爆款拆解：实战 Skill 精选

你：帮我把这段话改得不那么像AI写的

AI：[自动加载 avoid-ai-writing Skill]
    检测到以下 AI 痕迹：
    - "delve into" → 建议改为 "explore" 或 "dig into"
    - "it's important to note that" → 建议删除，直接陈述事实
    - 段落长度过于均匀，建议长短交替
    
    改写版本：
    （输出自然流畅的人工风格文本）

name: avoid-ai-writing
description: Audit and rewrite content to remove AI writing patterns ("AI-isms"). 
  Use this skill when asked to "remove AI-isms," "clean up AI writing," 
  "edit writing for AI patterns," "audit writing for AI tells," 
  or "make this sound less like AI."

## Words and phrases to replace

Words are organized into three tiers based on how reliably they signal AI-generated text.

- **Tier 1 — Always flag.** These words appear 5–20x more often in AI text than human text. Replace on sight.
- **Tier 2 — Flag in clusters.** Individually fine, but two or more in the same paragraph is a strong AI signal.
- **Tier 3 — Flag by density.** Common words that AI simply overuses. Only flag when they make up a noticeable fraction of the text.

## Context profiles

Pass an optional context hint to adjust rule strictness.

| Rule | linkedin | blog | technical-blog | investor-email | docs | casual |
|------|----------|------|----------------|----------------|------|--------|
| Em dashes | relaxed (2/post OK) | strict | strict | strict | relaxed | skip |
| Bold overuse | relaxed (bold hooks OK) | strict | strict | strict | relaxed | skip |
| Word table (full list) | strict | strict | partial | strict | relaxed | P0 only |

## Severity tiers

### P0 — Credibility killers (fix immediately)
- Cutoff disclaimers ("As of my last update")
- Chatbot artifacts ("I hope this helps!", "Great question!")
- Vague attributions without sources ("Experts believe")

### P1 — Obvious AI smell (fix before publishing)
- Word-list violations (delve, leverage, harness, robust, etc.)
- Template phrases and slot-fill constructions
- "Let's" transition openers

### P2 — Stylistic polish (fix when time allows)
- Generic conclusions ("The future looks bright")
- Compulsive rule of three
- Uniform paragraph length

## Output format

Return your response in four sections:

**1. Issues found** — A bulleted list of every AI-ism identified
**2. Rewritten version** — The full rewritten content
**3. What changed** — A brief summary of major edits
**4. Second-pass audit** — Re-read and identify any remaining AI tells

你：转小红书 https://mp.weixin.qq.com/xxxxx

AI：[加载 content-pipeline Skill]
    已抓取文章《早起打卡如何改变你的一天》
    正在生成多平台内容...
    
    ✅ 小红书轮播图 HTML（10 张卡片）
    ✅ 小红书发布文案（300 字 + 标签）
    ✅ 即刻发布文案
    ✅ 小宇宙播客脚本（15 分钟）
    ✅ 播客封面 HTML
    ✅ manifest.json（供一键分发）

## 两条输入路径

### Path A：日常素材收集 → 出稿
边干活边记录 → 说"出稿" → 写文章 → 排版 → 封面图 → 朋友圈文案 → manifest

### Path B：微信链接 → 多平台内容
微信链接 → 抓取文章 → 分析结构 → 生成小红书/即刻/播客/视频 → manifest → 分发

## Reference 文件索引

| 场景 | 读取文件 |
|------|----------|
| 出稿写说明书文章 | `references/manual-framework.md` |
| 出稿写深度长文 | `references/writing-style.md` |
| 出稿写教程文章 | `references/tutorial-framework.md` |
| 生成头图/配图 | `references/cover-template.md` |
| 生成小红书轮播图 | `references/xiaohongshu-format.md` |

按需读取，不要一次性加载所有 reference。

## 触发词

### 素材收集（Path A）
| 触发词 | 说明 |
|-------|------|
| `/story` | 查看当前素材状态 |
| "出稿" | 生成文章 + 排版 + 封面图 |
| "写个朋友圈" | 根据素材生成朋友圈文案 |

### 内容生成（Path B）
| 触发词 | 说明 |
|-------|------|
| `/xiaohongshu` + 链接 | 微信文章转小红书轮播图 |
| "转小红书" + 微信链接 | 同上 |
| "做成小红书" + 微信链接 | 同上 |

## Script Directory

| Script | Purpose |
|--------|---------|
| `scripts/fetch_wechat_article.py` | 微信文章抓取（Python，模拟微信UA） |
| `scripts/md2wechat_formatter.py` | Markdown转公众号HTML排版 |
| `scripts/distribute/distribute.ts` | 分发主编排器 |
| `scripts/distribute/platforms/*.ts` | 各平台发布模块 |

### 抓取文章
```bash
python3 "${SKILL_DIR}/scripts/fetch_wechat_article.py" "<URL>" --json

**值得借鉴的分工模式**：

| 层级 | 职责 | 示例 |
|------|------|------|
| SKILL.md | 决定"什么时候做" | 第1步抓取、第2步分析... |
| scripts/ | 决定"怎么做" | Python抓取、TypeScript分发 |

**聪明做法**：使用 `${SKILL_DIR}` 变量让脚本路径可移植，不管Skill装在哪都能找得到。

---

#### 技巧5：四级降级设计——自动化不是"全或无"

**原文引用**：

```markdown
## 四级降级

| 级别 | 模式 | 触发条件 |
|------|------|---------|
| L0 | API直推 | 公众号API直接推草稿箱，无需Chrome |
| L1 | 自动发布 | CDP完全自动化 |
| L2 | 辅助发布 | 登录态失效/选择器失效/`--preview` |
| L3 | 手动模式 | CDP连接失败 |

公众号优先L0（API），凭证缺失或失败时自动降级L3（手动）。

xiaohongshu-skills/
├── skills/
│   ├── 01-内容创作/           ← 23个Skill
│   ├── 02-账号运营/           ← 20个Skill
│   ├── 03-互动运营/           ← 8个Skill
│   ├── 03-数据分析/           ← 12个Skill
│   ├── 04-电商转化/           ← 17个Skill
│   ├── 05-平台规则/           ← 14个Skill
│   ├── 06-工具生态/           ← 14个Skill
│   ├── 07-营销推广/           ← 16个Skill
│   └── 08-增长策略/           ← 15个Skill

---
name: title-writing
description: Use when writing Xiaohongshu post titles, creating headlines for content, 
  improving click-through rates, or optimizing titles for search and discovery
---

## When to Use

**Use when**:
- Writing titles for new posts
- Optimizing low-performing content
- A/B testing headline variants

**Do NOT use when**:
- Writing full content body (use copywriting-skills)
- Creating cover text (use cover-design)

## 渐进式加载

**第一层（启动时）**：AI扫描139个Skill的name + description，
约139 × 80 = 11120 tokens。这是固定开销，每次对话都要加载。

**第二层（匹配时）**：用户说"帮我写个标题"，AI从139个
description中匹配到 `title-writing`，只加载这一个Skill的正文（约250 tokens）。

**第三层（执行时）**：如果AI需要参考更多案例，可以读取
references/下的模板文件（但title-writing本身不需要）。

## Related Skills

- **REQUIRED**: cover-design (cover text = title)
- **REQUIRED**: content-planning (title matches content type)
- copywriting-skills (write compelling body)
- hashtag-optimization (add tags for discovery)

## 技能层次

| 层次 | 说明 | 占比 |
|------|------|------|
| **Tier 1 核心** | 必备基础技能 | ~30% |
| **Tier 2 进阶** | 重要提升技能 | ~40% |
| **Tier 3 高级** | 专项深度技能 | ~20% |
| **Tier 4 专家** | 战略级技能 | ~10% |

# 查看 avoid-ai-writing 的完整规则
cat .claude/skills/avoid-ai-writing/SKILL.md

# 查看 content-pipeline 的工作流编排
cat .claude/skills/content-pipeline/SKILL.md | head -100

# 查看 xiaohongshu-skills 的分类结构
ls .claude/skills/xiaohongshu-skills/skills/

# 查看单个子Skill的设计
cat .claude/skills/xiaohongshu-skills/skills/01-内容创作/title-writing/SKILL.md