[FEA]: Consider adding a static_assert to cuda::std::is_invocable that fails where its used in host code on an extended lambda

### Is this a duplicate?

- [X] I confirmed there appear to be no [duplicate issues](https://github.com/NVIDIA/cccl/issues) for this request and that I agree to the [Code of Conduct](CODE_OF_CONDUCT.md)

### Area

libcu++

### Is your feature request related to a problem? Please describe.

Similar to how cuda::std::invoke_result has a static_assert warning about accessing extended lambda return type from host where its not available, it would make sense to add such assert to cuda::std::is_invocable. Right now it just always returns true, no matter what argument types are passed into it

### Describe the solution you'd like

It would be nice if is_invocable would fail to compile when used in host code on an extended lambda

### Describe alternatives you've considered

_No response_

### Additional context

_No response_