zmcp
local mcp server
Documentation
Shell Executor MCP Server
A Model Context Protocol (MCP) server that provides safe shell command execution capabilities for AI agents and tools like VS Code Copilot.
Features
- Safe Command Execution: Whitelist-based command filtering
- Comprehensive Tool Set: Support for file operations, text processing, development tools, and system utilities
- MCP Compatible: Full Model Context Protocol support
- Easy Integration: Simple JSON-RPC interface over stdio
Installation
Option 1: Direct Usage
git clone https://github.com/yourusername/shell-executor-mcp.git
cd shell-executor-mcp
pip install -e .Option 2: Install from PyPI (when published)
pip install shell-executor-mcpVS Code Integration
1. Global Configuration
Add to your VS Code settings.json or MCP configuration:
{
"mcpServers": {
"shell-executor": {
"command": "python",
"args": ["/path/to/shell-executor-mcp/src/mcp_server.py"],
"env": {
"MCP_API_TOKEN": "your-secret-token"
}
}
}
}2. Workspace Configuration
Create `.vscode/mcp_servers.json` in your workspace:
{
"shell-executor": {
"command": "python",
"args": ["${workspaceFolder}/path/to/mcp_server.py"],
"env": {
"MCP_API_TOKEN": "your-secret-token"
}
}
}3. Using with VS Code Copilot
Once configured, Copilot can discover and use the shell executor through MCP:
@shell-executor execute ls -la
@shell-executor execute git status
@shell-executor execute cargo buildManual Testing
Start the server:
python src/mcp_server.pySend test requests:
1. Initialize:
{"jsonrpc": "2.0", "id": 1, "method": "initialize", "params": {"protocolVersion": "2024-11-05", "capabilities": {}, "clientInfo": {"name": "test", "version": "1.0.0"}}}2. List tools:
{"jsonrpc": "2.0", "id": 2, "method": "tools/list", "params": {}}3. Execute command:
{"jsonrpc": "2.0", "id": 3, "method": "tools/call", "params": {"name": "execute_command", "arguments": {"command": "echo", "args": ["Hello World"]}}}Supported Commands
File System
- `ls`, `cat`, `head`, `tail`, `find`, `tree`, `wc`, `du`, `df`
- `file`, `stat`, `pwd`
Text Processing
- `grep`, `sed`, `awk`, `sort`, `uniq`, `cut`, `tr`
Development
- `git`, `npm`, `yarn`, `pip`, `python`, `node`, `cargo`, `make`
- `rustc`, `go`, `java`, `javac`
System Info
- `ps`, `uptime`, `whoami`, `id`, `uname`, `which`
Network
- `curl`, `wget`, `ping`, `dig`, `nslookup`
Archives
- `tar`, `zip`, `unzip`, `gzip`, `gunzip`
Utilities
- `echo`, `date`, `cal`, `bc`, `expr`, `basename`, `dirname`
- Hash tools: `md5`, `sha256sum`, `shasum`
Security
- Commands are filtered through a whitelist
- 30-second execution timeout
- No destructive operations (rm, mv, cp) allowed
- No privilege escalation (sudo, su) allowed
- Environment variables can be controlled
Configuration
Set environment variables:
- `MCP_API_TOKEN`: Authentication token (default: "MYSECRET")
License
MIT License - see LICENSE file for details.
Frequently asked questions
What is zmcp?
zmcp is local mcp server
How do I install zmcp?
Open the GitHub repository and follow its README. Most MCP servers are added to your client's MCP config, then called by your agent.
Is zmcp open source?
Yes — it is hosted on GitHub at https://github.com/kosiew/zmcp.
Related MCP tools
Cognee is the open-source AI memory platform for agents. Give your AI agents persistent long-term memory across sessions with a self-hosted knowledge graph engine.
Automate browser based workflows with AI
Hindsight: Agent Memory That Learns
A privacy-first app that strips AI watermarks from content you own.
Agent framework and applications built upon Qwen>=3.0, featuring Function Calling, MCP, Code Interpreter, RAG, Chrome extension, etc.
The power of Claude Code / GeminiCLI / CodexCLI + [Gemini / OpenAI / OpenRouter / Azure / Grok / Ollama / Custom Model / All Of The Above] working as one.
Run your own MCP server? See who uses it and what to fix.
Measure it with TrackMCP