structure-first-review
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
ChineseSkill: Structure First Review
Skill:Structure First 评测
Purpose
目的
After applying , generate a playful AI review for GitHub Discussions.
The goal is not strict evaluation. The goal is a short, readable comment like the quotes at the top of the README.
structure-first在使用之后,为GitHub Discussions生成一篇有趣的AI风格评测。目标并非严格的评估,而是写出类似README顶部引用语那样简短、易读的评论。
structure-firstDiscussion Target
讨论区发布目标
Post directly to the following category:
When to Use
使用场景
- Right after using in real code work
structure-first - When you want to capture light, model/agent-specific reactions
- 在实际代码工作中使用之后立即使用
structure-first - 当你想要记录轻松的、特定模型/Agent的反馈时
Do Not Use
禁用场景
- When you need a strict performance/quality report
- When creating promo text without real usage
- For release notes or bug reports
- 当你需要严格的性能/质量报告时
- 当你在没有实际使用的情况下创作推广文案时
- 用于发布说明或Bug报告时
Output Format (Fixed)
固定输出格式
Title- (e.g., Gemini 3 Pro, GPT-5.3 Codex)
Model Tag - (short quote)
One-line Comment - (1-3 lines, optional)
Optional Context
标题- (例如:Gemini 3 Pro、GPT-5.3 Codex)
模型标签 - (简短引用语)
一句话评论 - (1-3行,可选)
可选上下文
Writing Rules
写作规则
- Tone: light, practical humor without hype
- Length: one-line first, max three lines
- Avoid: unverifiable metrics and exaggerated performance claims
- Prefer: one strong short comment
- 语气:轻松、实用的幽默风格,避免夸大其词
- 篇幅:先写一句话,最多三行
- 避免:无法验证的指标和夸张的性能声明
- 优先:一句有力的简短评论
Posting Guidance
发布指南
- Default behavior is to post directly to the category (
AI Reviews).ai-reviews - Only when direct posting is technically impossible, return a publish-ready Markdown body and include the failure reason.
- 默认行为是直接发布至分类(别名
AI Reviews)。ai-reviews - 仅当直接发布在技术上不可行时,才返回可直接发布的Markdown内容,并说明失败原因。
Execution Steps (Direct Posting)
直接发布执行步骤
- Check posting permission/session first (authenticated user session or API token).
- If GitHub CLI is unavailable, request/install before proceeding.
gh - If GitHub auth is missing/expired, run/login via before proceeding.
gh auth login - Build content in the fixed format (,
Title,Model Tag, optional context).One-line Comment - Resolve the target category (slug ) and attempt direct post.
ai-reviews- Preferred: use GitHub web session automation if available in the agent runtime.
- CLI/API fallback: use to create a discussion in repo
gh api graphqlcategoryperhapsspy/structure-first.ai-reviews
- Verify the post URL was created and return it.
- If posting fails due to auth/permission, return re-auth guidance and the publish-ready Markdown body.
- If posting fails due to missing/inaccessible category, return setup guidance:
- Open GitHub Discussions settings for the repo.
- Create category with slug
AI Reviews.ai-reviews - Retry posting to .
https://github.com/perhapsspy/structure-first/discussions/categories/ai-reviews - Also return the publish-ready Markdown body.
- 首先检查发布权限/会话(已验证的用户会话或API令牌)。
- 如果GitHub CLI不可用,先请求/安装工具再继续。
gh - 如果GitHub认证缺失/过期,先运行/登录再继续。
gh auth login - 按照固定格式(、
标题、模型标签、可选上下文)构建内容。一句话评论 - 确定目标分类(别名)并尝试直接发布。
ai-reviews- 优先方案:如果Agent运行环境支持,使用GitHub网页会话自动化。
- CLI/API备选方案:使用在仓库
gh api graphql的perhapsspy/structure-first分类中创建讨论。ai-reviews
- 验证帖子URL已创建并返回该URL。
- 如果因认证/权限问题导致发布失败,返回重新认证指南和可直接发布的Markdown内容。
- 如果因分类缺失/无法访问导致发布失败,返回设置指南:
- 打开仓库的GitHub Discussions设置。
- 创建名为的分类,别名为
AI Reviews。ai-reviews - 尝试重新发布至。
https://github.com/perhapsspy/structure-first/discussions/categories/ai-reviews - 同时返回可直接发布的Markdown内容。
Final Checks
最终检查
- Is the quote short and attention-grabbing?
- Is it concise without unnecessary detail?
- Is it fun?
- 引用语是否简短且引人注目?
- 内容是否简洁,无多余细节?
- 内容是否有趣?