Argus MCP — Multi-Agent Security Oracle

Ask any AI to audit any smart contract. Argus runs three independent AI agents — Agent-α (DeepSeek-V3) on contract logic, Agent-β (Claude) on tokenomics, Agent-γ (rules engine) on deterministic checks — and returns a consensus verdict (SAFE / RISKY / SCAM) with per-agent confidence and reasoning. Every scan settles on-chain via Circle x402.

⚡ Powered by the Argus Arc testnet oracle · argusarc.dev · Circle Alliance Program

Tools

| Tool | Description | |---|---| | scan_contract | Full 3-agent consensus audit of any EVM contract address | | recent_scans | Latest scans from the Argus oracle | | agent_elo | Live ELO reputation leaderboard for the 3 agents | | pool_stats | Platform stats — assigned user wallets, available wallets |

Quick start with npx

npx @ogxavier/mcp

No build, no install — just run it. Point your MCP client at this command.

Claude Desktop

Add to claude_desktop_config.json (Claude → Settings → Developer → Edit Config):

{
  "mcpServers": {
    "argus": {
      "command": "npx",
      "args": ["-y", "@ogxavier/mcp"]
    }
  }
}

Restart Claude, then ask:

"Is 0xA0b86991c6218b36c1d19D4a2e9Eb0cE3606eB48 safe?" "Scan 0x675E25a39568713144905A946F8d9F74d3a7Aa79 for scams"

Claude will call scan_contract and cite the 3-agent consensus verdict.

ChatGPT (MCP Connector)

1. Open ChatGPT → your profile → Connectors → MCP
2. Add a connector of type stdio
3. Command: npx, Arguments: -y @ogxavier/mcp
4. Ask: "Use Argus to scan 0x... and tell me if it's a scam"

Works the same in Cursor, Windsurf, and any MCP-compatible client.

Configuration

| Env var | Default | Purpose | |---|---|---| | ARGUS_API_URL | https://argus-web-backend-production.up.railway.app | Argus backend endpoint |

{
  "mcpServers": {
    "argus": {
      "command": "npx",
      "args": ["-y", "@ogxavier/mcp"],
      "env": { "ARGUS_API_URL": "https://your-backend.example.com" }
    }
  }
}

Local development

npm install
npm run build        # compile to dist/
npm test             # smoke test against live backend
npm run dev          # run from source via tsx

How it works

sequenceDiagram
    participant U as User (Claude/ChatGPT)
    participant M as Argus MCP (stdio)
    participant B as Argus Backend
    U->>M: "is 0x... a scam?"
    M->>B: POST /debug/scan
    B->>B: 3 agents analyze in parallel
    B->>B: Consensus + on-chain settlement
    B-->>M: verdict + per-agent reasoning
    M-->>U: formatted audit report

License

MIT