Add PEEK_LOSS functionality to Q# / QASM

In qdk-stim, `PEEK_LOSS` is an operation that checks whether a qubit was lost without collapsing the state, recording a One (in case of loss) or a Zero.

Currently, all of the adaptive simulators (clifford, CPU and GPU full state) support it, but we can only use it from stim or directly from qir. It seems like this is a behavior that mimics real-world architectures and might be useful for researchers. In that case, we should consider how to implement this functionality for Q# and QASM.

Start by reviewing the existing PEEK_LOSS behavior in the adaptive simulators and how Q# and QASM currently expose operations. Determine the integration points for both languages and define completion as usable PEEK_LOSS support in Q# and QASM, with behavior matching the existing operation.

Written by the indexing model from the issue text.