← ClaudeAtlas

do-oop-auditlisted

Audit Python classes for OOP and data-modeling anti-patterns. Use when reviewing class design, boundaries, or hierarchy, or 'check my classes', 'lint class structure', 'scan for design issues'.
tomcounsell/ai · ★ 18 · AI & Automation · score 74
Install: claude install-skill tomcounsell/ai
# OOP / Data Modeling Audit **Goal:** surface structural weaknesses in Python class design — field semantics, object boundaries, inheritance, naming consistency, coupling — by running the 14 checks below against every class in the target path and reporting findings grouped by severity. Framework-agnostic: Django, SQLAlchemy, Pydantic, dataclasses, and vanilla Python. Findings only — the skill never modifies source files and pauses for human review. Also the right skill for requests like "review the data model", "are there OOP problems", or a class-design review after refactoring. ## Invocation ``` /do-oop-audit [path] [--framework django|sqlalchemy|pydantic|auto] [--severity critical|warning|info] ``` - `path`: Directory or file to audit. Defaults to current project root. - `--framework`: Force a framework context. Default is `auto` (detect from imports and base classes). - `--severity`: Minimum severity to include in the report. Default: show all. ## Quick start 1. **Enumerate**: Find all `.py` files in the target path containing `class ` definitions 2. **Detect framework**: Identify Django / SQLAlchemy / Pydantic / dataclasses / vanilla (see Framework Detection below) 3. **Check**: Run each of the 14 audit checks against every class and cross-class relationship 4. **Filter**: If `--severity` is set, exclude findings below the threshold 5. **Report**: Present findings grouped by severity 6. **Pause**: Wait for human review — never modify source files --- ## Framework