fix(i18n): sync dashscope-image-gen & xiaomi-tts zh-CN body with local (#13)

## Summary

- 用本地 `defaults/global-skills/` 中的最新中文 body 内容更新远程 `SKILL.zh-CN.md`
- 更新 zh-CN `source_hash` 以反映最新内容
- 解锁 en-US `translated_by`，允许 CI 自动重新翻译英文 body

## 变更文件

- `skills/dashscope-image-gen/SKILL.zh-CN.md` — 中文 body 同步至本地最新版本
- `skills/dashscope-image-gen/SKILL.md` — 更新 source_hash + 解锁 en-US
- `skills/xiaomi-tts/SKILL.zh-CN.md` — 中文 body 同步至本地最新版本
- `skills/xiaomi-tts/SKILL.md` — 更新 source_hash + 解锁 en-US

skills/dashscope-image-gen/SKILL.md

@@ -38,14 +38,14 @@ metadata:
       description: >-
         当用户希望使用阿里云 DashScope 的通义万相系列模型生成图片时使用此技能。支持多种模型层级（wan2.7-image-pro / wan2.7-image）的文生图，通过 OpenAI 兼容的 chat/completions API 同步生成图片。用户提到 生成图片、画图、文生图、创建图片、AI 绘画、生成插图、画一张、帮我画、设计图片、通义万相、万相、阿里云画图、dashscope 画图。
       body: ./SKILL.zh-CN.md
-      source_hash: sha256:d24415cd18ebf5d2
+      source_hash: sha256:135b99cdd33441fb
       translated_by: human
     en-US:
       name: DashScope Image Generation
       short_desc: Text-to-image generation using Alibaba Cloud Wan (通义万相) models
       description: "Use this skill when the user wants to generate images using Alibaba Cloud DashScope's Wan (通义万相) series models. Supports text-to-image with multiple model tiers (wan2.7-image-pro, wan2.7-image) via the OpenAI-compatible chat/completions API. Trigger keywords: generate image, draw, text-to-image, create image, AI painting, illustration, design picture, Wan, Tongyi Wanxiang, DashScope."
       body: ./SKILL.md
-      source_hash: sha256:d24415cd18ebf5d2
+      source_hash: sha256:135b99cdd33441fb
       translated_by: human
 market:
   icon: >-