konflux-devlake-mcp: Cursor rule for Cursor

Konflux DevLake MCP Server - Cursor Rules

Project Overview

This is a Model Context Protocol (MCP) server that provides tools for querying and analyzing data from the Konflux DevLake database. The server exposes tools for database operations, incident analysis, deployment tracking, and PR retest analysis.

Code Style & Formatting

Python Code Style

• Formatter: Use black with line length 100
  • Command: black --line-length 100 .
  • Always format code before committing
• Linter: Use flake8 for code quality
  • Command: flake8 .
  • Fix all linting errors before committing
• Type Hints: Use type hints for all function parameters and return types
• Docstrings: Use Google-style docstrings for all classes and functions

YAML Code Style

• Linter: Use yamllint for YAML file validation
  • Command: yamllint .
  • Configuration file: .yamllint (extends default, max line length 200)
  • Fix all linting errors before committing
  • Applies to: .github/workflows/*.yaml, k8s/*.yaml, and all other YAML files

Code Formatting Rules

• Maximum line length: 100 characters
• Use double quotes for strings (black default)
• Use trailing commas in multi-line collections
• Use f-strings for string formatting (not .format() or %)
• Use async/await for all database operations and tool calls
• NO EMOJIS: Do not use any emojis in code, comments, docstrings, or any project files

Architecture Patterns

Tool Creation Pattern

All tools must:

1. Inherit from BaseTool in tools/base/base_tool.py
2. Implement get_tools() method returning a list of Tool objects
3. Implement call_tool(name, arguments) method for async tool execution
4. Use log_tool_call() for all tool invocations
5. Return TOON-encoded strings for token efficiency (not JSON)

Tool Structure

class MyTool(BaseTool):
    def __init__(self, db_connection):
        super().__init__(db_connection)
        self.logger = get_logger(f"{__name__}.MyTool")
    
    def get_tools(self) -> List[Tool]:
        return [Tool(name="...", description="...", inputSchema={...})]
    
    async def call_tool(self, name: str, arguments: Dict[str, Any]) -> str:
        # Use toon_encode for output
        return toon_encode(result, {"delimiter": ",", "indent": 2, "lengthMarker": ""})