Skip to main content
Glama
martinhavel

cz-agents-mcp

by martinhavel
OverviewSchemaRelated ServersScoreDiscussions
TypeScript
Hybrid

get_dd_report

Generate a complete due-diligence report for any Czech IČO, returning company facts, statutory body sanctions checks, and a risk score with triggered red flags.

Instructions

Generate a complete due-diligence report for a Czech IČO. Returns company facts (name, address, legal form, VAT status, bank accounts), statutory body with per-member sanctions check, and a transparent risk score with all triggered red flags.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
icoYesCzech IČO — 7 or 8 digits.
depthNobasic = ARES + sanctions only; full = + ISIR insolvency + virtual-address probe.basic

Implementation Reference

  • packages/dd/src/server.ts:23-38 (registration)
    Tool registration via server.tool('get_dd_report', ...) — defines the tool name, description, input schema (ico string, depth enum with default 'basic'), and the async handler that validates the ICO and calls buildReport.
    server.tool(
  'get_dd_report',
  'Generate a complete due-diligence report for a Czech IČO. Returns company facts (name, address, legal form, VAT status, bank accounts), statutory body with per-member sanctions check, and a transparent risk score with all triggered red flags.',
  {
    ico: z.string().describe('Czech IČO — 7 or 8 digits.'),
    depth: z
      .enum(['basic', 'full'])
      .default('basic')
      .describe('basic = ARES + sanctions only; full = + ISIR insolvency + virtual-address probe.'),
  },
  async ({ ico, depth }) => {
    const clean = validateIcoInput(ico);
    const report = await buildReport(clean, clients, { depth });
    return wrap(JSON.stringify(report, null, 2));
  },
);
  • packages/dd/src/report.ts:30-190 (handler)
    buildReport is the core handler for get_dd_report. It orchestrates ARES data fetching, sanctions screening, insolvency checks, government address detection, prior bankruptcy search, and virtual address detection. Returns a DdReport with company info, VAT details, statutory body, insolvency data, sanctions matches, red flags, and risk score.
    export async function buildReport(
  ico: string,
  clients: DdClients,
  opts: ReportOptions = {},
): Promise<DdReport> {
  const depth = opts.depth ?? 'basic';
  const basicOnly = depth === 'basic';

  const [subject, bankAccounts, vr] = await Promise.all([
    safe(() => clients.ares.getByIco(ico)),
    safe(() => clients.ares.getBankAccounts(ico)),
    safe(() => clients.ares.getVrRecord(ico)),
  ]);

  const { members, mostRecentStatutoryChange } = extractStatutoryMembers(vr);

  const screenedMembers = await screenStatutory(members, clients.sanctions);

  // Govt-address detection on each statutory FO (úřad bydliště = bílý kůň indicator).
  // Cheap heuristic — runs even in basic depth.
  const govtAddrFlags: Array<{ name: string; signal: string; matched_token?: string }> = [];
  for (let i = 0; i < members.length; i++) {
    const m = members[i]!;
    if (!m.is_person) continue;
    const detect = detectGovtAddress(m.address);
    if (detect.is_govt_address) {
      const sm = screenedMembers[i];
      if (sm) {
        sm.registered_at_govt_office = {
          signal: detect.signal as 'marker' | 'known_address',
          matched_token: detect.matched_token,
        };
      }
      govtAddrFlags.push({ name: m.name, signal: detect.signal, matched_token: detect.matched_token });
    }
  }

  // Fáze 2: historical bankrupt-company check per statutory person.
  // For each surname, search ARES for other companies → check ISIR per company.
  // Conservative: only flag when an OTHER company (not the current ICO) had/has insolvency.
  const priorBankruptcyHits: Array<{ name: string; ico: string; company_name?: string; spisova_znacka?: string }> = [];
  if (!basicOnly && clients.isir) {
    await Promise.all(
      members.map(async (m, i) => {
        if (!m.is_person) return;
        const surname = m.surname;
        if (!surname || surname.length < 5) return; // Skip short common surnames
        const otherIcos = await findOtherCompaniesBySurname(clients.ares, surname, ico);
        for (const co of otherIcos.slice(0, 5)) {
          const status = await safe(() => clients.isir!.checkActiveInsolvency(co.ico));
          if (status?.has_active) {
            const sm = screenedMembers[i];
            if (sm) {
              if (!sm.prior_bankrupt_companies) sm.prior_bankrupt_companies = [];
              sm.prior_bankrupt_companies.push({
                ico: co.ico,
                name: co.name,
                spisova_znacka: status.spisova_znacka,
              });
            }
            priorBankruptcyHits.push({ name: m.name, ico: co.ico, company_name: co.name, spisova_znacka: status.spisova_znacka });
          }
        }
      }),
    );
  }

  // Person-level insolvency screen (full depth only) — uses ISIR person search
  // by name + DOB. Skip silently when ISIR client doesn't expose searchPersonInsolvency.
  if (!basicOnly && clients.isir?.searchPersonInsolvency) {
    await Promise.all(
      screenedMembers.map(async (m, i) => {
        if (!m.is_person) return;
        const dob = members[i]?.dob;
        try {
          const hits = await clients.isir!.searchPersonInsolvency!({ name: m.name, dob, onlyActive: true });
          if (hits.length > 0) {
            const top = hits[0]!;
            m.personal_insolvency = {
              spisova_znacka: top.spisova_znacka,
              phase: top.druh_stav_konkursu,
              url: top.url_detail,
            };
          }
        } catch {
          // Network/ISIR error — degrade gracefully, do not fail whole report
        }
      }),
    );
  }

  const companyMatch = screenCompany(ico, subject?.obchodniJmeno, clients.sanctions);

  const insolvency = !basicOnly && clients.isir
    ? await safe(() => clients.isir!.checkActiveInsolvency(ico))
    : null;

  const isVirtualAddress = !basicOnly
    ? await checkVirtualAddress(clients.ares, subject)
    : undefined;

  const flags = evaluateFlags({
    ico,
    subject: subject ?? null,
    vr: vr ?? null,
    vatPayer: !!subject?.dic,
    bankAccountsCount: bankAccounts?.length ?? 0,
    companySanction: companyMatch ?? undefined,
    statutorySanctions: screenedMembers
      .filter((m) => m.sanctions_match)
      .map((m) => ({
        name: m.name,
        match: rebuildSanctionsMatch(m.sanctions_match!),
      })),
    insolvency: insolvency ?? null,
    isVirtualAddress,
    mostRecentStatutoryChange,
    statutoryPersonalInsolvencies: screenedMembers
      .filter((m) => m.personal_insolvency)
      .map((m) => ({ name: m.name, spisova_znacka: m.personal_insolvency!.spisova_znacka })),
    statutoryGovtAddresses: govtAddrFlags,
    statutoryPriorBankruptcies: priorBankruptcyHits,
  });

  return {
    ico,
    retrieved_at: new Date().toISOString(),
    basic_only: basicOnly,
    company: {
      name: subject?.obchodniJmeno,
      legal_form: subject?.pravniForma,
      address: subject?.sidlo?.textovaAdresa,
      registered_on: subject?.datumVzniku,
      dissolved_on: subject?.datumZaniku,
      nace_codes: subject?.czNace,
      found: !!subject,
    },
    vat: {
      is_payer: !!subject?.dic,
      dic: subject?.dic,
      bank_accounts: (bankAccounts ?? []).map((a) => `${a.cisloUctu}/${a.kodBanky}`),
      financial_office: subject?.financniUrad,
    },
    statutory_body: screenedMembers,
    insolvency: insolvency
      ? {
          has_active_proceeding: insolvency.has_active,
          spisova_znacka: insolvency.spisova_znacka,
          started_on: insolvency.started_on,
        }
      : !basicOnly && clients.isir
        ? { has_active_proceeding: false, note: 'No record found' }
        : undefined,
    sanctions: {
      company_match: companyMatch ? toSummary(companyMatch) : undefined,
      any_statutory_match: screenedMembers.some((m) => m.sanctions_match),
    },
    red_flags: flags,
    risk_score: scoreFromFlags(flags),
  };
}
  • packages/dd/src/types.ts:58-97 (schema)
    DdReport interface — the typed return shape for get_dd_report. Contains company info, VAT data, statutory body (StatutoryMember[]), insolvency, sanctions matches, red flags, and risk score.
    export interface DdReport {
  ico: string;
  retrieved_at: string;
  basic_only: boolean;            // true if depth='basic' (skipped ISIR / chain)

  company: {
    name?: string;
    legal_form?: string;
    address?: string;
    registered_on?: string;
    dissolved_on?: string;
    nace_codes?: string[];
    found: boolean;
  };

  vat: {
    is_payer: boolean;
    dic?: string;
    bank_accounts: string[];
    financial_office?: string;
  };

  statutory_body: StatutoryMember[];

  insolvency?: {
    has_active_proceeding: boolean;
    spisova_znacka?: string;
    started_on?: string;
    note?: string;
  };

  sanctions: {
    company_match?: SanctionMatchSummary;
    /** True if any statutory member matched a list (details on each member). */
    any_statutory_match: boolean;
  };

  red_flags: RedFlag[];
  risk_score: { value: number; level: RiskLevel };
}
  • packages/dd/src/score.ts:182-203 (schema)
    scoreFromFlags computes the numeric risk score (0-100) and level (low/medium/high) from the triggered red flags. Any critical flag forces 'high' level.
    export function scoreFromFlags(flags: RedFlag[]): { value: number; level: RiskLevel } {
  const sum = flags.reduce((acc, f) => acc + f.weight, 0);
  const value = Math.min(100, sum);
  // Any critical flag forces 'high' regardless of numeric score — compliance
  // default: insolvency / sanctions / dissolution ARE high risk by definition.
  const hasCritical = flags.some((f) => f.severity === 'critical');
  const level: RiskLevel = hasCritical
    ? 'high'
    : value <= 20
      ? 'low'
      : value <= 50
        ? 'medium'
        : 'high';
  return { value, level };
}

function daysBetween(a: Date, b: Date): number {
  if (Number.isNaN(a.getTime()) || Number.isNaN(b.getTime())) return -1;
  const ms = b.getTime() - a.getTime();
  return Math.floor(ms / (1000 * 60 * 60 * 24));
}
  • packages/dd/src/score.ts:38-180 (helper)
    evaluateFlags generates all RedFlag entries based on company data, sanctions, insolvency, government address detection, and other signals. Each flag has a code, severity, weight, description, source, and evidence.
    export function evaluateFlags(input: ScoreInputs): RedFlag[] {
  const flags: RedFlag[] = [];

  if (input.insolvency?.has_active) {
    flags.push({
      code: 'INSOLVENCY_ACTIVE',
      severity: 'critical',
      weight: 50,
      description: 'Aktivní insolvenční řízení v ISIR.',
      source: 'isir',
      evidence: input.insolvency,
    });
  }

  if (input.companySanction && input.companySanction.confidence >= HIGH_CONFIDENCE_SANCTIONS) {
    flags.push({
      code: 'COMPANY_SANCTIONED',
      severity: 'critical',
      weight: 50,
      description: `Firma na sankčním seznamu (${input.companySanction.entity.source}).`,
      source: `sanctions:${input.companySanction.entity.source}`,
      evidence: input.companySanction,
    });
  }

  for (const s of input.statutorySanctions) {
    if (s.match.confidence >= HIGH_CONFIDENCE_SANCTIONS) {
      flags.push({
        code: 'STATUTORY_SANCTIONED',
        severity: 'critical',
        weight: 50,
        description: `Statutární zástupce na sankčním seznamu: ${s.name} (${s.match.entity.source}).`,
        source: `sanctions:${s.match.entity.source}`,
        evidence: s,
      });
    }
  }

  for (const g of input.statutoryGovtAddresses ?? []) {
    flags.push({
      code: 'STATUTORY_REGISTERED_AT_GOVT_OFFICE',
      severity: 'high',
      weight: 25,
      description: `Statutární osoba ${g.name} má trvalé bydliště evidované na úřadu (signál: ${g.signal}${g.matched_token ? `, ${g.matched_token}` : ''}). Klasický indikátor "bílého koně" — nestabilní bydlení nebo nominální statutář pro shell company.`,
      source: 'ares',
      evidence: g,
    });
  }

  for (const p of input.statutoryPriorBankruptcies ?? []) {
    flags.push({
      code: 'STATUTORY_PRIOR_BANKRUPT_COMPANY',
      severity: 'high',
      weight: 20,
      description: `Statutární osoba ${p.name} je pravděpodobně spojena s firmou ${p.company_name ?? p.ico} (IČO ${p.ico}), která má aktivní insolvenční řízení (${p.spisova_znacka ?? 'spis. zn. neznámá'}). Možný serial-bankrupt founder pattern. Volitelně ověřit přes ARES /historie.`,
      source: 'ares+isir',
      evidence: p,
    });
  }

  for (const p of input.statutoryPersonalInsolvencies ?? []) {
    flags.push({
      code: 'STATUTORY_PERSONAL_INSOLVENCY',
      severity: 'critical',
      weight: 50,
      description: `Statutární osoba v osobní insolvenci: ${p.name} (${p.spisova_znacka}). Dle § 13 ZSVR je nezpůsobilá řídit firmu.`,
      source: 'isir',
      evidence: p,
    });
  }

  if (input.subject?.datumZaniku) {
    flags.push({
      code: 'COMPANY_DISSOLVED',
      severity: 'critical',
      weight: 50,
      description: `Firma zanikla (${input.subject.datumZaniku}).`,
      source: 'ares',
      evidence: { datumZaniku: input.subject.datumZaniku },
    });
  }

  if (input.isVirtualAddress) {
    flags.push({
      code: 'VIRTUAL_ADDRESS',
      severity: 'medium',
      weight: 10,
      description: 'Adresa registrovaná u 50+ firem (pravděpodobně virtuální sídlo).',
      source: 'ares',
      evidence: { address: input.subject?.sidlo?.textovaAdresa },
    });
  }

  if (input.mostRecentStatutoryChange) {
    const days = daysBetween(new Date(input.mostRecentStatutoryChange), new Date());
    if (days >= 0 && days < 30) {
      flags.push({
        code: 'RECENT_STATUTORY_CHANGE',
        severity: 'medium',
        weight: 10,
        description: `Změna ve statutárním orgánu před ${days} dny.`,
        source: 'ares',
        evidence: { datumZapisu: input.mostRecentStatutoryChange },
      });
    }
  }

  if (input.subject?.datumVzniku) {
    const days = daysBetween(new Date(input.subject.datumVzniku), new Date());
    if (days >= 0 && days < 180) {
      flags.push({
        code: 'NEW_COMPANY',
        severity: 'low',
        weight: 5,
        description: `Firma registrována před ${days} dny (mladší 6 měsíců).`,
        source: 'ares',
        evidence: { datumVzniku: input.subject.datumVzniku },
      });
    }
  }

  if (input.vatPayer && input.bankAccountsCount === 0) {
    flags.push({
      code: 'NO_DPH_BANK_ACCOUNT',
      severity: 'low',
      weight: 5,
      description: 'Plátce DPH bez zveřejněného transparentního účtu.',
      source: 'ares',
    });
  }

  if (!input.subject) {
    flags.push({
      code: 'NOT_FOUND_IN_ARES',
      severity: 'high',
      weight: 30,
      description: 'IČO nenalezeno v ARES.',
      source: 'ares',
    });
  }

  return flags;
}

Tool Definition Quality

A3.9/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations are present, so the description bears full responsibility for behavioral disclosure. It omits critical traits: whether the operation is read-only, required permissions, handling of invalid IČOs, or potential side effects. 'Generate' suggests a non-idempotent action, but the actual behavior is unclear.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description consists of two concise sentences. The first sentence front-loads the purpose, and the second lists the report contents. No unnecessary words or redundancy.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given no output schema, the description adequately outlines the return categories. However, it lacks behavioral context (e.g., read-only hint, expected latency, error patterns). The presence of sibling tools could have been leveraged for completeness cues.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Both parameters have schema descriptions, and the tool description adds meaning by explaining the 'depth' enum values ('basic = ARES + sanctions only; full = + ISIR insolvency + virtual-address probe'), which is more detailed than the schema alone.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool generates a 'complete due-diligence report for a Czech IČO' and enumerates the returned data categories (company facts, statutory body, risk score). This distinguishes it from sibling tools 'get_risk_score' and 'get_statutory_chain,' which are more narrow in scope.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description implies the tool is for comprehensive reports but does not explicitly contrast with siblings or state when to use a simpler alternative like 'get_risk_score'. No when-not-to-use guidance is provided.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Install Server

Other Tools

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/martinhavel/cz-agents-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server