Back to Directory/Developer Tools

io.github.makosdDavid/package-risk

Package risk checks: maintenance, licence, advisories. Paid per call in USDC, no signup.

Developer ToolsJavaScriptv1.0.5

package-risk MCP connector

MCP tools for checking a package's maintenance status, licence, and security advisories before you depend on it - package_risk, package_licence, package_advisories. Paid per call in USDC on Base mainnet via x402.

No subscription, no API key. You pay from your own wallet, per call, only for what you use.

What this is (and isn't)

This is a thin client. The actual service is a stateless HTTP API at x402-package-risk.x402-package-risk.workers.dev. This connector never sees, holds, or forwards anyone else's funds - it only ever spends the wallet key you configure below, and only when you call one of its tools.

Setup

You need an EVM wallet with a small amount of USDC on Base mainnet (calls cost $0.005-$0.01 each). Never use a wallet holding significant funds for an automated agent key - keep this one funded lightly.

Add to your MCP client config (Claude Desktop, Claude Code, Cursor, etc.):

{
  "mcpServers": {
    "package-risk": {
      "command": "npx",
      "args": ["-y", "@makosdav/package-risk-mcp"],
      "env": {
        "EVM_PRIVATE_KEY": "0xyour-private-key-here"
      }
    }
  }
}

Tools

ToolPriceWhat it returns
package_risk$0.01Full verdict: maintenance, licence, advisories, deprecation
package_licence$0.005Licence expression and closed-source safety
package_advisories$0.005Open OSV advisories for the resolved version

All three take system (npm/pypi/go/maven/cargo/nuget), name, and an optional version.

How payment works

  1. Your agent calls a tool.
  2. This connector requests the resource; the server replies 402 Payment Required.
  3. @x402/fetch builds and signs a payment authorisation with your key.
  4. The request retries with payment attached; the server verifies via Coinbase CDP, returns the result, and settles on-chain.

No approval prompt happens here beyond what your MCP client itself asks for - if you want per-call confirmation, configure that in your agent framework, not here.

Installation

Source-derived launch command. Check the maintainer’s required arguments and credentials before running:

bash
npx -y @makosdav/package-risk-mcp

Set up in your AI client

Merge this template into ~/Library/Application Support/Claude/claude_desktop_config.json. Keep existing servers. Add any arguments, credentials, and permissions required by the maintainer; this template has not been install-tested.

json
{
  "mcpServers": {
    "io-github-makosddavid-package-risk": {
      "command": "npx",
      "args": [
        "-y",
        "@makosdav/package-risk-mcp"
      ]
    }
  }
}

Restart Claude Desktop completely for changes to take effect. Confirm the server appears connected in the client’s tool list, then try a read-only example from its documentation.

Claude Desktop setup reference

Package

@makosdav/package-risk-mcpnpm

Compatible MCP Clients

io.github.makosdDavid/package-risk works with any MCP-compatible client. Copy the config snippet from the Configuration section above and add it to the file shown for your client, then restart the application.

  • Claude Desktop~/Library/Application Support/Claude/claude_desktop_config.jsonRestart Claude Desktop completely for changes to take effect.
  • Cursor~/.cursor/mcp.jsonRestart Cursor for changes to take effect.
  • VS Code.vscode/mcp.jsonReload VS Code window for changes to take effect.
  • Windsurf~/.codeium/windsurf/mcp_config.jsonRestart Windsurf for changes to take effect.
  • Claude Code.mcp.jsonSave at the project root, then start Claude Code in that project and review the MCP server approval prompt. Keep real credentials out of shared files.

Learn More