メニュー
Quick photo essay pipeline for hot topics. Produces ~600 word image-driven articles using Claude Opus + Kimi K2.5 refinement. 7-step fast workflow.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Create distinctive, production-grade hackathon demo decks with bold design choices that stand out from typical technical presentations. Use when building pitch decks for hackathons, demo days, or competition submissions that need to combine technical depth with visual polish.
Generate card-based architecture diagrams as standalone HTML files. Warm glassmorphism style with CSS-only arrows, rounded cards, and backdrop-filter effects. Use when the user asks to create a pipeline, flow, or architecture diagram in card style.
Full hottrend article pipeline. Aggregates sources, drafts, reviews, refines, generates images, and packages for WeChat publication. 11-step multi-agent workflow.
Plan a series of hottrend articles. Interactive discussion to research topic, propose structure, and generate plan.md. Phase A of the series workflow.
Generate a single article in a hottrend series. Reads plan.md, creates context, runs full hottrend pipeline with series context injection. Phase B of series workflow.
A Stop hook that catches Claude when it claims a UI change is "done", "deployed", or "verified" without showing evidence that a real-browser UI test actually passed. Stops the curl + grep illusion of testing.
| name | hottrend-quick |
| description | Quick photo essay pipeline for hot topics. Produces ~600 word image-driven articles using Claude Opus + Kimi K2.5 refinement. 7-step fast workflow. |
从热点话题或新闻链接出发，通过 7 步精简 pipeline 快速产出 ~600 字 Photo Essay 风格图文速报。
/hottrend-quick <topic_or_urls> [options]
Examples:
/hottrend-quick "GPT-5 发布"
/hottrend-quick https://url1.com https://url2.com
/hottrend-quick "Apple WWDC 2026" --no-publish
执行此 skill 时，必须读取并严格按照以下 PROMPT 文件执行：
/home/ubuntu/codes/ai-writing/PROMPT-hottrend-quick.md
Step 1: 创建项目目录 (articles/YYYYMMDD-HHMMSS-hottrend-quick/)
Step 2: hottrend-aggregator → raw/aggregated.md + images/
Step 3: hottrend-quick-draft(Opus) → output/v1_quick.md
Step 4: hottrend-quick-refiner(Kimi K2.5) → output/v2_kimi.md
Step 5: title-generator(Opus) → 5 个备选标题
Step 6: de-ai-filter(Opus) → output/v3_final.md
Step 7: wechat-mcp → 公众号草稿箱
| Step | Model | Reason |
|---|---|---|
| quick-draft | Claude Opus | 初稿生成质量高 |
| quick-refiner | Kimi K2.5 | 口语化润色最自然 |
| title/de-ai | Claude Opus | 质量把关 |
| Step | Agent | Role |
|---|---|---|
| 2 | hottrend-aggregator | 多源抓取聚合（降低 Quality Gate） |
| 3 | hottrend-quick-draft | Photo Essay 风格初稿，图片驱动叙事 |
| 4 | hottrend-quick-refiner | Kimi K2.5 润色（via invoke_model.py） |
| 对比 | /hottrend-article | /hottrend-quick |
|---|---|---|
| 步骤 | 11 步 | 7 步 |
| 字数 | 2000-3000 字 | 500-700 字 |
| 图片 | 3-6 张 | 5-8 张 |
| 风格 | 深度综述 | Photo Essay 图文速报 |
| 模型 | 全 Claude Opus | Opus + Kimi K2.5 |
| 砍掉 | — | reader-critic, analyzer, refiner, diagram-gen, image-gen, wx-packager |
articles/YYYYMMDD-HHMMSS-hottrend-quick/
├── images/ # 抓取的图片
├── raw/
│ ├── aggregated.md # 聚合素材
│ └── source_*.md # 各来源原文
├── output/
│ ├── images -> ../images # 符号链接
│ ├── v1_quick.md # 图文速报初稿 (Opus)
│ ├── v2_kimi.md # Kimi K2.5 润色稿
│ └── v3_final.md # 最终稿
└── process/
├── pipeline.log # 执行日志
└── step4_input.md # Kimi 调用输入