feat(adk): improve skill middleware prompt for better LLM understanding

[hi-pender]

Summary

This PR optimizes the skill middleware prompts to improve LLM understanding and consistency.

Changes

1. Use English XML tags in Chinese prompts

• Changed Chinese XML tags (e.g., <技能>, <名称>) to English tags (<skill>, <name>)
• This improves LLM parsing and maintains technical consistency
• Removed duplicate toolDescriptionTemplateChinese variable since it's now identical to the English version

2. Use "Skill" terminology in Chinese prompts

• Replaced "技能" with "Skill" throughout Chinese prompts
• Added "Skill（技能）" explanation on first occurrence for clarity
• Maintains terminology consistency with the actual tool name

3. Clarify skill name usage

• Updated instruction to explicitly state: "Use the exact string inside <name> tag as the skill name"
• This helps when skill names contain special characters or paths

4. Add relative path conversion guidance

• Added clear instruction about converting relative paths to absolute paths using the base directory
• Moved this guidance to tool description (decoupled from skill content)

5. Remove product-specific rules

• Removed "Do not use this tool for built-in CLI commands" rule
• This rule was specific to certain products and not applicable to the generic skill middleware

Files Changed

• adk/middlewares/skill/prompt.go: Updated prompts and removed duplicate template
• adk/middlewares/skill/skill.go: Simplified renderToolDescription function

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 81.79%. Comparing base (0077549) to head (100b626).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #724      +/-   ##
==========================================
+ Coverage   81.77%   81.79%   +0.01%     
==========================================
  Files         129      129              
  Lines       12419    12416       -3     
==========================================
  Hits        10156    10156              
+ Misses       1508     1504       -4     
- Partials      755      756       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.