ai-unity-asset-pipelinelisted

# AI to Unity asset pipeline Use the installed `ai-unity-asset` command. Keep provider generation, mesh optimization, Unity import, and live visual proof as separate evidence lanes. ## Required flow 1. Run `ai-unity-asset doctor --project "$PWD"`. 2. For a new concept, run `ai-unity-asset prompt` and use ImageGen to create four orthographic views on a plain background. Preserve the generated image, prompt, and fingerprint sidecar as immutable references. Read [references/imagegen-prompts.md](references/imagegen-prompts.md) for fixed prompt fields and bulk rules. For era production, run `ai-unity-asset prompt-batch`; when ChatGPT To Codex is available, consume its queue exactly as documented in [references/chatgpt2codex-asset-factory.md](references/chatgpt2codex-asset-factory.md). 3. Generate a mesh only through a provider the user has authorized. Never read or print API keys. Paid providers and gated model access require explicit approval. 4. Run Blender normalization: bottom-center pivot, meters, UV preservation, material consolidation, LOD0/LOD1/LOD2, full wheel clearance, and FBX plus GLB export. 5. Use `ai-unity-asset catalog` for broad 1950-2999 production. Add an archetype record whenever an existing modular recipe fits; add Python only when a genuinely new topology recipe is required. 6. Publish into Unity with `ai-unity-asset publish` or use the complete `ai-unity-asset fleet` flow for the bundled Transport Tycoon vehicle f