[babyshake]

One aspect the report is very vague about is the nature of the monitoring Anthropic is doing on Claude Code. If they can detect attacks they can surely detect other things of interest (or value) to them. Is there any more information about this?