MarketNow MCP — Free Trust Layer for AI Agents

MarketNow MCP — Free Trust Layer for AI Agents

Verify who you're talking to and what you're about to call — before your agent executes anything. 100% free. No fees. No signup. No API keys.

Remote endpoint (zero install): https://www.marketnow.site/api/mcp — marketnow-mcp v1.13.0, 9 tools, JSON-RPC over streamable HTTP.

curl -s https://www.marketnow.site/api/mcp \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-06-18","capabilities":{},"clientInfo":{"name":"curl","version":"1.0"}}}'

Why

• AI agents connect to MCP servers blind. Tool poisoning, prompt injection via tool descriptions, and scam domains are live attack surfaces (see the OWASP MCP Cheat Sheet).
• Existing answers put a price tag on trust (per-call fees, paywalled ledgers). MarketNow trust tooling is free forever: the endpoint, the credential format, the registry lookups, and the 14 npm packages.
• Everything is fail-closed: if verification can't complete, the answer is "unknown", never "trusted".

The 9 remote tools

30-second setup

Remote (recommended — nothing to install):

{
  "mcpServers": {
    "marketnow": {
      "url": "https://www.marketnow.site/api/mcp"
    }
  }
}

Works with Cursor, Cline, Claude, Gemini CLI, and any MCP 2025-03-26+ client. See mcp.json for a drop-in config and llms-install.md for a guided install.

Local stdio (npm):

npx marketnow-mcp

Client configs

{
  "mcpServers": {
    "marketnow": { "url": "https://www.marketnow.site/api/mcp" }
  }
}

{
  "mcpServers": {
    "marketnow": { "type": "http", "url": "https://www.marketnow.site/api/mcp" }
  }
}

{
  "mcpServers": {
    "marketnow": { "url": "https://www.marketnow.site/api/mcp", "disabled": false, "autoApprove": [] }
  }
}

gemini-extension.json is already in this repo with the gemini-cli-extension topic — the extensions gallery crawler indexes it automatically.

REST API (no MCP needed)

# Scam-check any domain — free, no key
curl -s 'https://www.marketnow.site/api/scam-check?domain=example.com'

# Current CA public key for Agent Trust Card verification
curl -s 'https://www.marketnow.site/api/atc?action=ca-key'

Also live: /api/trust, /api/interceptor, /api/audit-report.json, /api/owasp.json, and the ATC schemas at /atc/unified and /atc/schema-3.0.

npm packages (14, ~6,300 downloads/month)

One-shot installer for the whole stack:

curl -fsSL https://marketnow.site/install.sh | bash

Agent Trust Card (ATC) — the free credential

ATC v3 is an open trust credential for agents: Ed25519-signed, RFC 8785-canonicalized, revocable. Mint one from npm, verify any of the 8 formats through the endpoint. Free. No per-card fee. No ledger rent. The verification pipeline runs 12 stages (signature → canonicalization → expiry → revocation → CA rotation → ...); get the full list via the marketnow_get_pipeline tool.

Facts

• Registry: 68,388 indexed MCP servers (deduped, multi-source) — search it with marketnow_search_skills
• Credential formats supported: 8 (ATC v3, JWT, W3C VC, A2A, EAT-AI, ZTA, MCP Card, X.509)
• Cost: $0 — endpoint, packages, registry search, CA key, revocation checks
• This repo = distribution storefront (docs + manifests for marketplaces). Canonical development: alicelabs-llc/universal-trust-adapter

Manifests in this repo (agent-first)

License

Docs and manifests in this repo: MIT. The npm packages carry their own licenses (AL-1.0 / MNNC-1.0) — see each package page.