qwen-web-arwaky: Skill for Claude Code

cleanup-consolidate-python

Purpose

Unified Python codebase cleanup skill combining dead code removal and file consolidation. First find and remove dead code, unused files, stubs, thin wrappers, and duplicates. Then detect overlapping files that share the same domain and merge them into single cohesive modules. The result is a cleaner codebase with fewer files, less bloat, and maximum signal-to-noise ratio.

CRITICAL: Two-Phase Approach — Phase 1 removes dead code. Phase 2 merges overlapping files. Never skip Phase 1 — consolidating files with dead code wastes effort.

Rules

• Never remove real logic — only remove code not relevant to FRD scope
• Always update __all__ — when removing functions/classes, remove from __all__ too
• Always update __init__.py — when deleting modules, remove their imports/re-exports
• Always run lint + tests after changes — verify no breakage
• Always snapshot before cleanup — git commit or stash before any deletion
• Respect # noqa — developer explicitly suppressed a lint; investigate intent before removing
• Respect # type: ignore — may indicate intentional dynamic typing
• Respect # pragma: no cover — code intentionally excluded from coverage; investigate why
• Respect decorator-registered code — @app.route, @pytest.fixture, @celery.task, @receiver etc. are NOT dead code even if never directly called
• Respect if TYPE_CHECKING: blocks — these imports are used by type checkers, not at runtime
• Respect try/except ImportError fallbacks — conditional imports for optional dependencies
• File with 0 inbound imports AND not an entry point = likely unused (verify with multi-pattern check)
• File with only re-exports in __init__.py = evaluate whether re-export adds value
• Single Responsibility (consolidation): each file should have ONE clear purpose
• Related classes/functions belong in the same file (consolidation)

When to Use