AgentSkillsCN

refactor-radar

识别代码异味、复杂度热点以及重构机会。按影响程度进行优先级排序，并提出切实可行的修复建议。可通过“/refactor”指令调用。

SKILL.md
--- frontmatter
name: refactor-radar
description: >
  Identify code smells, complexity hotspots, and refactoring opportunities.
  Prioritize by impact and suggest concrete fixes.
  Invoke with /refactor.

Refactor Radar

Commands

  • •/refactor <path> — Scan a file or directory for refactoring opportunities
  • •/refactor --hotspots — Find the most complex files in the project
  • •/refactor --duplicates — Find code duplication across the codebase

Code Smells Detected

Structural

  • •Long methods (50+ lines)
  • •God classes (300+ lines, 10+ methods)
  • •Deep nesting (3+ levels of if/for/while)
  • •Long parameter lists (5+ parameters)
  • •Feature envy (method uses another class's data more than its own)

Duplication

  • •Copy-paste blocks (3+ identical or near-identical blocks)
  • •Similar logic in different functions that should be unified
  • •Repeated patterns that could be abstracted

Complexity

  • •Cyclomatic complexity above 10
  • •Boolean logic with 3+ conditions
  • •Switch/match statements with 7+ cases
  • •Callback hell or deeply nested promises

Naming

  • •Single-letter variables outside loop counters
  • •Misleading names (variable does something different than name suggests)
  • •Inconsistent naming conventions within a file

Procedure

Phase 1: Scan

  1. •Read the target files
  2. •Calculate metrics (line count, complexity, nesting depth)
  3. •Identify code smells from the checklist above

Phase 2: Prioritize

Rate each finding by:

  • •Impact: How much does this hurt readability/maintainability? (HIGH/MEDIUM/LOW)
  • •Effort: How hard is the refactor? (HIGH/MEDIUM/LOW)
  • •Risk: Could the refactor break things? (HIGH/MEDIUM/LOW)

Priority = High Impact + Low Effort + Low Risk first.

Phase 3: Report

For each finding:

  • •Location (file:line range)
  • •Smell type
  • •Current state (brief description)
  • •Suggested refactor (concrete approach)
  • •Impact/effort/risk assessment

Phase 4: Refactor (if requested)

Only refactor when the user explicitly asks. Never auto-refactor.

MCMAP-Specific Patterns

  • •Python scripts in scripts/ — check for config duplication across files
  • •Cloud Flow JSON — check for duplicated action patterns across workflows
  • •Copilot instructions — check for inconsistent formatting across agents