Measure a repository against the CARL standard.

Paste a public GitHub repo. CARL clones it in memory, scores its agent-readiness across eight pillars with deterministic, evidence-based checks, and resolves a single level on the 0–6 scale — no code stored, no AI guesswork.