[bug] First-run launcher setup crashes before configuration

[zhongxuanwang-nv]

Symptom

First-run switchyard launch setup crashes before it can save launcher defaults.

Reproduction

switchyard launch claude --model openai/gpt-4o-mini \
  --api-key "$OPENROUTER_API_KEY" \
  --base-url "$OPENROUTER_BASE_URL"

Run with no saved Claude launch defaults so the setup path is entered.

Expected vs. actual

• Expected: Setup saves the provider, model, and credentials, then continues launching.
• Actual: The command fails with this traceback:

Switchyard is missing claude launch defaults. Starting setup.
Traceback (most recent call last):
  File "/Users/daniewang/Switchyard/.venv/bin/switchyard", line 10, in <module>
    sys.exit(main())
             ^^^^^^
  File "/Users/daniewang/Switchyard/switchyard/cli/switchyard_cli.py", line 1461, in main
    args.func(args)
  File "/Users/daniewang/Switchyard/switchyard/cli/switchyard_cli.py", line 390, in _cmd_launch_claude
    cmd_launch_claude(args)
  File "/Users/daniewang/Switchyard/switchyard/cli/launch_command.py", line 517, in cmd_launch_claude
    maybe_bootstrap_launch_config(
  File "/Users/daniewang/Switchyard/switchyard/cli/launch_command.py", line 303, in maybe_bootstrap_launch_config
    cmd_configure(configure_args)
  File "/Users/daniewang/Switchyard/switchyard/cli/configure_command.py", line 377, in cmd_configure
    skill_distillation = _apply_skill_distillation_args(
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/daniewang/Switchyard/switchyard/cli/configure_command.py", line 93, in _apply_skill_distillation_args
    if args.disable_skill_distillation:
       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AttributeError: 'Namespace' object has no attribute 'disable_skill_distillation'

Environment

• Switchyard version (or commit SHA): a2a939d
• Python version (python --version): 3.12.13
• OS / arch: macOS / arm64
• Install path (uv sync, pip install nemo-switchyard, source build, etc.): editable source install with uv
• Inbound format (Chat Completions / Anthropic Messages / Responses): N/A; failure occurs before launch
• Backend (OpenAI / Anthropic / NVIDIA Inference Hub / other): OpenRouter

Additional context

Fixed by #59.