Should AI agents show the extra data sent with requests?
A debate started after claims that Claude Code marks requests when a custom is used. The useful question is not whether this is spying or normal anti-abuse telemetry. The more practical question is whether AI should reveal in the same clear way they reveal .
show what an agent can access or do on a local machine. shows what the client adds when it sends work to the model. This matters more for AI agents than for ordinary apps because agents can read files, inspect code , run , call tools, and send local context to a remote model.
Metadata is not automatically bad. Vendors have real reasons to detect abuse, resale, model-copying attempts, suspicious gateways, policy problems, reliability issues, billing problems, and account boundary problems. Still, even if exact anti-abuse rules cannot be public, the whole request layer should not stay hidden.
Key points
- The debate began with claims about Claude Code and custom s.
- show what an agent can access or do locally.
- shows what extra information is attached when work is sent to a model.
- AI agents need more because they can handle files, , , tools, and local context.
- Vendors still need some private anti-abuse methods for safety, billing, and reliability.