Score a codebase for "slop" using the SlopCodeBench metrics (structural erosion and verbosity) from arXiv 2603.24755, with published human and agent baselines to compare against. Use when asked how sloppy / bloated / eroded a repo is, whether code looks AI-generated or agent-degraded, where complexity is concentrated, whether a refactor is worth doing, or for any request naming slop, erosion, verbosity, SlopCodeBench, scb-check, code quality metrics, complexity concentration, or code smells at repo scale. Also use when a cyclomatic-complexity question is really a "is this codebase in bad shape" question.
