Loading ToolBench
A comprehensive MCP server hub providing multiple tools for GUI automation, CLI operations, notebook management, Python script execution, and FTP file operations
MCPHub has 24 tools with significant definition gaps. Tool names follow verb_noun convention (start_task, get_status, run_command), which is good. However, descriptions are minimal (10-50 chars for most tools), parameter descriptions are sparse or missing, and input schemas lack proper type definitions and constraints. No output schemas are documented. Error handling is absent, tools return generic success/error dicts without recovery guidance. Security concerns: run_command and kill_process are destructive but lack permission gates or dry-run patterns. FTP tools accept file paths without sanitization, risking path traversal. Overall, definitions are skeletal and would not pass production code review.
将笔记添加到指定文件中。
在 FTP 服务器上创建目录。支持创建多级目录（父目录不存在时自动递归创建），支持绝对路径和相对路径，自动处理路径分隔符。
从 FTP 服务器下载文件或目录到本地。下载单个文件直接保存到配置的下载目录，下载目录时先递归下载到临时目录再打包为指定格式，支持绝对路径和相对路径。
在指定目录下递归搜索匹配文件
获取磁盘使用情况
获取环境变量，可按前缀过滤
获取当前任务的执行历史。
获取网络配置信息
No output schemas documented for any tool. LLMs cannot plan downstream calls or extract required fields. Baseline: 100% of A+ tools document return types.
Destructive tools (run_command, kill_process, install_package) lack permission gates, dry-run patterns, or confirmation steps. No audit trail. Agents can execute arbitrary commands without safeguards.
FTP tools (upload_file, download_file, move_file, create_directory) accept file paths without sanitization. Risk of path traversal attacks. No validation against '../' or absolute paths outside allowed directories.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 47 | <=2025-11-25 | v2 |
获取运行中的进程列表
根据sessionId查询脚本执行状态
获取当前任务的运行状态。
获取系统信息
通过 pip 安装指定的 Python 包
强制终止指定 PID 的进程
列出目录内容
在 FTP 服务器上移动文件到目标目录。支持绝对路径和相对路径，目标目录不存在时自动创建，移动后验证文件是否存在于目标位置。
读取指定文件中的笔记。
执行任意命令行指令（带安全检查）
运行指定的Python脚本
异步运行指定的Python脚本
以UTF-8编码保存Python脚本内容
启动一个GUI自动化任务（后台异步运行）。
停止当前正在运行的任务。
通过 FTP 上传文件或目录到服务器。支持上传单个文件或整个目录，自动在 FTP 服务器上为指定用户创建专属目录，上传目录时递归处理所有子目录和文件，支持大文件流式传输。
Most tool descriptions are under 20 characters (e.g., 'Get system info', 'Get disk usage'). Baseline: 194 chars average. Descriptions lack context for LLM selection, WHEN to use the tool, and what it returns.
No error handling or recovery guidance. Tools return generic {'success': bool, 'err_msg': str} without categorizing errors as retryable, user-fixable, or fatal. LLMs cannot self-correct.