[Experimental] Ivan Human UI: Anti-AI-Vibe skill for Codex

[flychicken067]

Summary (EN)

• add new experimental skill ivan-human-ui
• package a concrete anti-AI-vibe UI polish workflow for Codex
• include measurable eval loop (prompts, rubric, results log, and now 2 real eval reports: case-b + case-c)
• include deterministic verifier scripts for testimonial tracks and dashboard polish checks
• include contributor-friendly README and direct install example

摘要（中文）

• 新增实验技能 ivan-human-ui
• 提供一套可执行的“去 AI 味”UI 打磨流程（适配 Codex）
• 内置可量化评测闭环（prompts、rubric、结果记录），并已补 2 个真实评测案例（case-b + case-c）
• 提供可重复验证脚本（testimonial 轨道可见性 + dashboard 打磨约束）
• 补充了贡献者友好的 README 与安装示例

Install Example

$skill-installer install https://github.com/openai/skills/tree/main/skills/.experimental/ivan-human-ui

Maintenance

I can maintain this skill after merge and respond quickly to follow-up review comments or fixes.

Path

skills/.experimental/ivan-human-ui

[flychicken067]

Maintainers, could you please review this experimental skill contribution when you have a moment?

Highlights:

• skill path: skills/.experimental/ivan-human-ui
• includes concrete constraints + eval loop + real eval evidence
• includes deterministic testimonial visibility checker

If anything should be adjusted for openai/skills standards, I can update quickly. Thank you!

维护者您好，方便的话请帮忙 review 这个 experimental skill 贡献。
如需按仓库规范调整，我可以马上跟进修改，感谢！

[flychicken067]

Quick update: I just pushed an additional real eval case (case-b) with deterministic verifier evidence, and updated the PR description with install + maintenance details.

Could maintainers please take a look when available? @alistair-openai @vb-openai

If this should be split/re-scoped to match openai/skills standards, I can revise quickly. Thanks!

[flychicken067]

Superseded by #316 (curated path): skills/.curated/ivan-human-ui. Closing this experimental PR to keep one review thread and reduce maintainer overhead. Happy to re-open if preferred.