@agentresources/mcp

@agentresources/mcp

Operational Model Context Protocol server for Agent Resources — plus the ar CLI (ar doctor, ar skills get, ar mcp).

Wraps the operational AR gateway routes (Trust Card lookup, signed telemetry, memory verbs, signing-key management) behind a stdio-transport MCP server. Drop into Claude Code, Cursor, VS Code, Windsurf, or any MCP host.

Install

npm install -g @agentresources/mcp     # gives you `ar` + `agent-resources-mcp`
# or, no install:
npx @agentresources/mcp                 # starts the MCP server on stdio

Python users: pip install agent-resources-mcp ships the same MCP surface (CLI parity with ar is a follow-up).

What it exposes (MCP tools)

Tool	Purpose
trust_card_lookup	Public Trust Card fetch + optional anchor proof
telemetry_ingest	Submit (signed) telemetry spans
memory	remember / recall / forget / improve over the unified telemetry-as-memory store
signing_keys	List / register / revoke an agent's signing keys

All authed tools require AR_SESSION_TOKEN in the environment. Get one with walletLogin from @agentresources/sdk.

CLI

ar doctor                  # six-check self-diagnostic (gateway, LiteLLM, signer key, dry-run, secrets, vulns)
ar doctor --offline --quick --json
ar skills get agent-resources
ar skills get verify-trust-card
ar mcp                     # equivalent to `agent-resources-mcp`
ar version

ar doctor checks

• Gateway reachable (/api/health)
• LiteLLM Tier-1 responding (skipped when offline / no token)
• Trust Card signer key configured (env presence only — never printed)
• Trust Card issuance dry-run (/.well-known/trust-card/0x000...0)
• Secrets scan of cwd (no plaintext private keys / API tokens / Slack tokens / GitHub PATs)
• Dependency vulnerability scan (pnpm audit → npm audit fallback)

Exit code = number of failed checks. Warns do not contribute to exit. Use --json for machine-readable output.

MCP host integration (Claude Code example)

.mcp.json at your project root:

{
  "mcpServers": {
    "agent-resources": {
      "command": "npx",
      "args": ["@agentresources/mcp"],
      "env": {
        "AR_API_URL": "https://api.agentresources.xyz",
        "AR_SESSION_TOKEN": "<your wallet session token>"
      }
    }
  }
}

Cursor: .cursor/mcp.json. VS Code: .vscode/mcp.json. Windsurf: .windsurf/mcp.json. OpenCode: opencode.json. OpenClaw: .claw.json.

Embedding the server

import { createMcpServer } from "@agentresources/mcp";
import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js";

const server = createMcpServer({
  apiUrl: "https://api.agentresources.xyz",
  token: process.env.AR_SESSION_TOKEN,
});
await server.connect(new StdioServerTransport());

License

MIT. See LICENSE.

Skills are free

AR skills are free, portable capability packs. Paid lifecycle products — Trust Card issuance, KYA, scans, on-chain attestations — are billable services and are never packaged as skills.

Crypto-scam notice

Agent Resources has not issued, sold, or pre-sold any token. There is no AR token, presale, airdrop, points programme, or staking programme. Anyone offering AR tokens or asking you to send funds to participate in AR is running a scam. See PHILOSOPHY.md for the full disclaimer and agentresources.xyz/legal/terms#10 for the legal notice.