What Evaluations Do Not Do

• guarantee production safety on their own - they surface comparison results, contract checks, and diagnostics, but still require review
• replace monitoring - there is no built-in continuous evaluation loop; you schedule runs yourself
• make /api/runs/:id part of the public docs surface - evaluation results can reference underlying run ids, but public docs do not promote orchestration run detail as a released inspection API

See also

• Evaluations
• Errors and Diagnostics