← 返回 Skills 市场
556
总下载
0
收藏
7
当前安装
2
版本数
在 OpenClaw 中安装
/install mcp-bridge
功能描述
Use mcp-bridge-openclaw CLI to connect to and manage Model Context Protocol (MCP) servers with auto-reconnection and retry logic. Install via npm install -g...
使用说明 (SKILL.md)
mcp-bridge-openclaw
CLI tool for connecting to MCP servers with built-in resilience.
Installation
npm install -g mcp-bridge-openclaw
Verified publisher: npm user jaggu37
Commands
Connect to MCP server
mcp-bridge --config config.json
List available servers
mcp-bridge --config config.json --list
Run with verbose logging
mcp-bridge --verbose --config config.json
Configuration
Create config.json:
{
"servers": {
"filesystem": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-filesystem", "/tmp"],
"env": {}
}
}
}
Security tip: Use environment variables for tokens instead of plaintext in config:
{
"servers": {
"github": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-github"],
"env": {
"GITHUB_TOKEN": "${GITHUB_TOKEN}"
}
}
}
}
Then run: GITHUB_TOKEN=your_token mcp-bridge --config config.json
Programmatic Usage
import { MCPBridge } from 'mcp-bridge-openclaw';
const bridge = new MCPBridge({
configPath: './config.json',
onServerConnect: (name) => console.log(`Connected to ${name}`),
});
await bridge.connect();
await bridge.disconnect();
Key Features
- Auto-reconnect on disconnect
- Configurable retry logic
- Type-safe JSON config
- CLI + programmatic API
- Multiple server support
安全使用建议
This skill appears to do what it says, but it executes and downloads code from npm (both during global install and via npx when launching server implementations). Before installing: 1) inspect the npm package and its GitHub repository (publisher, recent commits, issues). 2) Avoid global install if possible; consider installing in an isolated environment/container. 3) Pin versions and review package.json/maintainers. 4) Limit and scope any tokens (e.g., GITHUB_TOKEN) used by server implementations and do not store secrets in plaintext config files. 5) Review config.json to ensure the configured command/args are trusted (they can run arbitrary commands). If you need higher assurance, request the package source code or a reproducible build before use.
功能分析
Type: OpenClaw Skill
Name: mcp-bridge
Version: 1.0.1
The skill bundle contains metadata and documentation for the 'mcp-bridge-openclaw' CLI tool, which is designed to manage Model Context Protocol (MCP) servers. The instructions in SKILL.md are consistent with the stated purpose, providing standard installation, configuration, and usage examples (including security best practices for environment variables) without any evidence of malicious intent, data exfiltration, or prompt injection.
能力评估
Purpose & Capability
Name, description, and runtime instructions all describe a CLI/SDK for connecting to MCP servers. The npm package + config-based server definitions (including running server implementations via npx) are coherent with the stated purpose.
Instruction Scope
SKILL.md instructs installing the npm package and running the CLI with a config.json. The config lets you specify arbitrary commands/args (e.g., npx calls) and environment variables, which is expected for a bridge but means the skill's runtime can execute arbitrary third‑party code and read the provided config file and any env vars you pass.
Install Mechanism
There is no formal install spec in the skill bundle (it's instruction‑only), but SKILL.md tells users to run npm install -g. Installing a global npm package and using npx to fetch server implementations at runtime is typical here but increases supply‑chain risk because code is fetched from npm/GitHub during install or execution.
Credentials
The manifest declares no required env vars. SKILL.md recommends using environment variables for tokens (example: GITHUB_TOKEN). That is reasonable, but tokens grant access to external services and should be scoped and protected; the skill does not request unrelated credentials.
Persistence & Privilege
always is false and the skill is user-invocable. The skill does not request elevated platform privileges in the manifest. No indication it modifies other skills or system-wide settings.
如何使用
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install mcp-bridge - 安装完成后,直接呼叫该 Skill 的名称或使用
/mcp-bridge触发 - 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.1
- Added homepage and repository links to SKILL.md for improved discoverability.
- Noted verified npm publisher (`jaggu37`) in the documentation.
- Updated config example to emphasize security best practices using environment variables for tokens.
- Simplified configuration and usage examples for clarity.
- Minor edits for improved documentation and security guidance.
v1.0.0
Initial release of mcp-bridge.
- Provides CLI and programmatic tools to connect and manage Model Context Protocol (MCP) servers.
- Supports auto-reconnection and configurable retry logic.
- Manages multiple servers through a JSON config file.
- Easy installation via npm.
- Includes verbose logging and server listing options.
元数据
常见问题
mcp-bridge 是什么?
Use mcp-bridge-openclaw CLI to connect to and manage Model Context Protocol (MCP) servers with auto-reconnection and retry logic. Install via npm install -g... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 556 次。
如何安装 mcp-bridge?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install mcp-bridge」即可一键安装,无需额外配置。
mcp-bridge 是免费的吗?
是的,mcp-bridge 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
mcp-bridge 支持哪些平台?
mcp-bridge 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。
谁开发了 mcp-bridge?
由 jaggu1999(@jaggu1999)开发并维护,当前版本 v1.0.1。
推荐 Skills