Sonarqube Py

A Python MCP server for SonarQube — query projects, issues, quality gates, coverage, and security hotspots directly from any MCP-compatible AI agent.

Inspired by the official SonarQube MCP Server (Java/Kotlin), reimplemented in Python for lightweight deployment without JVM dependency.

Install

uvx sonarqube-mcp-py
# or
pip install sonarqube-mcp-py

Configuration

Set environment variables:

export SONARQUBE_URL=https://your-sonarqube-instance
export SONARQUBE_TOKEN=squ_your_token_here
export SONARQUBE_VERIFY_SSL=false  # optional, for self-signed certs

Transport

• stdio (default): for mcp.json integration with Claude Desktop, Cursor, Kiro CLI, etc.
• HTTP: set MCP_TRANSPORT=streamable-http and optionally MCP_PORT=8959

Available Tools

Tool	Description
search_sonarqube_projects	Search for projects
get_project_quality_gate_status	Get quality gate status (OK/ERROR)
get_component_measures	Get metrics (bugs, coverage, code smells, etc.)
list_quality_gates	List all quality gate definitions
search_sonar_issues	Search issues by severity and status
search_security_hotspots	Find security hotspots to review
show_security_hotspot	Get hotspot details
show_rule	Get rule description and examples
search_metrics	List all available metrics
list_pull_requests	List analyzed pull requests
get_file_coverage_details	Get per-file coverage data
search_files_by_coverage	Find files with lowest coverage
get_duplications	Get code duplication details

MCP Client Configuration

stdio (mcp.json)

{
  "sonarqube": {
    "command": "sonarqube-mcp-py",
    "env": {
      "SONARQUBE_URL": "https://your-instance",
      "SONARQUBE_TOKEN": "squ_xxx"
    }
  }
}

HTTP (standalone service)

{
  "sonarqube": {
    "url": "http://localhost:8959/mcp"
  }
}

Start in HTTP mode:

MCP_TRANSPORT=streamable-http MCP_PORT=8959 sonarqube-mcp-py

License

MIT