Skip to content
Limitguard logo

Limitguard MCP server

Which MCP servers are Limitguard's own, from the official registry and from asking its own domain, whether they answer, and what they offer. A program checked each of these, nobody copied them from a product page.

read 2026-10-21

← Back to Limitguard

Whose server it is
  • Registry nameai.limitguard.api/limitguardverified in the MCP registry

The registry name sits under limitguard.ai. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.

When we asked itchecked 2026-10-21
  • Stateanswerschecked 2026-10-21
  • Tools10
  • Server calls itselflimitguard
  • Server version0.1.0
  • Protocol version2025-06-18

On 2026-10-21 the server answered initialize and tools/list without a sign-in. First answer in 691 ms.

Should an agent connect itNot proven either way yet
  • Whether its tools do the product's job has not been judged yet.

Once connected

  • Parameters are untyped or undescribed in 10 of 10 tools: check arguments before calling.

Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.

Measured dailydaily checks: 12, the first on 2026-10-21
  • INITE MCP score83 of 100answers 100 · definitions 83 · context 96 · use 8 · connects 100
  • Use1 GitHub stars
  • Answered100% of days12 of 12 daily checks
  • Median first answer717 ms
  • Space in an agent's context≈ 3,434 tokensthe definitions of all its tools (10), which an agent carries on every turn
  • Tool definitionsB · 83 of 100
  • Typed results10 of 10 toolsdeclared outputSchema: results come typed, read once
  • Parameters without a description: in 10 of 10

What an agent can do with it

  • not sure · check Dutch and Belgian lead companies before client delivery
  • not sure · remove leads whose company is not in the register

Each of the product's use cases, put to Jev against this server's tool list: can an agent do it with these tools?

Worth a look

None of the five warning signs in these definitions.

A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (25%), doing the product's job (25%), definitions (20%), context (10%), use (10%) and how readily a client connects (10%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →

Tools it offers10 tools · read 2026-10-21
  • check_entity

    Company check via the Limitguard API. Company check on a business: sanctions screening and country risk, the Dutch KVK register for an NL company, and the website domain when you send it. Returns trust_score (0-100, 100 = best)…

  • get_trust_score

    Look up your own most recent trust score for an entity you checked before, from your stored checks: score, level, when, which product, trend and how many checks are on record. Runs no new check and calls no data source. Free ($0). Returns…

  • verify_wallet

    Screen a wallet: OFAC SDN address match, on-chain signals (contract check, native and USDC balance, transaction count, first seen on Base) and named risk rules with up to 3 advice items. On Base, also reports any ERC-8004 agent the wallet…

  • get_risk_score

    Quick risk score from sanctions screening and country risk only. Faster than check_entity: no register or domain lookup. Use when you only need basic risk evaluation. Returns risk_score (0-100, 100 = riskiest), risk_level (high means high…

  • get_compliance_report

    Per-entity report built from one real check. Registry identity, sanctions and PEP screens, domain signals, risk score with the rules that fired, correlations, every finding as a ranked action, a per-source status table (ok / unavailable /…

  • sanctions_preview

    Free yes/no sanctions preview against the local OFAC SDN, EU and UN lists. Returns possible_match, lists_checked, list_dates and list_note only, never an entry. 10 per caller per UTC day, shared with POST /v1/sanctions/preview, and key…

  • wallet_sanctions_preview

    Free yes/no check of a wallet address against the OFAC SDN list. Returns wallet_address (canonical), chain_id, sanctioned, lists, as_of (the list's publication date), status (ok or stale), source and upgrade, which names the full wallet…

  • sanctions_screen

    Sanctions screen against the local OFAC SDN, EU and UN lists: matched entries. Each match carries its list, entry id, entry type, programmes, countries, listing date and match score (exact normalised or word-order-insensitive name match…

  • check_agent_wallet

    Check a counterparty agent's EVM wallet in one call: OFAC SDN digital-currency address list match, Base USDC and ETH balance, ERC-8004 identity registration (and, given an agent id, that agent's owner and payment wallet) and open ERC-8004…

  • verify_lead

    Verify a NL/BE sales lead against the registers in one call: real and active, VAT, mail server, IBAN, sanctions; a 0-100 lead score. $0.10 by x402, $0.27 with an API key. Returns verdict (real_active, real_inactive, not_found, ambiguous…

How to connect

Add the server to your MCP client config. Claude Desktop, Cursor and most other clients read this shape:

{
  "mcpServers": {
    "limitguard": {
      "type": "http",
      "url": "https://api.limitguard.ai/mcp"
    }
  }
}

Add it

Claude Code, in a terminal

claude mcp add --transport http limitguard https://api.limitguard.ai/mcp

Add to CursorAdd to VS Code

claude.ai, ChatGPT and other clients: add https://api.limitguard.ai/mcp as a custom connector in the settings.

Each line is something a program found when it went looking: the registry entry, the server's own answer, a spec that parses. A missing line means we did not find it, not that it does not exist.