AI Code Archaeology – Health Check Before Inheriting AI Spaghetti

Analyze AI-written code to uncover hidden architecture traps and modification costs

Problem

A V2EX user inherited a store system heavily built with AI coding; issues never stopped, modifications became harder, and AI kept adding unnecessary corner-case defenses. He's torn between rewriting or patching. Another user found an AI-generated Controller injecting another Controller with @EventListener.

Solution

Upload a codebase to automatically analyze AI-generated code patterns: unreasonable dependency injections, over-defensive programming, circular dependencies, duplicate logic; output a modifiability score and refactoring suggestions; optionally trace which AI sessions touched each file if git history exists.

Users

Developers and tech leads inheriting AI-generated projects

MVP

Support zip upload or git repo connection; identify common AI code smells (e.g., Controllers injecting each other, unnecessary @Lazy); generate a visual dependency graph and issue list; output rewrite vs. maintain cost estimation.

Revenue

Pay-per-analysis ($19 each) or subscription ($29/month for unlimited analyses, with self-hosted option).

Why now

The first wave of vibe-coded projects is entering maintenance, technical debt is surfacing, and no dedicated tool exists for AI-specific code smells.