значный ИНН).
| Name | Required | Description | Default |
|---|---|---|---|
| inn | Yes |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description must disclose behavioral traits. It only states it returns the current director, but fails to mention aspects like error handling, data freshness, auth requirements, or side effects. This is insufficient.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single sentence, very concise and front-loaded with the key action and constraint. Every word earns its place, though it could be slightly expanded for additional clarity without losing brevity.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the simple single-parameter tool and the existence of an output schema, the description is minimally adequate but lacks details on edge cases or error responses. It covers the basic purpose but not the full context of usage.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The schema has 0% description coverage. The tool description adds value by specifying that the INN must be exactly 10 digits. However, it doesn't fully compensate for missing schema descriptions, e.g., no mention of format beyond digit count.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool retrieves the current director of a legal entity by INN. It specifies the exact resource (director), action (get), and constraint (10-digit INN for legal entities). This distinguishes it from siblings like get_founders or search_by_inn.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description does not provide any guidance on when to use this tool versus alternatives such as get_founders or search_by_inn. It only describes what it does without contextual or comparative usage advice.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Учредители юр.лица по ИНН (только 10-значный ИНН).
| Name | Required | Description | Default |
|---|---|---|---|
| inn | Yes |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries full burden for behavioral disclosure. It only mentions the TIN length constraint but does not disclose whether this is a read-only operation, error handling, or any side effects.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single sentence that conveys the core purpose and a key constraint without any unnecessary words. It is efficiently structured and front-loaded.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the presence of an output schema (not shown), the description does not need to detail return values. However, it lacks information on error handling, input validation details, or the structure of the founders data, which could be incomplete for an agent.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema has 0% description coverage, leaving the 'inn' property undocumented. The description adds the crucial constraint that only a 10-digit TIN is accepted, which is valuable beyond the schema definition.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool retrieves founders of a legal entity by TIN and specifies the requirement of a 10-digit TIN. This distinguishes it from siblings like get_director which targets a different role.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No explicit guidance on when to use this tool over alternatives like search_by_inn or get_director. It does not state prerequisites or exclusions, leaving the agent to infer usage context.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Полная карточка (все секции: реквизиты, ОКВЭД, учредители, директор).
Хотя бы один из inn / ogrn обязателен. Если переданы оба — используется inn.
| Name | Required | Description | Default |
|---|---|---|---|
| inn | No | ||
| ogrn | No |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden. It discloses that at least one of inn/ogrn is required and inn takes precedence, but does not mention read-only nature, authentication needs, or error handling for missing parameters.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is two sentences and front-loads the purpose. It is concise with no unnecessary words, though structure could be improved with bullet points for clarity.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the presence of an output schema and only two parameters, the description covers the primary requirements: purpose and parameter constraints. It lacks mention of error cases or usage limitations, but is largely complete for a simple tool.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema has 0% description coverage, but the description adds meaning by clarifying that inn and ogrn are alternative identifiers, at least one is mandatory, and inn is used if both are provided. This significantly compensates for the schema's lack of parameter descriptions.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description explicitly states 'Full card (all sections: details, OKVED, founders, director)', clearly indicating it retrieves a complete company card. This distinguishes it from sibling tools like get_director and get_founders, which target specific sections.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies usage for retrieving full card data, but does not explicitly contrast with siblings or state when to use this tool over alternatives like search_by_inn or get_director. No exclusion criteria or context is provided.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Диагностика: сервер жив, сообщает версию и размер локального слепка.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description takes full burden. It discloses that the tool reports version and snapshot size, which is useful. However, it does not mention read-only nature, safety, or potential side effects, though for a ping tool these are minimal.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single, focused sentence that front-loads the main purpose ('Диагностика') and immediately specifies what the tool reports. No wasted words.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given zero parameters, an output schema exists (not shown but referenced), and a set of sibling tools, the description is complete for this simple tool: it tells the agent exactly what the tool returns and its role as a health check.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
There are zero parameters, so the description need not add parameter info. Baseline for zero parameters is 4, and the description appropriately focuses on the tool's output rather than parameters.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool is for diagnostics: checking server liveness, version, and snapshot size. It uses a specific verb ('диагностика') and resource ('сервер'), and is distinct from sibling tools that handle cards, directors, or searches.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies usage for health checks but does not explicitly state when to use this tool vs alternatives, nor does it mention exclusions or prerequisites. Given the clear purpose, usage is implied but not guided.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Карточка юр.лица или ИП по ИНН (10 цифр — ООО/АО, 12 — ИП/физлицо).
| Name | Required | Description | Default |
|---|---|---|---|
| inn | Yes |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations exist, so the description must cover behavioral traits. It only states the searchby TIN and format, omitting any details on error handling, rate limits, or return behavior.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is extremely concise (one short phrase) and front-loaded with the core purpose, containing no wasted words.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a simple lookup tool with an output schema, the description is minimally complete but lacks details on error cases and what the card contains, leaving some gaps.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema has no parameter descriptions (0% coverage), but the description adds critical semantic meaning by explaining the TIN length and entity mapping, significantly aiding correct usage.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool returns a card of a legal entity or individual entrepreneur by TIN, and distinguishes it from siblings by specifying TIN format (10/12 digits) and entity types.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies this tool is for TIN-based lookup, distinct from siblings like search_by_name or search_by_ogrn, but provides no explicit when-to-use or when-not-to-use guidance.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Fuzzy-поиск юр.лиц по названию через FTS5.
| Name | Required | Description | Default |
|---|---|---|---|
| limit | No | максимум результатов (1..50). | |
| query | Yes | строка запроса (минимум 2 символа). | |
| only_active | No | фильтровать только записи со статусом 'active'. |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries full burden. It mentions fuzzy search via FTS5 but omits details like matching behavior, result ordering, or handling of misspellings.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Description is a single sentence, concise but lacking structure. It could benefit from additional context without being lengthy.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given no annotations and only input schema provided, the description omits important behavioral traits and result format expectations. Even with an output schema, more context would be helpful.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100% with descriptions for all parameters. The description adds no new meaning beyond 'fuzzy', so baseline 3 applies.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
Description clearly states it is a fuzzy search of legal entities by name using FTS5, which is specific and distinguishes from siblings like search_by_inn and search_by_ogrn.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No explicit guidance on when to use this tool versus alternatives. It implicitly targets name-based searches, but without stating exclusions or context.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Карточка по ОГРН (13 цифр) или ОГРНИП (15 цифр).
| Name | Required | Description | Default |
|---|---|---|---|
| ogrn | Yes |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description must fully disclose behavior. It only states what the tool does (retrieve a card) but omits details about error handling, side effects, or what happens for invalid inputs. This is insufficient for an agent to fully anticipate behavior.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single sentence, which is highly concise and front-loaded. However, it could be structured with separate sentences for clarity, but for a simple tool it is appropriately sized.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the existence of an output schema (not shown), the description may not need to detail return values. It provides the parameter format but lacks any mention of error handling or usage context. For a simple lookup tool, it is minimally adequate.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The schema has 0% description coverage, so the description must compensate. It adds meaningful format constraints: 13 digits for OGRN and 15 digits for OGRNIP, which is valuable beyond the plain string type in the schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states that the tool retrieves a card by OGRN (13 digits) or OGRNIP (15 digits), making the purpose and resource specific. It distinguishes itself from sibling tools like search_by_inn and search_by_name by indicating the identifier type.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description offers no explicit guidance on when to use this tool versus alternatives or when not to use it. The context is only implied by the tool name and description, with no mention of prerequisites or exclusions.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
v0.1.2bulk_cardsget_directorget_foundersget_full_cardpingsearch_by_innsearch_by_namesearch_by_ogrnScored across 8 tools
Each tool targets a distinct query type: bulk export, specific fields (director, founders), full card, health check, and three search methods (INN, name, OGRN). No overlapping purposes.
Most tools use a verb_noun pattern (get_director, get_founders, search_by_inn, etc.). Ping and bulk_cards are minor deviations but still clear.
Eight tools cover a complete set of operations for a business registry: multiple search methods, specific field lookups, bulk export, and health check. No extraneous tools.
Covers essential read operations for a registry: search by various identifiers, retrieval of full cards and specific fields, bulk export. Missing filtering or advanced search (e.g., by region) but acceptable for a focused server.
MCP server for US nursing facility search and ownership lookup (NursingHomeDatabase).
MCP server for Russian books search, details, and recommendation candidates.
MCP server for medicaid-intelligence
Agent-native MCP server over 49M+ US public and government records, privacy-first, always current.
