ShadowRun
ShadowRun
一个面向 MCP 服务器的本地优先 stdio 代理。它位于 AI 代理(Claude Code、Cursor 等)和真实 MCP 服务器之间,在破坏性工具调用(写入、删除、SQL 变更等)执行之前将其拦截,并在内存中保留,直到你从本地仪表板提交或丢弃它们。
与天真的“假成功”拦截器不同,ShadowRun 在拦截的调用被解决之前不会回复代理。伪造响应(例如针对 INSERT ... RETURNING id)会让代理基于尚不存在的数据行动,并使其信念状态与现实产生分歧——因此拦截的调用只是等待(带有超时)而不是。
范围(v0.1)
这有意保持狭窄:一个面向本地 MCP 服务器(Postgres、文件系统等)的单开发者 CLI,与 Claude Code / Cursor 一起使用。它不是团队策略执行网关,没有身份验证,也没有跨重启的持久化——在依赖它处理超出自己机器范围的事情之前,请参阅下面的限制。
Related MCP server: truecopy
安装与运行
npm install
npm run build将你的代理配置指向代理而不是真实服务器:
{
"mcpServers": {
"postgres": {
"command": "node",
"args": [
"/path/to/shadowrun-mcp/dist/index.js",
"npx", "-y", "@modelcontextprotocol/server-postgres",
"postgresql://localhost:5432/devdb"
]
}
}
}打开 http://127.0.0.1:4040 查看待处理的变更,并提交或丢弃每一个。只读调用(get*、list*、search* 等)会立即通过,永远不会出现在仪表板中。
分类工作原理
参见 src/interceptor.ts — DEFAULT_CONFIG。如果工具调用的名称匹配变更动词模式(write、delete、create 等)且不首先匹配安全读取模式,或者任何字符串参数包含 SQL 变更关键字(insert into、drop table 等),则该工具调用会被拦截。这是一个启发式方法,而非保证——请参阅限制。
限制
启发式分类器。 基于工具名称和参数文本的正则表达式会遗漏巧妙命名的变更工具,并可能过度拦截名称奇怪的读取工具。请将其视为减速带,而非安全边界。
仪表板无身份验证。 它仅绑定到
127.0.0.1,但任何能够访问该端口的本地进程(或理论上进行 DNS 重新绑定的恶意页面)都可以提交/丢弃。不要在共享或不受信任的机器上运行此程序。仅内存存储。 重启代理会丢弃所有待处理的变更——代理的原始调用只会超时。
单开发者,本地使用。 没有共享的审计日志或团队范围的策略执行。如果你需要这些,请考虑使用托管的 MCP 网关。
许可证
AGPL-3.0
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Servers
- Alicense-qualityDmaintenanceA lightweight stdio proxy that intercepts and rewrites MCP tool annotations to bypass security approval prompts in AI CLIs like Codex and Claude Code. It transparently passes through all tool operations while marking them as safe to ensure a seamless automation experience.1110MIT
- AlicenseAqualityAmaintenanceSupply-chain gate for AI agent skills and MCP servers: poison-scan tool definitions, hash-pin the vetted set into a lock file and verify drift in CI. truecopy-mcp is a drop-in stdio proxy that filters a live server's tools/list down to its pinned, unmodified, unpoisoned tools.1222351MIT
- Alicense-qualityAmaintenanceZero-dependency local proxy that wraps any MCP server to redact secrets, strip hidden-Unicode prompt injection, and block writes to protected paths like ~/.ssh and .env.4MIT
- Alicense-qualityBmaintenanceA least-privilege enforcement proxy for MCP servers. It sits between MCP clients and upstream servers, enforcing tool policies, hiding denied tools, requiring human approval for risky actions, and providing a structured audit trail.MIT
Related MCP Connectors
Scans MCP servers for tool poisoning, prompt injection and supply chain risks.
Security scanner for MCP servers. Detect vulnerabilities, prompt injection, and tool poisoning.
A MCP server built for developers enabling Git based project management with project and personal…
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/pranavgawasproject/shadowrun-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server