# 知识库维护系统 - 详细设计

> 版本: 1.0.0  
> 状态: 草稿

---

## 目录

1. [技能接口总览](#1-技能接口总览)
2. [技能1-link-parser](#2-技能1-link-parser)
3. [技能2-bilibili-metadata](#3-技能2-bilibili-metadata)
4. [技能2-zhihu-metadata](#4-技能2-zhihu-metadata)
5. [技能3-bilibili-content](#5-技能3-bilibili-content)
6. [技能3-ocr-processor](#6-技能3-ocr-processor)
7. [技能4-deep-analyzer](#7-技能4-deep-analyzer)
8. [技能5-kb-workflow](#8-技能5-kb-workflow)
9. [数据格式定义](#9-数据格式定义)
10. [配置文件规范](#10-配置文件规范)
11. [扩展指南](#11-扩展指南)

---

## 1. 技能接口总览

### 1.1 调用约定

所有技能遵循以下调用约定：

```yaml
输入: 
  - 命令行参数
  - stdin JSON (可选)
  - 环境变量

输出:
  - stdout JSON
  - 0: 成功
  - 1: 失败 (错误信息输出到 stderr)

状态码:
  - 0: 成功
  - 1: 一般错误
  - 2: 参数错误
  - 3: 网络错误
  - 4: 权限错误
```

### 1.2 通用错误响应

```json
{
  "success": false,
  "error": "错误描述",
  "error_code": "ERROR_CODE",
  "details": {}
}
```

---

## 2. 技能1: link-parser

**功能**: 解析链接，提取平台+类型+ID，处理短链跳转

### 2.1 命令行接口

```bash
# 基本用法
python link-parser.py --url "https://www.bilibili.com/video/BV1xx411c7mD"

# 支持短链
python link-parser.py --url "https://b23.tv/xxxxx"

# 批量处理
python link-parser.py --batch input.txt

# 输出文件
python link-parser.py --url "..." --output result.json
```

### 2.2 输入参数

```yaml
参数:
  --url, -u     # 必需，单个URL
  --batch, -b   # 可选，批量文件（每行一个URL）
  --output, -o  # 可选，输出文件路径
  --format      # 输出格式: json/yaml (默认: json)
```

### 2.3 输出格式

**成功**:
```json
{
  "success": true,
  "data": {
    "platform": "bilibili",
    "type": "video",
    "id": "BV1xx411c7mD",
    "source_url": "https://www.bilibili.com/video/BV1xx411c7mD",
    "canonical_url": "https://www.bilibili.com/video/BV1xx411c7mD",
    "short_link_resolved": false,
    "title": null,
    "metadata": {}
  },
  "metadata": {
    "processed_at": "2024-01-15T10:00:00",
    "parser_version": "1.0.0"
  }
}
```

**列表类型**:
```json
{
  "success": true,
  "data": {
    "platform": "bilibili",
    "type": "up",
    "id": "7770724",
    "source_url": "https://space.bilibili.com/7770724",
    "is_list_type": true,
    "list_info": {
      "total_items": 156,
      "item_type": "video"
    }
  },
  "metadata": {
    "requires_user_input": true,
    "input_options": {
      "scope": ["subscribe_only", "new_only", "all"],
      "subscription": ["daily", "weekly", "custom", "none"]
    }
  }
}
```

### 2.4 支持的链接格式

| 类型 | 平台 | 正则示例 | 输出 |
|------|------|----------|------|
| 短链 | B站 | `b23.tv/[a-zA-Z0-9]+` | resolved_url + 解析 |
| 视频 | B站 | `bilibili.com/video/BV[a-zA-Z0-9]+` | `{type: video, id: BV...}` |
| 专栏 | B站 | `bilibili.com/read/cv[0-9]+` | `{type: article, id: cv...}` |
| 动态 | B站 | `bilibili.com/opus/[0-9]+` | `{type: opus, id: ...}` |
| UP主 | B站 | `space.bilibili.com/[0-9]+` | `{type: up, id: ...}` |
| 收藏夹 | B站 | `space.bilibili.com/[0-9]+/favlist?fid=[0-9]+` | `{type: favlist, id: ...}` |
| 合集 | B站 | `space.bilibili.com/[0-9]+/lists/[0-9]+?type=season` | `{type: playlist, id: ...}` |
| YouTube | YouTube | `(youtu.be/[a-zA-Z0-9]+|youtube.com/watch?v=[a-zA-Z0-9]+)` | `{type: video, id: ...}` |
| 知乎 | 知乎 | `zhihu.com/question/[0-9]+` | `{type: question, id: ...}` |
| 知乎 | 知乎 | `zhihu.com/answer/[0-9]+` | `{type: answer, id: ...}` |
| 知乎 | 知乎 | `zhihu.com/people/[a-zA-Z0-9-]+` | `{type: user, id: ...}` |
| 知乎 | 知乎 | `zhihu.com/collection/[0-9]+` | `{type: collection, id: ...}` |

### 2.5 复用代码

```python
# 复用现有代码
from shortlink_resolver import resolve_shortlink, extract_bilibili_id
```

---

## 3. 技能2: bilibili-metadata

**功能**: 反爬获取B站视频/专栏元数据

### 3.1 命令行接口

```bash
# 基本用法
python bilibili-metadata.py --id "BV1xx411c7mD"

# 指定输出目录
python bilibili-metadata.py --id "BV1xx" --output ./metadata/

# 批量处理
python bilibili-metadata.py --batch ids.txt --output ./

# 视频+专栏支持
python bilibili-metadata.py --id "cv43780186" --type "article"
```

### 3.2 输入参数

```yaml
参数:
  --id, -i        # 必需，内容ID
  --type, -t      # 内容类型: video/article/opus (默认: video)
  --output, -o    # 输出目录 (默认: ./)
  --batch, -b     # 批量文件 (每行: id [type])
  --config, -c    # 配置文件路径
```

### 3.3 输出格式

**视频元数据**:
```json
{
  "success": true,
  "data": {
    "platform": "bilibili",
    "type": "video",
    "id": "BV1xx411c7mD",
    "source_url": "https://www.bilibili.com/video/BV1xx411c7mD",
    "title": "Python完整教程",
    "description": "这是一个Python教程...",
    "author": {
      "id": 7770724,
      "name": "UP主名字",
      "face": "https://i0.hdslb.com/xxx.jpg"
    },
    "duration": 9000,
    "duration_text": "2:30:00",
    "publish_time": "2024-01-01T12:00:00",
    "tags": ["python", "tutorial", "编程"],
    "stats": {
      "view": 100000,
      "like": 5000,
      "coin": 2000,
      "favorite": 1000,
      "share": 500
    },
    "subtitle": {
      "has_subtitle": true,
      "languages": ["zh-CN"],
      "url": "https://api.bilibili.com/..."
    },
    "pic": "https://i0.hdslb.com/xxx.jpg",
    "quality_score": null
  },
  "metadata": {
    "fetched_at": "2024-01-15T10:00:00",
    "source": "bilibili_api"
  }
}
```

**专栏文章元数据**:
```json
{
  "success": true,
  "data": {
    "platform": "bilibili",
    "type": "article",
    "id": "cv43780186",
    "source_url": "https://www.bilibili.com/read/cv43780186",
    "title": "文章标题",
    "description": "文章摘要...",
    "author": {
      "id": 123456,
      "name": "作者名字",
      "face": "https://i0.hdslb.com/xxx.jpg"
    },
    "publish_time": "2024-01-01T12:00:00",
    "stats": {
      "view": 50000,
      "like": 1000,
      "comment": 200
    },
    "word_count": 5000,
    "images_count": 10,
    "category": "技术"
  },
  "metadata": {
    "fetched_at": "2024-01-15T10:00:00",
    "source": "bilibili_api"
  }
}
```

**UP主/收藏夹列表**:
```json
{
  "success": true,
  "data": {
    "platform": "bilibili",
    "type": "up",
    "id": "7770724",
    "source_url": "https://space.bilibili.com/7770724",
    "name": "UP主名字",
    "description": "UP主介绍...",
    "follower": 100000,
    "following": 100,
    "archive_count": 156,
    "content_stats": {
      "total_views": 10000000,
      "total_likes": 500000
    },
    "subscription": null
  },
  "metadata": {
    "fetched_at": "2024-01-15T10:00:00"
  }
}
```

### 3.4 复用代码

```python
# 复用现有代码
from bilibili_extractor import BilibiliExtractor
```

---

## 4. 技能2: zhihu-metadata

**功能**: 获取知乎元数据

### 4.1 命令行接口

```bash
python zhihu-metadata.py --id "266309957" --type "question"
python zhihu-metadata.py --id "2003427891925509507" --type "answer"
```

### 4.2 输入参数

```yaml
参数:
  --id, -i        # 必需，内容ID
  --type, -t      # 内容类型: question/answer/article/user/collection (默认: question)
  --output, -o    # 输出目录
  --config, -c    # 配置文件路径
```

### 4.3 输出格式

**问题元数据**:
```json
{
  "success": true,
  "data": {
    "platform": "zhihu",
    "type": "question",
    "id": "266309957",
    "source_url": "https://www.zhihu.com/question/266309957",
    "title": "问题标题",
    "description": "问题描述...",
    "author": {
      "id": "user-xxx",
      "name": "作者名字",
      "url_token": "user-xxx"
    },
    "answer_count": 50,
    "followers": 1000,
    "browse_count": 50000,
    "created_time": "2024-01-01T12:00:00",
    "updated_time": "2024-01-15T10:00:00",
    "tags": ["标签1", "标签2"]
  },
  "metadata": {
    "fetched_at": "2024-01-15T10:00:00"
  }
}
```

**回答元数据**:
```json
{
  "success": true,
  "data": {
    "platform": "zhihu",
    "type": "answer",
    "id": "2003427891925509507",
    "source_url": "https://www.zhihu.com/question/266309957/answer/2003427891925509507",
    "question_id": "266309957",
    "question_title": "问题标题",
    "author": {
      "id": "user-xxx",
      "name": "作者名字",
      "headline": "作者简介"
    },
    "vote_up": 1000,
    "comment_count": 50,
    "created_time": "2024-01-01T12:00:00",
    "updated_time": "2024-01-15T10:00:00",
    "content_summary": "回答摘要..."
  },
  "metadata": {
    "fetched_at": "2024-01-15T10:00:00"
  }
}
```

### 4.4 复用代码

```python
# 复用现有代码
from zhihu_extractor import ZhihuDrissionExtractor
```

---

## 5. 技能3: bilibili-content

**功能**: 下载低分视频 + 提取字幕/专栏全文

### 5.1 命令行接口

```bash
# 基本用法（视频）
python bilibili-content.py --id "BV1xx411c7mD" --output ./content/

# 视频+字幕
python bilibili-content.py --id "BV1xx" --output ./ --with-subtitle

# 专栏全文
python bilibili-content.py --id "cv43780186" --type article --output ./

# 低分辨率（用于OCR）
python bilibili-content.py --id "BV1xx" --quality 480 --output ./
```

### 5.2 输入参数

```yaml
参数:
  --id, -i        # 必需，内容ID
  --type, -t      # 内容类型: video/article/opus (默认: video)
  --output, -o    # 输出目录 (默认: ./)
  --quality       # 视频质量: 360/480/720/1080 (默认: 480)
  --with-subtitle # 是否提取字幕 (默认: true)
  --subtitle-lang # 字幕语言 (默认: zh-CN)
  --format        # 输出格式: json/ndjson (默认: json)
```

### 5.3 输出格式

**视频内容**:
```json
{
  "success": true,
  "data": {
    "platform": "bilibili",
    "type": "video",
    "id": "BV1xx411c7mD",
    "video_file": {
      "path": "content/BV1xx_video.mp4",
      "size_bytes": 52428800,
      "duration": 9000,
      "quality": "480p"
    },
    "subtitle": {
      "format": "srt",
      "path": "content/BV1xx_subtitle.srt",
      "language": "zh-CN",
      "text_length": 50000,
      "lines_count": 1500
    },
    "danmaku": {
      "path": "content/BV1xx_danmaku.xml",
      "count": 5000
    },
    "frames": {
      "output_dir": "content/BV1xx_frames",
      "count": 900,
      "fps": 0.1
    }
  },
  "metadata": {
    "fetched_at": "2024-01-15T10:00:00",
    "content_quality": "preview"
  }
}
```

**专栏内容**:
```json
{
  "success": true,
  "data": {
    "platform": "bilibili",
    "type": "article",
    "id": "cv43780186",
    "content": {
      "markdown": "# 文章标题\n\n内容...",
      "markdown_path": "content/cv43780186.md",
      "html": "<html>...</html>",
      "word_count": 5000,
      "images": [
        {
          "url": "https://...",
          "path": "content/images/img_001.jpg",
          "description": "图片描述"
        }
      ]
    },
    "author": {
      "name": "作者名字"
    },
    "publish_time": "2024-01-01T12:00:00"
  },
  "metadata": {
    "fetched_at": "2024-01-15T10:00:00"
  }
}
```

### 5.4 复用代码

```python
# 复用现有代码
from video_stream_extractor import VideoStreamExtractor
from bilibili_extractor import BilibiliExtractor
```

---

## 6. 技能3: ocr-processor

**功能**: 通用字幕OCR（被技能3调用）

### 6.1 命令行接口

```bash
# 基本用法
python ocr-processor.py --input ./frames/ --output ./ocr.json

# 指定引擎
python ocr-processor.py --input ./frames/ --engine paddleocr --output ./

# 只处理部分帧
python ocr-processor.py --input ./frames/ --start 0 --end 100 --output ./
```

### 6.2 输入参数

```yaml
参数:
  --input, -i     # 必需，帧图片目录或单个图片
  --output, -o    # 输出文件 (默认: ocr.json)
  --engine        # OCR引擎: paddleocr/tesseract (默认: paddleocr)
  --region        # 字幕区域: bottom/full/auto (默认: bottom)
  --fps           # 抽帧频率 (默认: 0.1，即每10秒1帧)
  --start         # 开始帧号 (默认: 0)
  --end           # 结束帧号 (默认: -1，即全部)
  --confidence    # 置信度阈值 (默认: 0.7)
```

### 6.3 输出格式

```json
{
  "success": true,
  "data": {
    "frames": {
      "frame_0001.jpg": {
        "texts": [
          {
            "text": "字幕文本",
            "confidence": 0.95,
            "bbox": [x1, y1, x2, y2]
          }
        ],
        "timestamp": 0.0
      }
    },
    "merged_text": "合并的字幕文本...",
    "statistics": {
      "total_frames": 100,
      "frames_with_text": 80,
      "total_texts": 500,
      "unique_texts": 200
    }
  },
  "metadata": {
    "engine": "paddleocr",
    "processed_at": "2024-01-15T10:00:00"
  }
}
```

### 6.4 复用代码

```python
# 复用现有代码
from ocr_processor import OCRProcessor
```

---

## 7. 技能4: deep-analyzer

**功能**: AI深度分析，按内容类型使用不同模板

### 7.1 命令行接口

```bash
# 基本用法
python deep-analyzer.py --metadata meta.json --content content.json --output ./analysis.md

# 指定类型
python deep-analyzer.py --metadata meta.json --type tutorial --output ./

# 使用自定义模板
python deep-analyzer.py --metadata meta.json --template my_template.md --output ./

# 带OCR结果
python deep-analyzer.py --metadata meta.json --content content.json --ocr ocr.json --output ./
```

### 7.2 输入参数

```yaml
参数:
  --metadata, -m  # 必需，元数据JSON文件
  --content, -c   # 必需，内容JSON文件
  --output, -o    # 必需，输出Markdown文件
  --type, -t      # 内容类型: auto/tutorial/documentation/news (默认: auto)
  --template      # 自定义模板路径
  --ocr           # OCR结果JSON文件 (可选)
  --model         # AI模型 (默认: 配置中指定)
  --tags          # 预设标签 (可选)
```

### 7.3 模板变量

**通用变量**:
```jinja2
{{ title }}           # 标题
{{ source }}          # 来源 (bilibili/zhihu)
{{ url }}             # 原始URL
{{ author }}          # 作者/UP主
{{ duration }}        # 时长
{{ date }}            # 发布日期
{{ tags }}            # 标签列表
{{ summary }}         # 详细摘要
{{ key_points }}      # 核心要点
{{ quality_score }}   # 质量评分 (1-100)
```

**教程视频额外变量**:
```jinja2
{{ prerequisites }}   # 前置知识
{{ outline }}        # 分段大纲
{{ pitfalls }}       # 常见坑点
{{ practice_advice }} # 实践建议
```

### 7.4 输出格式

**教程类输出**:
```markdown
---
title: "Python完整教程"
source: "bilibili"
url: "https://..."
author: "CodeMaster"
duration: "2:30:00"
date: "2024-01-01T00:00:00"
tags:
  - python
  - tutorial
  - 编程入门
quality_score: 95
---

## 🎯 一句话总结

本教程系统讲解Python基础语法，适合零基础学习者。

## 📚 前置知识

- 无需编程基础
- 电脑基本操作能力

## 📝 步骤详解

### 环境配置 [00:00-15:00]

首先下载Python...

![frame_0015: Python官网下载页面](./images/frame_0015.jpg)

```bash
# 安装Python
pip install python
```

### 基础语法 [15:00-45:00]

...

## 💡 关键要点

1. Python使用缩进定义代码块
2. 变量不需要声明类型
3. ...

## ⚠️ 常见坑点

- Python 2 与 3 的区别
- 缩进问题导致报错

## 🤔 实践建议

- 跟随视频动手编码
- 完成课后练习

## 📚 相关资源

- [Python官网](https://www.python.org/)
- [官方文档](https://docs.python.org/)
```

### 7.5 复用代码

```python
# 复用现有代码
from deep_analyzer import DeepAnalyzer
from ai_client import AIClient
```

---

## 8. 技能5: kb-workflow

**功能**: 工作流编排 + OpenClaw cron定时 + 手动触发 + 归档管理

### 8.1 命令行接口

```bash
# 手动触发处理
python kb-workflow.py --trigger manual

# 定时任务触发
python kb-workflow.py --trigger scheduled

# 只处理待处理列表
python kb-workflow.py --trigger manual --scope pending

# 强制刷新订阅
python kb-workflow.py --trigger manual --refresh-subscriptions

# 生成汇报
python kb-workflow.py --report

# 查看状态
python kb-workflow.py --status
```

### 8.2 输入参数

```yaml
参数:
  --trigger, -t   # 触发方式: manual/scheduled (默认: manual)
  --scope         # 处理范围: all/pending/metadata/content (默认: all)
  --limit         # 最大处理数量 (默认: 5)
  --refresh-subs  # 刷新订阅
  --report        # 生成汇报
  --status        # 显示状态
  --config, -c    # 配置文件路径
```

### 8.3 汇报文件格式

```json
{
  "run_id": "run_20240115_030000",
  "run_time": "2024-01-15T03:00:00",
  "trigger": "scheduled",
  "scope": "all",
  "processed": 5,
  "success": 4,
  "failed": 1,
  "details": {
    "total_inbox": 12,
    "pending": 2,
    "metadata_ready": 3,
    "content_ready": 5,
    "analyzed": 2,
    "archived": 10
  },
  "items": [
    {
      "id": "BV1xx411c7mD",
      "type": "video",
      "status": "success",
      "duration_seconds": 45
    },
    {
      "id": "BV2yy522c8nE",
      "type": "video",
      "status": "failed",
      "error": "download timeout"
    }
  ],
  "subscriptions": {
    "checked": 3,
    "updates_found": 5,
    "new_items_added": 2
  },
  "storage": {
    "inbox_size_mb": 128,
    "archive_size_mb": 512
  },
  "generated_at": "2024-01-15T03:05:00"
}
```

### 8.4 工作流步骤

```python
def run_workflow(trigger, scope, limit):
    # 1. 扫描收件箱
    pending_items = scan_pending()
    metadata_items = scan_metadata()
    content_items = scan_content()
    
    # 2. 检查订阅更新
    if trigger == "scheduled":
        check_subscriptions()
    
    # 3. 按优先级处理
    all_items = pending_items + metadata_items + content_items
    
    for item in all_items[:limit]:
        try:
            if item.status == "pending":
                metadata = run_skill2(item)
                update_item(item, metadata, status="metadata_ready")
            elif item.status == "metadata_ready":
                content = run_skill3(item)
                update_item(item, content, status="content_ready")
            elif item.status == "content_ready":
                analysis = run_skill4(item)
                archive_item(item, analysis)
                update_item(item, status="archived")
        except Exception as e:
            record_failure(item, str(e))
    
    # 4. 生成汇报
    generate_report()
```

### 8.5 复用代码

```python
# 复用现有代码
from pipeline import Pipeline
from git_sync import GitSync
```

---

## 9. 数据格式定义

### 9.1 收件箱目录结构

```
inbox/
├── pending/
│   └── bilibili_video_BV1xx_2024-01-15T100000.json
├── metadata/
│   └── bilibili_video_BV1xx_2024-01-15T100000.json
├── content/
│   └── bilibili_video_BV1xx_2024-01-15T100000.json
└── analyzed/
    └── bilibili_video_BV1xx_2024-01-15T100000.json
```

### 9.2 元数据文件格式

```json
{
  "platform": "bilibili",
  "type": "video",
  "id": "BV1xx411c7mD",
  "source_url": "https://www.bilibili.com/video/BV1xx411c7mD",
  "status": "pending",
  "scope": "all",
  "title": "Python完整教程",
  "author": {"id": 7770724, "name": "UP主名字"},
  "duration": 9000,
  "publish_time": "2024-01-01T12:00:00",
  "tags": ["python", "tutorial"],
  "stats": {"view": 100000, "like": 5000},
  "subscription": {
    "enabled": true,
    "scope": "new_only",
    "frequency": "daily",
    "last_checked": "2024-01-15T00:00:00",
    "items_count": 156,
    "processed_count": 45
  },
  "created_at": "2024-01-15T00:00:00",
  "updated_at": "2024-01-15T10:00:00",
  "retry_count": 0
}
```

### 9.3 归档目录结构

```
archive/
├── 2024/
│   ├── 01/
│   │   ├── video/
│   │   │   ├── bilibili_BV1xx_2024-01-15.md
│   │   │   └── index.json
│   │   ├── article/
│   │   │   └── index.json
│   │   └── index.json
│   └── index.json
└── index.json
```

---

## 10. 配置文件规范

### 10.1 配置文件位置

优先级:
1. `./config.yaml` (当前目录)
2. `~/.config/obsidian-kb/config.yaml`
3. `~/.obsidian-kb/config.yaml`

### 10.2 完整配置示例

```yaml
# AI 配置
ai:
  provider: "openai"
  model: "gpt-4"
  api_key: "${OPENAI_API_KEY}"
  temperature: 0.3
  max_tokens: 8192

# 存储配置
storage:
  inbox_dir: "~/work/inbox"
  archive_dir: "~/ObsidianVault/knowledge-base"
  temp_dir: "/tmp/obsidian-kb"
  max_concurrent: 3
  cleanup_after_days: 7

# 输出配置
output:
  format: "markdown"
  encoding: "utf-8"
  include_timestamp: true
  include_source: true

# Git 同步配置
git:
  enabled: true
  vault_path: "~/ObsidianVault"
  remote: "origin"
  branch: "main"
  auto_commit: true

# Bilibili 提取器配置
bilibili:
  enabled: true
  cookie: "${BILIBILI_COOKIE}"
  use_playwright: true
  headless: true
  timeout: 30
  subtitle:
    enabled: true
    language: "zh-CN"

# 知乎提取器配置
zhihu:
  enabled: true
  use_drissionpage: true
  cookie_file: "~/.obsidian-kb/zhihu_cookies.json"

# OCR 配置
ocr:
  engine: "paddleocr"
  confidence_threshold: 0.7
  region: "bottom"
  fps: 0.1
  max_frames: 500

# 工作流配置
workflow:
  schedule: "0 3 * * *"
  max_items_per_run: 5
  subscription_check_interval: 24

# 汇报配置
reporting:
  enabled: true
  output_dir: "~/work/reports"
  format: "json"
  retention_days: 30

# 日志配置
logging:
  level: "INFO"
  format: "[%(asctime)s] %(levelname)s - %(message)s"
  file: "~/work/logs/obsidian-kb.log"
```

---

## 11. 扩展指南

### 11.1 新增平台步骤

以新增「小红书」为例：

#### 步骤1: 创建链接解析器扩展

```python
# link-parser 的 platform_registry 中添加
PLATFORM_REGISTRY = {
    "xiaohongshu": {
        "name": "小红书",
        "patterns": [
            r"xiaohongshu.com/notes/([a-zA-Z0-9]+)",
        ],
        "types": ["note", "user", "collection"]
    }
}
```

#### 步骤2: 创建元数据获取器

```bash
# 创建文件: skills/xiaohongshu-metadata/
```

```python
class XiaohongshuMetadata:
    def extract(self, url, config):
        # 实现获取逻辑
        pass
```

#### 步骤3: 更新工作流配置

```yaml
xiaohongshu:
  enabled: true
  cookie: "${XIAOHONGSHU_COOKIE}"
```

### 11.2 新增分析模板

```jinja2
# templates/podcast.md
---
title: "{{ title }}"
source: "{{ source }}"
url: "{{ url }}"
author: "{{ author }}"
duration: "{{ duration }}"
date: "{{ date }}"
tags:
{{- for tag in tags }}
  - {{ tag }}
{{- endfor }}
---

## 🎧 播客摘要

{{ summary }}

## 💡 核心要点

{{ key_points }}
```

---

## 附录A: 技能清单

| # | 技能名 | 主要文件 | 复用代码 |
|---|--------|----------|----------|
| 1 | link-parser | `link-parser.py` | shortlink_resolver.py |
| 2 | bilibili-metadata | `bilibili_metadata.py` | bilibili_extractor.py |
| 2 | zhihu-metadata | `zhihu_metadata.py` | zhihu_extractor.py |
| 3 | bilibili-content | `bilibili_content.py` | video_stream_extractor.py |
| 3 | ocr-processor | `ocr_processor.py` | ocr_processor.py |
| 4 | deep-analyzer | `deep_analyzer.py` | deep_analyzer.py |
| 5 | kb-workflow | `kb_workflow.py` | pipeline.py, git_sync.py |

---

## 附录B: OpenClaw Cron 配置示例

```yaml
# 定时任务配置
schedule: "0 3 * * *"
command: "python /path/to/skills/kb-workflow/kb_workflow.py --trigger scheduled"
```

---

蒋先生，详细设计文档已完成！

**包含内容：**
- ✅ 7个技能的完整接口规范（命令行参数、输入输出格式）
- ✅ 数据格式详细定义（收件箱、归档、汇报）
- ✅ 配置文件规范
- ✅ 复用代码对照表
- ✅ 扩展指南（新增平台/模板步骤）

**下一步：**
可以进入 **write-plan** 阶段，生成具体实现计划吗？