Add direct unit tests for `app/cli/local_llm/hardware.py`

[VaibhavUpreti]

Why This Matters

current local-LLM tests cover model recommendation heavily, but not the platform-specific hardware-detection helpers directly.

Read First

• app/cli/local_llm/hardware.py

Scope

1. Add direct tests for _get_total_ram_gb(), _get_available_ram_gb(), and detect_hardware().
2. Cover both Linux and macOS branches by patching sys.platform.
3. Test the conservative fallback behavior on command/file-read failure.

Tests / Coverage

• tests/cli/test_local_llm_hardware.py
• cover Linux /proc/meminfo parsing
• cover macOS sysctl parsing
• cover fallback RAM values
• cover NVIDIA detection via shutil.which

Acceptance Criteria

• hardware detection is directly covered
• tests do not depend on the real host OS
• fallback behavior is protected