How does the command approval process work?
When a command is run for the first time, the server prompts the user to allow it once, allow it and remember the approval, or deny execution. Approved commands are stored persistently.
@codelion
基本信息
Dynamic Shell Command MCP Server
使用下面的配置,将此服务器添加到你的 MCP 客户端。
{
"mcpServers": {
"dynamic-shell-server": {
"command": "python",
"args": [
"-m",
"venv",
"venv"
]
}
}
}`command`: The command to execute
`command`: The command to revoke approval for
Dynamic Shell Server is a Model Context Protocol (MCP) server that enables secure execution of shell commands with a dynamic approval system. It allows running arbitrary commands while maintaining security through user approval and audit logging. This server is designed for developers using MCP-compatible clients like Claude Desktop.
Install by cloning the repository, setting up a Python virtual environment, and installing dependencies. Run standalone with python dynamic_shell_server.py or integrate with Claude Desktop by adding its configuration to claude_desktop_config.json. When a command is executed for the first time, the user is prompted to allow it once, allow and remember, or deny execution.
npm install only after approvalWhen a command is run for the first time, the server prompts the user to allow it once, allow it and remember the approval, or deny execution. Approved commands are stored persistently.
Data is stored in the directory `~/.config/mcp-shell-server/`, containing `approved_commands.json` (list of approved commands) and `audit.log` (execution history).
The server does not use shell execution, preventing shell injection attacks. It also requires user approval for first-time command execution and logs all commands.
Yes, commands have a 5-minute timeout protection to prevent runaway processes.
Yes, the `revoke_command_approval` tool allows revoking approval for any command that was previously approved.


Routebase keeps your API's design, docs, contract tests, mocks and monitoring in one living OpenAPI spec so they can't drift apart. Agents work on that same source of truth, under your team's permissions.

SSH MCP server for AI agents: remote commands, file transfer, log search and server audits through OpenSSH.

Know when your AI agents, cron jobs and CI pipelines fail, stall or go quiet. Your agent sets up monitors from a plain description, tells you what is broken and why, reads run history and traces, and writes notes on inci

Read and edit DB Planner database schemas (DBML/SQL), Mermaid diagrams and board layouts from any MCP client. 38 tools, OAuth sign-in, read and write as separate scopes enforced server-side.
评论