[Feature Request] Add image generation/editing capability for UI asset extraction from mockups

Bug Description

## Summary Claude Code's coding ability is outstanding — but it currently has no image generation/editing capability, which blocks a real game-development workflow that competing tools (OpenAI Codex) already handle well. ## Use case I'm developing a Cocos2d-x casino game. A common task: take a flattened UI mockup (e.g. a "NEWS" popup screenshot with neon frames, tabs, buttons, stylized logo text) and split it into clean, alpha-transparent PNG assets ready for production use in CocosStudio — backplate 9-slice frames, tab states, list buttons, close button, scrollbar, icons, logo art. ## What happened - Claude correctly identified and decomposed every UI component from the mockup (excellent analysis: component tree, 9-slice suggestions, asset list). - But when asked to actually produce the cutout PNGs, it could only offer procedural drawing (PIL gradients/shapes). The results were usable for simple geometric pieces but clearly below production quality for illustrative art (stylized logo text, 3D-ish cartoon icons). - Codex's image editing feature produced clean, production-ready transparent cutouts from the same mockup (regenerated assets with proper alpha edges). ## Request Give Claude Code image generation/editing capability (or an official integration) so it can: 1. Re-generate individual UI elements from a flattened mockup as clean transparent PNGs (background removal + asset regeneration). 2. Match the style/palette of the source mockup. 3. Output at specified sizes suitable for game engines (9-slice aware would be a bonus). Claude already does the analysis half of this workflow better than anything else — it just can't finish the job. Closing this gap would make Claude Code a complete tool for game UI development. ## Environment - Claude Code on Windows 11 (desktop app), Opus 4.8 / Fable 5 - Project: Cocos2d-x 3.x + CocosStudio assets

Environment Info

• Platform: win32
• Terminal: windows-terminal
• Version: 2.1.219
• Feedback ID: 69e9480f-af3d-4750-8e58-6729ebf1190f

Errors

[]