Problem
slop-scan is great at catching code-level patterns, but there's a category of AI slop that lives outside source files, in READMEs, docs, and the overall repo composition. Maybe it is already thought for next releases, I just wanted to share my experience with vibe-coded projects:
Exaggerated README
Made-up metrics ("reduces load time by 40%"), buzzword-heavy lists ("seamless", "blazing fast", "enterprise-grade"), emoji-stuffed feature grids, and sections that don't actually say anything useful. AI-generated READMEs love listing stuff that are generic enough to apply to literally any project.
Weird Docs
Architecture sections that reference components that don't exist in the source tree, config examples that don't match real options, and "API reference" sections describing endpoints with no implementation behind them.
Right now slop-scan can't see this because it discovers files matching the JS/TS extensions. I hope it will be added in the next versions.
Problem
slop-scan is great at catching code-level patterns, but there's a category of AI slop that lives outside source files, in READMEs, docs, and the overall repo composition. Maybe it is already thought for next releases, I just wanted to share my experience with vibe-coded projects:
Exaggerated README
Made-up metrics ("reduces load time by 40%"), buzzword-heavy lists ("seamless", "blazing fast", "enterprise-grade"), emoji-stuffed feature grids, and sections that don't actually say anything useful. AI-generated READMEs love listing stuff that are generic enough to apply to literally any project.
Weird Docs
Architecture sections that reference components that don't exist in the source tree, config examples that don't match real options, and "API reference" sections describing endpoints with no implementation behind them.
Right now slop-scan can't see this because it discovers files matching the JS/TS extensions. I hope it will be added in the next versions.