Measuring AI agent autonomy: Towards a scalable approach with code inspection

Open in new window