MCP-Server zur Überprüfung russischer Vertragspartner (juristische Personen und Einzelunternehmer) über öffentliche Daten des Föderalen Steuerdienstes (FNS): EGRUL/EGRIP, EFRSB, „Transparent Business“, FSSP, KAD.
Bereit zur Einbindung in Claude Desktop, Cursor, Claude Code, Cline und jeden anderen Client, der mit dem Model Context Protocol (MCP) kompatibel ist.
Ein KI-Agent (Claude, Cursor, etc.) weiß normalerweise nichts über russische Vertragspartner: Das EGRUL wird von Suchmaschinen nicht ordnungsgemäß indexiert, Daten im „Transparent Business“ des FNS liegen hinter POST-Anfragen und CAPTCHAs, das EFRSB liefert HTML. Dieser MCP-Server stellt dem Agenten sieben Tools zur Verfügung, mit denen er in einem einzigen Aufruf ein vollständiges Bild erhält:
Wer ist das: Name, Adresse, OKVED, Geschäftsführer.
Status: aktiv, in Liquidation, Insolvenz, liquidiert, Reorganisation.
Ist die Zusammenarbeit sicher: Massenadresse, Massengeschäftsführer, Disqualifikation, Insolvenz, Steuerschulden, Nichtabgabe von Berichten, Vollstreckungsverfahren, Schiedsgerichtsverfahren.
Das Haupt-Tool check_contractor(identifier) akzeptiert eine INN oder OGRN und gibt einen aggregierten Bericht mit einem Urteil (safe_to_proceed / manual_review_required / high_risk_do_not_proceed / impossible_contractor_defunct) sowie eine Liste konkreter Empfehlungen zurück.
Related MCP server: mcp-egrul
pip install atomno-mcp-fns-checkOder über uv / pipx:
uv pip install atomno-mcp-fns-check
# или
pipx install atomno-mcp-fns-checkatomno-mcp-fns-check --version
# → atomno-mcp-fns-check 0.1.1
atomno-mcp-fns-check --help
# → полный список флагов: --transport / --host / --port / --log-levelStandardmäßig startet das Paket als stdio-MCP-Server: Der Agent kommuniziert über stdin/stdout JSON-RPC mit ihm. Direkt aus der Shell können Sie ihn nicht „testen“ – verbinden Sie ihn mit einem MCP-Client. Für Netzwerkszenarien ist das Flag --transport {http,sse,streamable-http} mit --host/--port verfügbar.
Bearbeiten Sie mcp.json (Cursor → Settings → Cursor Settings → MCP):
{
"mcpServers": {
"fns-check": {
"command": "atomno-mcp-fns-check"
}
}
}Starten Sie Cursor neu. Fragen Sie im Chat: „Überprüfe den Vertragspartner INN 7707083893“ – der Agent ruft check_contractor selbstständig auf.
Bearbeiten Sie claude_desktop_config.json (macOS: ~/Library/Application Support/Claude/, Windows: %APPDATA%\Claude\):
{
"mcpServers": {
"fns-check": {
"command": "atomno-mcp-fns-check"
}
}
}Starten Sie Claude Desktop neu.
claude mcp add fns-check atomno-mcp-fns-checkIn cline_mcp_settings.json:
{
"mcpServers": {
"fns-check": {
"command": "atomno-mcp-fns-check",
"disabled": false,
"autoApprove": []
}
}
}Tool | Zweck | Eingabe | Quellen |
| Haupt-Tool. Vollständige Prüfung anhand eines Identifikators + determiniertes Urteil und Empfehlungen |
| alle 5 |
| Basis-Karte EGRUL |
| egrul.nalog.ru |
| Basis-Karte nach OGRN/OGRNIP |
| egrul.nalog.ru |
| Lebensstatus mit Anreicherung |
| EGRUL + EFRSB |
| OKVED-Codes mit Erläuterung |
| EGRUL + OKVED-2 Wörterbuch |
| Aktueller Geschäftsführer (+ Historie gemäß Open Data) |
| EGRUL |
| 8 Risikoprüfungen (4 Basis + 4 erweitert) |
| alle 5 |
Verwendete öffentliche Quellen:
egrul.nalog.ru — EGRUL/EGRIP, Vertragspartner-Karte.
bankrot.fedresurs.ru — EFRSB (Einheitliches föderales Register für Insolvenzinformationen).
pb.nalog.ru — Transparent Business des FNS (Steuerschulden, Nichtabgabe von Berichten).
fssp.gov.ru — Datenbank der Vollstreckungsverfahren des FSSP.
kad.arbitr.ru — Kartothek der Schiedsgerichtsverfahren.
Lokale Auszüge der FNS-Register — Massenadressen, Massengeschäftsführer, disqualifizierte Personen (werden per Skript atomno-mcp-fns-etl aus den Open Data des FNS geladen).
check_contractor{
"identifier": "7707083893",
"identifier_type": "inn",
"inn": "7707083893",
"ogrn": "1027700132195",
"card": {
"name": {"full": "ПАО СБЕРБАНК", "short": "СБЕРБАНК"},
"status": "active",
"address": {"full": "117997, Г.Москва, УЛ. ВАВИЛОВА, Д. 19", "is_mass_address": false},
"director": {"full_name": "Греф Г. О.", "position": "Президент"},
"okved_main": {"code": "64.19", "name": "Денежное посредничество прочее"}
},
"legal_status": {"status": "active", "status_label_ru": "Действующее", "sources_checked": ["egrul", "efrsb"]},
"risks": {"overall_risk_level": "low", "overall_risk_score": 0, "flags": [], "errors": []},
"verdict_action": "safe_to_proceed",
"verdict_reason_ru": "Статус «Действующее», уровень риска — low (score 0/100). Препятствий к заключению сделки по открытым источникам не найдено.",
"recommendations": [
"По открытым источникам препятствий к заключению сделки не обнаружено. Соблюдайте стандартные меры должной осмотрительности (ст. 54.1 НК РФ): копия устава, приказ на руководителя, договор."
],
"sources": {"sources_queried": ["efrsb", "egrul", "fssp", "kad", "pb_fns", "registries"]},
"tier": "open",
"checked_at": "2026-04-24T20:15:00Z"
}EGRUL ist die einzige blocking-Quelle. Wenn sie nicht verfügbar ist, löst check_contractor einen SourceUnavailableError aus (der Agent erhält eine menschenlesbare Nachricht).
Andere Quellen werden nach dem Best-Effort-Prinzip eingebunden: CAPTCHA bei FSSP, Antibot bei KAD, 5xx bei pb.nalog.ru – alles wird in risks.errors[] gesammelt und führt nicht zum Abbruch des Berichts. Das übergeordnete Urteil wird zu manual_review_required.
Alle Einstellungen erfolgen über Umgebungsvariablen. Es sind keine Anmeldedaten erforderlich (die Quellen sind öffentlich).
Variable | Beschreibung | Standardwert |
| Pfad zur SQLite-Cache-Datei für Karten |
|
| Pfad zur SQLite-Datei der Register (Massenadressen/Geschäftsführer/Disqualifikationen) |
|
| TTL der zwischengespeicherten Karten in Stunden |
|
| HTTP-Timeout in Sekunden |
|
| User-Agent des HTTP-Clients |
|
| Logging-Level (DEBUG/INFO/WARNING/ERROR) |
|
Vorlage: .env.example.
Die Register für Massenadressen / Geschäftsführer / disqualifizierte Personen sind CSV/XML-Exporte der Open Data des FNS. Das Paket enthält einen integrierten Mini-Seed (registries_seed.json, synthetische Testdatensätze) – dies reicht aus, damit die Tools „out of the box“ funktionieren und Flags für Test-INNs anzeigen.
Für Produktionsprüfungen aktualisieren Sie die Register mit vollständigen Auszügen über die CLI atomno-mcp-fns-etl:
atomno-mcp-fns-etl --registry mass_addresses --source ./fns_open_data/ulm.csv --commit
atomno-mcp-fns-etl --registry mass_directors --source ./fns_open_data/uchredt.csv --commit
atomno-mcp-fns-etl --registry disqualified --source ./fns_open_data/disqualified.csv --commitOpen Data Quellen:
mass_addresses → nalog.gov.ru/opendata/7707329152-massreg/
mass_directors → nalog.gov.ru/opendata/7707329152-massuchredt/
disqualified → service.nalog.ru/disqualified.do
Standardmäßig arbeitet die CLI im --dry-run-Modus (parst und druckt ein Sample); für das Schreiben ist ein explizites --commit erforderlich. Meta-Felder <registry>.last_etl, <registry>.last_etl_source, <registry>.last_etl_count werden automatisch gespeichert – nutzen Sie diese für das Cron-Monitoring der Datenaktualität.
git clone https://github.com/atomno-labs/mcp-fns-check
cd mcp-fns-check
python -m venv .venv
source .venv/bin/activate # Linux/macOS
# .venv/Scripts/activate # Windows
pip install -e ".[dev]"
pytest -v --cov=src/atomno_mcp_fns_checkExterne APIs werden in Tests niemals direkt aufgerufen – nur über respx (Mocking von httpx) + lokale Fixtures in tests/fixtures/.
Keine Historie für Geschäftsführer – Das FNS liefert keine Historie der Wechsel über die Search-API; eine vollständige Historie wird nach dem Laden des Open Data-Auszugs des EGRUL verfügbar sein (geplant für v0.5+).
FSSP / KAD blockieren manchmal durch CAPTCHA / Antibot. In diesem Fall fällt die Prüfung in errors[] und das Gesamturteil wird zu manual_review_required.
Transparent Business liefert nur den Fakt („hat Schulden“ / „keine Berichte“), ohne Betrag. Der Betrag muss bei der IFNS angefragt werden.
Pro-Tier (gehostetes Backend in atomno-mcp-fns-check-server – geschlossenes Backend) beseitigt diese Einschränkungen durch: 24h Redis-Cache, Proxy-Rotation zur Umgehung von CAPTCHAs, vollständiger Open Data-Auszug des EGRUL, Batch-Prüfungen bis zu 100 INNs, KI-Zusammenfassung via LLM. Das Backend selbst ist nicht veröffentlicht.
Alle Quellen sind öffentlich zugängliche Daten des FNS und verbundener Register. Die Nutzung ist gemäß 149-FZ „Über Information“ legal.
Juristische Personen und Einzelunternehmer fallen nicht unter das 152-FZ (Über personenbezogene Daten).
Die Namen natürlicher Personen als Geschäftsführer werden vom FNS im EGRUL offen veröffentlicht; in ausgehenden Antworten wird die INN der natürlichen Person maskiert (Format XXX*****YY).
Keine Schreibvorgänge in irgendeine externe API.
Keine Anmeldedaten / Token erforderlich – die Quellen sind vollständig öffentlich.
Der Dienst ist ein Aggregator und eine benutzerfreundliche Schnittstelle für öffentliche Daten des FNS. Er ist nicht mit dem FNS Russlands, dem EFRSB, KAD oder FSSP verbunden. Die Nutzung erfolgt auf eigenes Risiko.
Die Informationen in den Antworten des Dienstes ersetzen keine vollständige rechtliche oder finanzielle Bewertung. Die Entscheidung über den Abschluss eines Vertrags mit einem Vertragspartner treffen Sie.
MIT – siehe LICENSE.
GitHub: atomno-labs/mcp-fns-check
Weitere MCP-Server unter der Marke atomno: Katalog atomno.ru (bald verfügbar)
MCP-Spezifikation: modelcontextprotocol.io
Главный тул: полная проверка контрагента по одному идентификатору.
Принимает ИНН (10 цифр — юр.лицо, 12 — ИП/физлицо) или ОГРН (13 цифр — юр.лицо, 15 — ИП). Тип определяется строго по длине и контрольной цифре.
| Name | Required | Description | Default |
|---|---|---|---|
| identifier | Yes | ИНН (10/12 цифр) или ОГРН/ОГРНИП (13/15 цифр). Пример: '7707083893' (Сбербанк) или '1027700132195'. | |
| include_extended_risks | No | Запускать ли 4 расширенные проверки (Прозрачный бизнес, ФССП, КАД). По умолчанию True. | |
| lawsuits_threshold_rub | No | Порог фильтрации арбитражных дел в рублях. По умолчанию 1 000 000 ₽ — отсекает мелкие споры. |
| 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 bears full responsibility. It discloses identifier validation logic and parameter defaults (extended risks, lawsuit threshold), but does not elaborate on what the 'full check' entails, side effects, or authentication requirements. The output schema exists but is not referenced.
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 concise (three sentences) with the main purpose front-loaded. Every sentence adds critical information: identifier formats, validation, and parameter defaults. No redundant or vague phrasing.
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 output schema, return values are covered. The description sufficiently explains input validation and parameter behavior for a 3-param tool. It does not mention prerequisites (e.g., API keys) or how this tool relates to siblings beyond being the 'main' one, but overall it is adequate for agent invocation.
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?
With 100% schema coverage, the description adds significant value beyond the schema. It explains identifier format and validation rules, specifies what extended risks include (4 checks: transparent business, FSSP, KAD), and clarifies the lawsuit threshold filters small disputes. Default values are contextualized.
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 full counterparty check using a single identifier (INN or OGRN). It specifies validation criteria (digit length, control digit) and implicitly differentiates from sibling tools (check_inn, check_ogrn) that only handle one 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 advises when to use this tool (for a full check) and provides input validation rules. It does not explicitly state when to use alternatives or when not to use this tool, but the context of being the 'main tool' implies it is the default for comprehensive checks.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Агрегированный риск-чек по контрагенту (8 проверок, SPEC §3.5).
Базовые 4 проверки (всегда):
mass_address — массовый юр.адрес ФНС;
mass_director — массовый руководитель (≥10 действующих компаний);
disqualified_director — руководитель в реестре дисквалифицированных;
bankruptcy_records — активные дела в ЕФРСБ.
Расширенные 4 проверки (include_extended=True):
tax_debts — индикатор задолженности по налогам (Прозрачный бизнес);
no_reporting — нет налоговой отчётности > 1 года (Прозрачный бизнес);
enforcement_proceedings — открытые исп.производства (Банк данных ФССП);
active_lawsuits — активные арбитражные дела где контрагент ответчик
(КАД), фильтр по сумме иска через lawsuits_threshold_rub.
Любая внешняя проверка может вернуть ошибку (CAPTCHA, 5xx, timeout) — она
попадёт в errors[] и не сорвёт остальные проверки.
| Name | Required | Description | Default |
|---|---|---|---|
| inn | Yes | ИНН контрагента (10 или 12 цифр). | |
| include_extended | No | запускать ли расширенные 4 проверки. По умолчанию True. | |
| lawsuits_threshold_rub | No | порог фильтра арбитражных дел в рублях. По умолчанию 1 000 000 ₽ (отсечь мелкие споры). |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations, so description carries full burden. Discloses possible errors (CAPTCHA, 5xx, timeout) and that errors are collected without failing other checks. Reveals behavioral traits beyond annotations.
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?
Well-structured with bullet points for readability, but slightly lengthy. Front-loaded with purpose. Could be slightly more concise.
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?
Covers purpose, parameter details, error handling, and references SPEC. Output schema exists, so return values are handled. Adequate for a tool with 8 sub-checks and error handling.
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?
Input schema has 100% coverage, but description adds value by listing the specific checks controlled by include_extended and explaining the lawsuit filter threshold.
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?
Clearly states it performs a aggregated risk check with 8 checks, referencing SPEC §3.5. Differentiates from siblings like check_contractor, check_inn by being comprehensive.
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?
Explains basic vs extended checks with include_extended, and describes error handling. Does not explicitly contrast with sibling tools, but context is clear when to use.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Полная карточка контрагента по ИНН (10 цифр — юр.лицо, 12 — ИП).
| Name | Required | Description | Default |
|---|---|---|---|
| inn | Yes | ИНН контрагента. Пример: '7707083893' (Сбербанк). | |
| include_extended | No | зарезервировано для S2.x (Прозрачный бизнес). |
| 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 describes the output as a 'full card' but does not disclose behavioral traits such as auth requirements, rate limits, or side effects. The description adds minimal behavioral context beyond the input format.
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, well-structured sentence that front-loads the main purpose. Every word adds value, with no filler. It is appropriately concise.
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 that an output schema exists, the description does not need to detail return values. However, the description is terse and does not specify what the 'full card' includes, leaving some ambiguity. It adequately covers the basic purpose but could be more comprehensive.
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 description coverage is 100%, so baseline is 3. The description adds meaningful context for the 'inn' parameter by specifying the digit length for legal entities vs IPs. For 'include_extended', it clarifies it is reserved for future use, adding value beyond 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 the tool returns a full counterparty card by INN, and specifies the length distinction (10 digits for legal entities, 12 for individual entrepreneurs). This is a specific verb and resource, and it helps distinguish from sibling tools like check_ogrn (which checks 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?
The description implies usage for obtaining a counterparty card based on INN but does not explicitly state when to use this tool versus alternatives like check_contractor or check_for_red_flags. No when-not conditions or alternative recommendations are provided.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Карточка контрагента по ОГРН (13 цифр) или ОГРНИП (15 цифр).
| Name | Required | Description | Default |
|---|---|---|---|
| ogrn | Yes | ||
| include_extended | No |
| 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 the full burden. It does not disclose any behavioral traits, such as whether the operation is read-only, requires authentication, or has rate limits. The description only states what the tool does, not its side effects or safety profile.
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?
Single sentence that front-loads the purpose. No wasted words, but could be slightly more informative without losing conciseness.
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 output schema exists, return values need not be described. However, the description is minimal; it lacks context about the optional parameter and does not clarify the scope of 'counterparty card' relative to sibling tools.
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 0%, so the description must compensate. It adds meaning to the 'ogrn' parameter by specifying digit length (13 or 15). However, it does not mention the 'include_extended' parameter, leaving it unexplained.
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 it returns a counterparty card by OGRN/OGRNIP, distinguishing it from siblings like check_inn which uses a different identifier. However, it doesn't elaborate on what the card contains, preventing a 5.
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 like check_inn or check_contractor. The description only implies usage when an OGRN is available.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
История смены руководителей и учредителей по ИНН.
На стадии S3 возвращает только текущего руководителя из egrul.nalog.ru плюс предупреждение о неполноте данных. Полная история (с датами назначения/увольнения) появится на стадии S5 после загрузки Open Data slice ЕГРЮЛ.
| Name | Required | Description | Default |
|---|---|---|---|
| inn | Yes | ИНН контрагента (10 или 12 цифр). | |
| depth_years | No | на S3 не используется (история пока без дат). | |
| include_founders | No | на S3 не используется (ЕГРЮЛ-search не отдаёт учредителей). |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided, but description discloses important behavioral traits: returns only current director at S3 with data incompleteness warning, and full history only at S5. Details about parameter non-usage at S3 add transparency. Doesn't mention auth or rate limits, but given no annotations, this is good.
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?
Three sentences with front-loaded purpose. No wasted words; each sentence provides distinct, necessary information (purpose, stage behavior, parameter notes).
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 3 parameters and existing output schema, the description covers stage-dependent behavior and parameter constraints. It doesn't detail output schema, but that's covered by the output schema itself. Sufficient for the 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?
Schema coverage is 100%, and the description adds value by noting that depth_years and include_founders are not effective at S3. This clarifies parameter behavior beyond the schema 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?
Clearly states 'История смены руководителей и учредителей по ИНН' (History of changes of directors and founders by TIN), specifying the verb (get history) and resource (by TIN). This distinguishes it from sibling tools like check_inn or get_legal_status.
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?
Provides stage-specific behavior: at S3 returns only current director with a warning, full history at S5. Also explains that depth_years and include_founders are not used at S3. While it doesn't explicitly list when not to use or compare to siblings, it gives clear usage context.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Текущий жизненный статус: active / liquidating / bankruptcy / liquidated / reorganizing / excluded_inactive.
Минимум один из inn или ogrn обязателен.
| 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 must disclose behavioral traits. It lists possible output values, which is helpful, but it does not mention whether the tool is read-only, requires authentication, or has any side effects. There is no indication of error behavior if required parameters are missing.
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 with two sentences. The first sentence presents the output values upfront, and the second provides a critical usage requirement. No extraneous information is included.
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?
While the output schema exists (context signals indicate it), the description does not leverage or reference it. It explains the return values but omits context about error handling, prerequisites (e.g., authentication), or how this tool fits into a workflow alongside siblings. The missing enforcement of the required parameter in the schema adds ambiguity.
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 description coverage is 0%, so the description must compensate. It only adds that at least one parameter is required, without explaining what 'inn' and 'ogrn' represent, their formats, or constraints. This minimal addition does not adequately compensate for the lack of schema 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 tool name 'get_legal_status' combined with the description clearly indicates that the tool retrieves the current legal status from a list of enumerated values (active, liquidating, etc.). It is distinguishable from sibling tools like 'check_inn' and 'check_ogrn' which likely validate identifiers, and 'get_directors_history' which retrieves different data. However, the description does not explicitly contrast with these siblings.
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 states that at least one of 'inn' or 'ogrn' is required, which provides a key usage constraint. However, it does not explain when to use this tool versus alternatives such as 'check_inn' or 'check_for_red_flags'. The guideline is present but incomplete.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Перечень кодов ОКВЭД основной + дополнительные с расшифровкой по справочнику ОКВЭД-2.
| Name | Required | Description | Default |
|---|---|---|---|
| inn | No | ||
| ogrn | No | ||
| include_history | No |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, and the description does not disclose behavioral traits such as read-only nature, authorization needs, or rate limits. For a tool with no annotations, more behavioral context is needed.
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 short sentence, which is concise but omits essential information about inputs and usage. It is underspecified for a tool with multiple optional parameters.
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 lack of param documentation and annotations, the description is insufficient. The output schema exists but the description does not cover input semantics or invocation requirements.
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 description coverage is 0%, and the description does not explain the purpose of any parameter (inn, ogrn, include_history). The agent cannot infer how to specify input correctly.
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 list of OKVED codes with interpretations. However, it does not specify that the user must provide either INN or OGRN as input, which is crucial for clarity.
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 guidance on when to use this tool versus siblings like check_inn or get_directors_history. No alternatives or exclusions are mentioned.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Диагностический тул: проверяет, что сервер запущен и доступен.
| 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 provided, but the description discloses that the tool performs a read-only check without side effects. It is adequate for a ping tool; additional details like return type are covered by the output schema.
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?
Single sentence with no wasted words. Front-loaded purpose, efficient and structured.
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 tool's simplicity (no parameters, no side effects), the description is complete. Output schema exists to explain return values, so no further description needed.
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?
No parameters exist, so baseline score of 4 applies. Description adds no parameter info, but none is needed.
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 a diagnostic tool that checks server availability, using specific verb 'проверяет' and resource 'сервер'. It distinguishes well from sibling tools which focus on business data.
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 verifying server connectivity, which is distinct from siblings. However, it does not explicitly state when to use or when not to use, but the context is clear enough for a simple diagnostic tool.
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.
v0.1.1check_contractorcheck_for_red_flagscheck_inncheck_ogrnget_directors_historyget_legal_statusget_okvedspingScored across 8 tools
Each tool targets a distinct aspect of contractor verification: comprehensive checks by identifier, risk flags, specific identifier lookups, director history, legal status, OKVED codes, and a diagnostic ping. No overlap in purpose.
Tools follow a consistent snake_case pattern with clear verb prefixes (check_, get_, ping). Minor deviation with 'check_for_red_flags' using 'for', but overall pattern is predictable.
8 tools is well-scoped for the domain of Russian contractor verification, covering all essential operations without being excessive or too sparse.
Covers core lifecycle: identifier validation, comprehensive checks, risk assessment, director history, legal status, and OKVED codes. Minor gaps like batch operations, but the main workflows are complete.
Russian company lookup (EGRUL/INN), Cyrillic search, RU page to Markdown. Pay per call in USDC.
Remote MCP server to enrich company profiles with structured B2B data and confidence scores.
MCP server for Russian books search, details, and recommendation candidates.
MCP server offering regulator-sourced legitimacy checks on investment entities by name or URL.
