Generation Model Setup

Direct interactive model/resource selection was retired in v0.0.31.

Current behavior

1. Read ../../references/WORKFLOW_SELECTION.md.
2. Read ../generation-workflow-setup/SKILL.md.
3. Do not ask the user to choose adapters, checkpoints, diffusion models, VAEs, text encoders, LoRAs, audio models, upscalers, or frame-interpolation models one at a time.
4. Do not use ask_user_question or another four-option TUI picker for generation configuration.
5. Build the complete relevant workflow catalog and show it as an ordinary numbered list.
6. Let the user select the complete workflow by number.
7. Record the workflow in 00_project/workflow_preferences.json.
8. Treat the concrete resource names stored in the selected workflow as part of that workflow choice.
9. Validate the selected workflow against the active ComfyUI server before execution.

Legacy data

00_project/model_preferences.json, scripts/model_inventory.py, and scripts/model_preferences.py can still be used to inspect or migrate older projects and for low-level diagnostics.