Hi, Thanks for the response. I understand that the on-device Foundation Models only access data explicitly provided via prompts or tool-calling implementations, and that they are not intended for code generation. That’s helpful. If external models are integrated into Apple Intelligence either via Private Cloud Compute or third-party APIs does the system ever scan the entire codebase automatically? Or does it still rely solely on developer-provided input? Best Regards, Soumen

Thanks for your understanding. Kindly note the feature request number: FB20976003