Python Code Guardian MCP Server
Enables automated Python code quality checks including linting, complexity analysis, typo detection, structure validation, duplicate detection, and test coverage, with integration into Cursor IDE and CLI.
README
Python Code Guardian MCP Server
A comprehensive Model Context Protocol (MCP) server for Python code quality checks. This tool helps teams maintain high code quality standards through automated checks for linting, modularity, typos, structure, and test coverage.
🚀 Features
✅ Comprehensive Code Analysis
- Linting with Pylint - PEP 8 compliance, code style, and best practices
- Complexity Analysis - Cyclomatic complexity and maintainability checks
- Duplicate Code Detection - Identifies redundant code blocks
- Typo Detection - Catches typos in comments, docstrings, variable names, and strings
- Structure Validation - File organization, naming conventions, and docstrings
- Test Coverage - Ensures 75% minimum test coverage
✅ Flexible Integration
- Pre-commit hook support
- On-demand checks via Cursor IDE
- Custom rules configuration
- Detailed reports with line numbers and suggestions
✅ Developer Friendly
- Summary statistics
- Auto-fix capabilities
- Configurable severity levels
- Python 3.8+ support
📦 Installation
Prerequisites
- Python 3.8 or higher
- pip (Python package manager)
Option 1: Install from PyPI (When Published)
Once published to PyPI, you can install it directly:
pip install python-code-guardian-mcp
Option 2: Install from GitHub
Install directly from the GitHub repository:
pip install git+https://github.com/priyanshi9692/py-sage-mcp.git
Option 3: Install from Source (Recommended for Local Use)
Clone the repository and install in development mode:
git clone https://github.com/priyanshi9692/py-sage-mcp.git
cd py-sage-mcp
pip install -e .
This allows you to use the MCP server without publishing to PyPI. Changes to the code will be immediately available.
Option 4: Install with Development Dependencies
For development and testing:
git clone https://github.com/priyanshi9692/py-sage-mcp.git
cd py-sage-mcp
pip install -e ".[dev]"
Option 5: Use Without Installation (Direct Path)
You can use the MCP server directly without installing by pointing Cursor to the source directory. See the Quick Start section for configuration details.
Verify Installation
python-code-guardian --version
python-code-guardian --help
Note: You don't need to publish to PyPI to use this MCP server. Options 2, 3, and 5 allow you to use it directly from GitHub or your local machine.
🎯 Quick Start
1. Integrate MCP Server with Cursor IDE
Add the following configuration to your Cursor settings file:
For global configuration: ~/.cursor/config.json
For workspace configuration: .cursor/settings.json (in your project root)
Configuration for Installed Package (Options 1-4)
If you installed the package using any of the installation methods above:
{
"mcpServers": {
"python-code-guardian": {
"command": "python",
"args": [
"-m",
"python_code_guardian.server"
]
}
}
}
If using a virtual environment, use the full path to the Python executable:
{
"mcpServers": {
"python-code-guardian": {
"command": "/path/to/venv/bin/python",
"args": [
"-m",
"python_code_guardian.server"
]
}
}
}
Configuration for Direct Path (Option 5 - No Installation)
If you want to use the MCP server without installing it, point directly to the source:
{
"mcpServers": {
"python-code-guardian": {
"command": "/path/to/py-sage-mcp/venv/bin/python",
"args": [
"-m",
"python_code_guardian.server"
],
"env": {
"PYTHONPATH": "/path/to/py-sage-mcp/src"
}
}
}
}
Example with absolute path:
{
"mcpServers": {
"python-code-guardian": {
"command": "/Users/priyanshijajoo/Desktop/python-code-guardian-mcp/venv/bin/python",
"args": [
"-m",
"python_code_guardian.server"
],
"env": {
"PYTHONPATH": "/Users/priyanshijajoo/Desktop/python-code-guardian-mcp/src"
}
}
}
}
After adding the configuration, restart Cursor IDE to activate the MCP server.
2. Use via Cursor IDE Chat
Once configured, you can use the MCP server by asking the AI:
- "Check the code quality of this file"
- "Run code guardian on the current project"
- "Fix linting issues in this function"
- "What's the test coverage of this module?"
3. Command Line Usage
# Check entire project
python-code-guardian check --path ./src
# Check specific file
python-code-guardian check --path ./src/module.py
# Check with specific checks only
python-code-guardian check --path ./src --checks lint complexity
# Check with auto-fix
python-code-guardian check --path ./src --fix
# Get summary report
python-code-guardian check --path ./src --format summary
# Validate PR changes
python-code-guardian validate-pr --pr feature-branch --base main
# Initialize configuration file
python-code-guardian init-config
4. Setup Pre-commit Hook
# Install pre-commit if not already installed
pip install pre-commit
# Install the hook
python-code-guardian install-hook
# Test the hook
pre-commit run --all-files
⚙️ Configuration
Create a .code-guardian.yaml file in your project root:
python-code-guardian init-config
Customize the configuration:
rules:
pylint:
enabled: true
max_line_length: 100
disable:
- C0111 # missing-docstring for specific cases
complexity:
enabled: true
max_complexity: 10
max_function_length: 50
typos:
enabled: true
check_variables: true
check_comments: true
custom_dictionary:
- "customword"
- "anotherterm"
structure:
enabled: true
require_docstrings: true
test_coverage_threshold: 75
naming_convention: "snake_case"
duplicates:
enabled: true
min_lines: 6
ignore_patterns:
- "tests/*"
🛠️ MCP Tools Available
The server exposes the following MCP tools:
check_code
Runs all configured checks on specified files or directories.
Parameters:
path(string): File or directory path to checkchecks(array, optional): Specific checks to runfix(boolean, optional): Attempt to auto-fix issuesconfig_path(string, optional): Path to custom config file
Example:
{
"path": "src/",
"checks": ["lint", "complexity", "typos"],
"fix": false
}
get_report
Returns detailed analysis report for the last check.
Parameters:
format(string): "detailed" or "summary"
validate_pr
Validates all changes in a pull request.
Parameters:
base_branch(string): Base branch to compare againstpr_branch(string): PR branch with changes
configure_rules
Updates or views current configuration rules.
Parameters:
action(string): "get" or "set"rules(object): Rules configuration object
fix_issues
Automatically fix detected issues where possible.
Parameters:
path(string): File or directory path to fixissue_types(array): Types of issues to fix
📊 Output Format
Detailed Report
================================================================================
CODE QUALITY REPORT
================================================================================
FILE: src/module.py
--------------------------------------------------------------------------------
[ERROR] Line 45, Column 0 - Pylint (C0301)
Line too long (105/100)
[WARNING] Line 78, Column 4 - Complexity (CC=12)
Function 'process_data' has complexity 12 (max: 10)
Suggestion: Consider breaking into smaller functions
[INFO] Line 12, Column 8 - Typo
Variable name 'caluculate' might be misspelled
Suggestion: Did you mean 'calculate'?
[WARNING] Line 25, Column 0 - Duplicates
Duplicate code block found (6 lines)
Also in: src/other_module.py:line 45
Suggestion: Extract to a shared function
--------------------------------------------------------------------------------
SUMMARY STATISTICS
--------------------------------------------------------------------------------
Files Checked: 15
Total Issues: 23
- Errors: 5
- Warnings: 12
- Info: 6
Test Coverage: 78.5% ✓
Issues by Category:
- Linting: 8
- Complexity: 5
- Typos: 3
- Structure: 4
- Duplicates: 3
================================================================================
🎨 Custom Rules
Create custom rules by extending the base checker:
# custom_rules.py
from python_code_guardian.checkers import BaseChecker
class MyCustomChecker(BaseChecker):
async def check(self, file_path, config):
issues = []
# Your custom logic
with open(file_path, 'r') as f:
for i, line in enumerate(f, 1):
if 'TODO' in line:
issues.append(self.create_issue(
file_path=file_path,
line=i,
column=0,
severity="info",
code="TODO-CHECK",
message="TODO comment found",
suggestion="Consider creating a ticket"
))
return {"issues": issues}
Register in .code-guardian.yaml:
custom_checkers:
- path: ./custom_rules.py
class: MyCustomChecker
🔧 CI/CD Integration
GitHub Actions
Add to your .github/workflows/ci.yml:
- name: Check code quality
run: python-code-guardian check --path ./src
🆘 Troubleshooting
"pylint not found"
pip install pylint
"Coverage check failed"
pip install coverage pytest-cov
"Permission denied"
pip install --user python-code-guardian-mcp
MCP Server Not Working in Cursor
- Verify the configuration JSON is correct
- Check that Python path is correct (especially for virtual environments)
- Restart Cursor IDE after configuration changes
- Check Cursor's MCP server logs for errors
Python Version Issues
Ensure you're using Python 3.8+:
python --version
# If using Python 3, might need to use python3
python3 --version
🤝 Contributing
Contributions are welcome! Please see CONTRIBUTING.md for guidelines.
- Fork the repository
- Create your feature branch (
git checkout -b feature/amazing-feature) - Commit your changes (
git commit -m 'Add amazing feature') - Push to the branch (
git push origin feature/amazing-feature) - Open a Pull Request
📄 License
MIT License - see LICENSE file for details.
🆘 Support
- 📖 Documentation: README
- 🐛 Issues: GitHub Issues
- 💬 Discussions: GitHub Discussions
🙏 Acknowledgments
Built with:
Made with ❤️ for the global Python community
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。