Potential fix for code scanning alert no. 2: Information exposure through an exception

[Exohayvan]

Potential fix for https://github.com/AetherMesh-AI/aethermesh-core/security/code-scanning/2

To fix this safely without changing endpoint behavior, keep returning stable reason_code values but replace client-facing reason text derived from exceptions with generic, non-sensitive messages. Log detailed exception information on the server for debugging.

Best single approach here:

1. In src/aethermesh_core/runtime_service.py, add a module logger.
2. In preflight_local_result_report exception handlers:
   • Log the exception with logger.exception(...) (includes traceback in server logs).
   • Call _reject_local_result_report with a generic user-safe reason string instead of str(exc).
3. Keep existing persistence and response schema intact (status, attempt_id, reason_code, etc.) so functionality remains the same from an API-contract perspective, while removing sensitive exception content from responses.

This addresses both variants (line 1309 and 1311 exception sources), since both currently pass str(exc) to the returned rejection payload.

Suggested fixes powered by Copilot Autofix. Review carefully before merging.