Vibe-Code Cleanup — Production Refactor Skill

Vibe-Code Cleanup — Production Refactor Skill

A safe, incremental cleanup workflow for AI-generated / vibe-coded fullstack apps. The goal is to make the codebase production-ready without breaking anything that already works.

When to Use

• Use when a rapidly built app works but has broken imports, duplicated logic, dead code, unclear environment variables, or fragile release hygiene.
• Use before launch or handoff to convert exploratory code into a maintainable production baseline.
• Use when cleanup must preserve existing behavior and avoid broad rewrites of routes, APIs, auth, data models, or integrations.

Core Philosophy

Surgery, not demolition. Remove only what is provably dead. Preserve everything else.

Never:

• Rewrite working systems for cosmetic reasons
• Rename routes, slugs, or API endpoints that may be indexed or cached
• Change tool inputs/outputs, API contracts, DB schema, or auth flow
• Delete files you haven't verified are unused
• Make broad sweeping changes in a single commit

Always:

• Make small, targeted, reversible changes
• Validate after every meaningful batch of changes
• Prefer shared helpers over copy-pasted blocks
• Keep backward compatibility

Step 1 — Reconnaissance (read before touching)

Before changing anything, map the codebase:

# List all pages/routes
find . -type f \( -name 'page.js' -o -name 'page.jsx' -o -name 'page.ts' -o -name 'page.tsx' \)
find pages -type f \( -name '*.js' -o -name '*.jsx' -o -name '*.ts' -o -name '*.tsx' \) | rg -v '/_' | sort

# Find broken imports (TS projects)