Building evals for AI coding agents, on your repo. Tests pass. Nobody's measuring the rest. http://stet.sh email ben@benr.build