AI Eval

Agent skills for evaluating AI models for a real job: define the workload, draw candidates from OpenRouter, run wide and shallow, publish findings.