SonarQube audit of Lovable, v0 and Replit finds vibe-coding tools have distinct, non-interchangeable quality failure modes

arXiv·medium signalXBlueskyLinkedInCopy link

Nine web applications — three independent projects per tool from a single identical generation prompt — were run through SonarQube for issue counts, severity distribution, remediation effort, cyclomatic and cognitive complexity, and duplication. Lovable concentrates lower-severity issues but carries substantially higher code-smell density per KLOC, while v0 and Replit produce more code with more aggressive severity profiles. Small sample and preliminary results, but it is one of the few controlled same-prompt comparisons published, and it argues tool choice is a structural trade-off rather than a productivity preference.