mcp-shellcheck
MCP server that provides shell script linting via ShellCheck. Allows AI agents to analyze shell scripts for common errors, stylistic issues, and potential bugs.
README
ShellCheck MCP Server
A Model Context Protocol (MCP) server that provides shell script linting via ShellCheck. Allows AI agents to analyze shell scripts for common errors, stylistic issues, and potential bugs.
MCP Server Profile
{
"name": "mcp-shellcheck",
"description": "A Model Context Protocol (MCP) server that provides shell script linting via ShellCheck for AI coding assistants",
"tools": [
{
"name": "shellcheck",
"description": "Run ShellCheck on a shell script to find bugs, stylistic issues, and potential errors",
"inputSchema": {
"type": "object",
"properties": {
"file_path": { "type": "string", "description": "Path to the shell script file to check" },
"script_content": { "type": "string", "description": "Raw shell script content to check" },
"shell": { "type": "string", "description": "Shell type to check", "enum": ["bash", "sh", "dash", "ksh", "ash"] },
"check_sourced": { "type": "boolean", "description": "Enable checks for sourced files" },
"enable_all": { "type": "boolean", "description": "Enable all optional checks" },
"exclude": { "type": "string", "description": "Comma-separated list of warning codes to exclude" },
"severity": { "type": "string", "description": "Minimum severity to report", "enum": ["error", "warning", "info", "style"] }
}
}
},
{
"name": "shellcheck_info",
"description": "Get information about the ShellCheck version and server capabilities",
"inputSchema": {
"type": "object",
"properties": {}
}
}
]
}
Quick Install
# Run via uvx from PyPI
uvx --from mcp-shellcheck shellcheck-mcp-server
# Run via uvx from GitHub release
uvx --from https://github.com/Ev3lynx727/mcp-shellcheck/releases/download/v0.1.3/mcp_shellcheck-0.1.3-py3-none-any.whl shellcheck-mcp-server
# One-liner install (install.sh)
curl -fsSL https://raw.githubusercontent.com/Ev3lynx727/mcp-shellcheck/main/install.sh | sh
# Install from PyPI
pip install mcp-shellcheck
# Clone and dev install
git clone https://github.com/Ev3lynx727/mcp-shellcheck.git
cd mcp-shellcheck && pip install -e .
Features
- File-based analysis: Check shell scripts by file path
- Inline script checking: Analyze raw shell script content directly
- Multiple shell support: bash, sh, dash, ksh, ash
- Configurable checks: Exclude specific warnings, set severity levels
- Structured output: JSON-formatted results for easy parsing
- OpenCode integration: Ready to use with OpenCode agents
- Production-ready: Async, tested, validated, logged
Requirements
- Python 3.10+
- ShellCheck installed on the system
Installing ShellCheck
Recommended (always latest):
pip install shellcheck-py
The shellcheck-py package provides a pre-built shellcheck v0.11.0 binary on your PATH with no system dependencies.
System package managers (may ship older versions):
# Ubuntu/Debian
sudo apt-get install shellcheck
# macOS
brew install shellcheck
# Fedora/RHEL
sudo dnf install ShellCheck
# Arch Linux
sudo pacman -S shellcheck
Tools
shellcheck
Run ShellCheck on a shell script to find bugs, stylistic issues, and potential errors.
Parameters:
| Parameter | Type | Required | Description |
|---|---|---|---|
file_path |
string | No* | Path to the shell script file |
script_content |
string | No* | Raw shell script content |
shell |
string | No | Shell type: bash, sh, dash, ksh, ash (default: bash) |
check_sourced |
boolean | No | Enable checks for sourced files (default: false) |
enable_all |
boolean | No | Enable all optional checks (default: false) |
exclude |
string | No | Comma-separated codes to exclude (e.g., "SC1090,SC2148") |
include |
string | No | Comma-separated codes to include (e.g., "SC2086,SC2164") |
severity |
string | No | Minimum severity: error, warning, info, style |
*Either file_path or script_content must be provided.
Common error codes:
| Code | Description | Severity |
|---|---|---|
| SC1090 | Can't follow non-constant source | info |
| SC2086 | Double quote to prevent globbing | warning |
| SC2164 | Use cd with || exit | warning |
| SC2006 | Use $(...) instead of legacy backticks | style |
shellcheck_info
Get ShellCheck version and server capabilities.
Parameters: None
Configuration
OpenCode
{
"mcp": {
"shellcheck": {
"type": "local",
"command": [
"uv",
"run",
"--with", "mcp",
"python3",
"/path/to/mcp-shellcheck/shellcheck_mcp_server.py"
],
"enabled": true,
"timeout": 60000
}
}
}
OpenCode (uvx from GitHub release)
{
"mcp": {
"shellcheck": {
"type": "local",
"command": [
"uvx",
"--from", "https://github.com/Ev3lynx727/mcp-shellcheck/releases/download/v0.1.3/mcp_shellcheck-0.1.3-py3-none-any.whl",
"shellcheck-mcp-server"
],
"enabled": true,
"timeout": 60000
}
}
}
Claude Desktop
{
"mcpServers": {
"shellcheck": {
"command": "python3",
"args": ["/path/to/mcp-shellcheck/shellcheck_mcp_server.py"]
}
}
}
Cursor
{
"mcpServers": {
"shellcheck": {
"command": "uvx",
"args": ["shellcheck-mcp-server"]
}
}
}
VS Code (Copilot)
Add to .vscode/mcp.json:
{
"servers": {
"shellcheck": {
"command": "python3",
"args": ["/path/to/mcp-shellcheck/shellcheck_mcp_server.py"]
}
}
}
Examples
Check a File
// Input
{ "file_path": "/path/to/deploy.sh" }
// Output
{
"success": false,
"message": "Found 3 issue(s)",
"results": [
{
"line": 15,
"column": 10,
"code": "SC2086",
"message": "Double quote to prevent globbing",
"severity": "warning"
}
],
"exit_code": 1
}
Check Script Content
// Input
{ "script_content": "#!/bin/bash\ncat `ls *.txt`", "shell": "bash" }
Exclude Specific Warnings
{ "file_path": "/path/to/script.sh", "exclude": "SC1090,SC2148" }
Filter by Severity
{ "file_path": "/path/to/script.sh", "severity": "error" }
Troubleshooting
| Problem | Solution |
|---|---|
| ShellCheck not found | Install via pip install shellcheck-py or system package manager |
| MCP package not installed | pip install mcp |
| Server not connecting | Verify shellcheck --version, test with python3 shellcheck_mcp_server.py |
| Timeout errors | Increase timeout: "timeout": 120000 in MCP config |
Development
pip install -e ".[dev]"
pytest
ruff check .
See CHANGELOG.md for release history and ARCHITECTURE.md for design docs.
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。