System Cleaner

Universal CLI for system cleanup and audit. Analyzes ML caches, Python dependencies, application leftovers, and security on macOS and Linux.

Problems Solved

• Unused ML models (Hugging Face, PyTorch, TensorFlow) consuming disk space
• Dependency conflicts in Python projects
• Removed application leftovers
• Security issues (file permissions, secrets in code)
• Build artifacts and unused dependencies

Features

• System scanning: caches, logs, application leftovers
• ML cache analysis: Hugging Face, PyTorch, TensorFlow
• Dependency analysis: conflicts, unused, outdated
• Security checks: permissions, SSH keys, sensitive files
• Cleanup recommendations with size estimates
• Cross-platform: macOS and Linux
• Reports: Markdown and JSON

Requirements

• Python 3.12+ (3.12, 3.13)
• uv (installed automatically)
• macOS or Linux

Installation

git clone https://github.com/FUYOH666/Cleaner-OS.git
cd Cleaner-OS
uv sync
uv run python -m syscleaner health

Run directly: syscleaner or system-cleaner

Usage

# Full scan
syscleaner scan --all

# Categories
syscleaner scan --caches
syscleaner scan --security
syscleaner scan --projects
syscleaner scan --dependencies
syscleaner scan --ml-cache

# Save and report
syscleaner scan --all --save-results results.json
syscleaner report --format markdown --output report.md --from-scan results.json

# Health check
syscleaner health

Configuration

Copy config.yaml.example to config.yaml and edit. Optional; defaults work out of the box.

scan:
  exclude_paths: [~/Library/Mail/, ~/Library/Messages/, ...]
  min_size_mb: 10
  check_security: true
  check_project_artifacts: true
  check_dependencies: true
  check_ml_cache: true

security:
  sensitive_patterns: ["*.env", "*credentials*", ...]
  check_ssh_permissions: true
  check_file_permissions: true

cleanup:
  safe_to_delete_patterns: ["**/__pycache__", "**/.DS_Store", ...]

What It Scans

1. Caches — macOS: ~/Library/Caches/, Linux: ~/.cache/
2. App leftovers — Compare Application Support with installed apps
3. Security — SSH permissions, sensitive files, world-readable configs
4. Hidden files — Large hidden files/dirs in home
5. Project artifacts — __pycache__, node_modules, dist, build
6. ML caches — Hugging Face, PyTorch, TensorFlow (unused models >30 days)
7. Dependencies — Conflicts, unused, outdated via uv pip check

Report Formats

Markdown: Summary, ML caches, dependencies, caches, security, cleanup recommendations.

JSON: Structured data for automation.

Security

• No automatic deletions; analysis and recommendations only
• Fail-fast on config errors
• Path validation, structured logging

Contributing

1. Fork, create branch, make changes
2. Run uv run ruff check ., uv run pyright, uv run pytest
3. Open Pull Request

License

MIT. See LICENSE.

Author

Aleksandr Mordvinov — GitHub | scanovich.ai