---

name: code-cleanup description: | Detect and clean redundant code in Python and TypeScript projects. Finds unused imports, dead functions, stale comments, debug statements, and unused dependencies. Use when asked to "clean code", "find dead code", "remove unused imports", "check redundant code", or "清理代码".

Code Cleanup Skill

Scan projects for redundant code and generate cleanup reports.

Quick Start

# Scan for dead code (unused imports, functions, variables)
python scripts/scan_dead_code.py backend/

# Scan for stale code (old TODOs, debug statements, commented code)
python scripts/scan_stale_code.py src/

# Check unused dependencies
python scripts/scan_unused_deps.py .

Scripts

Script	Detects	Confidence
scan_dead_code.py	Unused imports, functions, classes, variables	High
scan_stale_code.py	Commented code, old TODOs, debug statements	Medium
scan_unused_deps.py	Unused packages in requirements.txt/package.json	Low

Usage

scan_dead_code.py

python scripts/scan_dead_code.py <dir> [--format json|text] [--ignore "test_*"]

Detects:

• Unused imports (high confidence, safe to remove)
• Unused functions/classes not starting with _ (medium confidence)
• Unused top-level variables (medium confidence)

scan_stale_code.py

python scripts/scan_stale_code.py <dir> [--todo-days 90] [--no-debug]

Detects:

• Commented-out code blocks
• TODOs/FIXMEs older than N days (default: 90)
• Debug statements: print(, console.log(, debugger, pdb

scan_unused_deps.py

python scripts/scan_unused_deps.py <dir> [--type python|node|all]

Checks:

• requirements.txt vs actual imports
• package.json dependencies vs actual imports

Output Format

All scripts output JSON by default:

{
  "target": "backend/",
  "issues": [
    {
      "type": "unused_import",
      "file": "services/user.py",
      "line": 3,
      "code": "import os",
      "confidence": "high",
      "suggestion": "Remove unused import: os"
    }
  ],
  "summary": {
    "total": 15,
    "high": 10,
    "medium": 5,
    "low": 0
  }
}

Confidence Levels

Level	Meaning	Action
high	Definitely unused	Safe to remove
medium	Likely unused	Review before removing
low	Possibly unused	Manual inspection needed

Cleanup Workflow

1. Run scan scripts on target directory
2. Review JSON output, filter by confidence
3. For each issue: remove, keep, or defer
4. Apply changes (Claude can help edit files)
5. Run tests to verify nothing broke

Safety Notes

• Scripts are read-only (scan only, no modifications)
• Always backup before bulk cleanup
• Run tests after cleanup
• Some "unused" code may be used dynamically (reflection, __getattr__, lazy imports)
• Ignore __init__.py re-exports

References

• Detection patterns: references/detection-patterns.md
• Cleanup examples: references/cleanup-examples.md