AgentSkillsCN

pyscn-mcp

使用 MCP 工具分析 Python 代码质量——包括复杂度、代码重复、死代码与耦合度。适用于用户询问代码质量、重构、可维护性、重复代码,或技术债务等问题时。

SKILL.md
--- frontmatter
name: pyscn-mcp
description: Analyze Python code quality using MCP tools - complexity, clones, dead code, coupling. Use when user asks about code quality, refactoring, maintainability, duplicates, or technical debt.

Python Code Quality Analysis with pyscn MCP

Use the pyscn MCP tools for Python code quality analysis.

Available Tools

ToolPurpose
get_health_scoreOverall code health score (0-100) with grade
analyze_codeComprehensive analysis (complexity, dead code, clones, coupling, deps)
check_complexityCyclomatic complexity of functions
detect_clonesDuplicate code detection
find_dead_codeUnreachable code detection
check_couplingClass coupling (CBO) metrics

Tool Selection Guide

User RequestTool
"How healthy is this code?"get_health_score
"Analyze code quality"analyze_code
"Find complex functions"check_complexity
"Find duplicate code"detect_clones
"Find dead code"find_dead_code
"Check class coupling"check_coupling

Common Parameters

  • path (required): Path to Python file or directory
  • recursive (analyze_code): Recursively analyze directories (default: true)
  • analyses (analyze_code): Array of analyses to run - complexity, dead_code, clone, cbo, deps

Examples

Quick Health Check

Use get_health_score for a quick overview:

  • Returns score 0-100 with letter grade (A-F)
  • Category breakdowns for maintainability, reliability, etc.

Detailed Analysis

Use analyze_code with specific analyses:

  • analyses: ["complexity"] - Only complexity
  • analyses: ["clone"] - Only duplicates
  • analyses: ["dead_code"] - Only dead code
  • analyses: ["complexity", "dead_code"] - Multiple

Complexity Thresholds

Use check_complexity with:

  • min_complexity: Minimum to report (default: 1)
  • max_complexity: Maximum allowed (default: 0 = no limit)

Clone Detection

Use detect_clones with:

  • similarity_threshold: 0.0-1.0 (default: 0.8)
  • min_lines: Minimum lines to consider (default: 5)

Always explain results and suggest improvements based on findings.