We previously open-sourced a rubric for assessing APIs for agent-readiness (carefully designed by some colleagues who are deeply involved with the OpenAPI initiative). We hosted a nice web app for scoring an API according to this rubric (a mix of deterministic and LLM-based assessment), and learned people were hacking it into their CI/CD pipelines for regression testing and longitudinal tracking. So we just released a nice CLI to make that easier, with a decent free tier.