返回 Skill 列表
extension
分类: 其它无需 API Key

ScopeBlind protect-mcp

MCP安全网关。为任意MCP服务器提供每工具策略控制、Ed25519签名决策收据和人工审批门禁。影子模式记录所有操作...

person作者: tomjwxfhubclawhub

protect-mcp — MCP Security Gateway

What This Skill Does

Wraps any MCP server as a transparent stdio proxy with per-tool security policies and cryptographic audit trail. Every tool call decision is logged and optionally Ed25519-signed.

Quick Start

# Shadow mode — log everything, block nothing
npx protect-mcp -- node your-server.js

# Enforce mode — apply per-tool policies
npx protect-mcp --policy policy.json --enforce -- node your-server.js

# Initialize signing (generates Ed25519 keypair)
npx protect-mcp init

Policy Example

{
  "tools": {
    "db_write": { "decision": "deny" },
    "file_read": { "decision": "allow", "rateLimit": { "maxCalls": 30, "windowSecs": 60 } },
    "deploy": { "decision": "require_approval" }
  }
}

Pre-built Policy Packs

protect-mcp ships CVE-anchored policy packs:

# List available policies
npx protect-mcp policies

# Apply the Clinejection prevention policy
npx protect-mcp --policy clinejection --enforce -- node your-server.js

Verify Receipts

Receipts are independently verifiable offline — no ScopeBlind dependency:

npx @veritasacta/verify receipt.json
npx @veritasacta/verify --self-test

OWASP MCP Top 10 Coverage

| Risk | Control | |------|---------| | MCP-01 Rug Pulls | Signed tool manifests; policy pins allowed tools | | MCP-03 Tool Poisoning | Per-tool allow/deny/rate-limit policies | | MCP-04 Tool Arg Injection | Argument inspection + approval gates | | MCP-07 Auth/AuthZ | Trust-tier gating | | MCP-08 Logging & Audit | Ed25519-signed receipts — verifiable offline | | MCP-09 Excessive Agency | Shadow mode reveals actual tool usage |

Links

  • npm: https://npmjs.com/package/protect-mcp
  • IETF Draft: https://datatracker.ietf.org/doc/draft-farley-acta-signed-receipts/
  • Docs: https://scopeblind.com/docs/protect-mcp
  • OWASP Mapping: https://scopeblind.com/docs/owasp