models-eval-playbook

Run model evaluations, diagnose MLflow runs, report stats, and compare models using the models-eval-playbook repo. Use when the user asks to evaluate an LLM, run the playbook, check eval results, diagnose MLflow run errors, compare models, or generate a report from experiment data. Triggers on mentions of running evals, playbook, MLflow run IDs, experiment IDs, model comparison, or eval suites (enterprise, hallucination, model, vision).