宜博 / lemonboss

Commit 3c8952e467de88621ecff0120a9b11c03a35ad91

Authored by yibo
1 parent 5454b2cd

feat: track template agent skills and sync latest updates

Inline Side-by-side
Showing 10 changed files with 613 additions and 12 deletions
.gitignore
  View file @3c8952e
... ... @@ -62,6 +62,15 @@ tmp/
62 62  
63 63 # === 运行时文件 (不提交) ===
64 64 .claude/
  65 +!system/templates/*/.claude/
  66 +system/templates/*/.claude/*
  67 +!system/templates/*/.claude/skills/
  68 +!system/templates/*/.claude/skills/**
  69 +!system/templates/*/.claude/*.lock
  70 +!system/agents/*/.claude/
  71 +system/agents/*/.claude/*
  72 +!system/agents/*/.claude/skills/
  73 +!system/agents/*/.claude/skills/**
65 74 .pylibs/
66 75 nohup.out
67 76 data/agents/
... ...
system/agents/master/.claude/scheduled_tasks.json
  View file @3c8952e
... ... @@ -5,6 +5,7 @@
5 5 "cron": "*/5 * * * *",
6 6 "prompt": "执行系统巡检：读取 plan.md 和 todo.md，检查所有任务状态。已完成的更新标记，卡住的分析原因后自主处理（重新分发或修复），未开始的按优先级立即分发。不要请示用户，直接行动。",
7 7 "createdAt": 1775231859266,
  8 + "lastFiredAt": 1775235913481,
8 9 "recurring": true
9 10 },
10 11 {
... ... @@ -12,6 +13,22 @@
12 13 "cron": "*/30 * * * *",
13 14 "prompt": "生成系统状态日报，统计所有 Agent 任务完成率和异常情况，写入 memory/daily/ 今日文件。",
14 15 "createdAt": 1775231859272,
  16 + "lastFiredAt": 1775235574165,
  17 + "recurring": true
  18 + },
  19 + {
  20 + "id": "7868eeb3",
  21 + "cron": "*/5 * * * *",
  22 + "prompt": "执行系统巡检：读取 plan.md 和 todo.md，检查所有任务状态。已完成的更新标记，卡住的分析原因后自主处理（重新分发或修复），未开始的按优先级立即分发。不要请示用户，直接行动。",
  23 + "createdAt": 1775235820776,
  24 + "recurring": true,
  25 + "lastFiredAt": 1775235971536
  26 + },
  27 + {
  28 + "id": "695f0a30",
  29 + "cron": "*/30 * * * *",
  30 + "prompt": "生成系统状态日报，统计所有 Agent 任务完成率和异常情况，写入 memory/daily/ 今日文件。",
  31 + "createdAt": 1775235825837,
15 32 "recurring": true
16 33 }
17 34 ]
... ...
system/templates/claude-youtube-skills/.claude/skills/skills.md 0 → 100644
  View file @3c8952e
  1 +# 我的技能
  2 +
  3 +## 技能清单
  4 +（尚未添加技能）
  5 +
  6 +## 说明
  7 +- 每个技能是 skills/ 目录下的一个子文件夹
  8 +- 每个技能文件夹内必须包含 SKILL.md 定义文件
... ...
system/templates/claude-youtube-skills/.claude/skills/youtube-competitor-analyzer/SKILL.md 0 → 100644
  View file @3c8952e
  1 +---
  2 +name: youtube-competitor-analyzer
  3 +description: "Find and analyze YouTube competitor channels using YouTube Data API v3. Discover competitors through keyword search, category matching, content similarity, and related channel discovery. Compare metrics, content strategies, and market positioning. Use when users want to (1) Find competitors for their YouTube channel, (2) Analyze competitor performance metrics, (3) Compare their channel against competitors, (4) Identify content gaps and opportunities, (5) Benchmark against similar creators, (6) Generate competitive analysis reports. Requires user's YouTube Data API v3 key."
  4 +---
  5 +
  6 +# YouTube Competitor Analyzer
  7 +
  8 +Find and analyze competitor channels using YouTube Data API v3.
  9 +
  10 +## Setup
  11 +
  12 +1. **API Key**: Get from [Google Cloud Console](https://console.cloud.google.com/apis/credentials) - enable YouTube Data API v3
  13 +2. **Input**: User's channel (for context) OR list of competitor channels to analyze directly
  14 +
  15 +## API Endpoints
  16 +
  17 +```
  18 +BASE_URL = https://www.googleapis.com/youtube/v3
  19 +
  20 +# Search channels
  21 +GET /search?part=snippet&q={query}&type=channel&maxResults={n}&key={API_KEY}
  22 +
  23 +# Channel details
  24 +GET /channels?part=snippet,statistics,contentDetails,brandingSettings&id={ids}&key={API_KEY}
  25 +
  26 +# Channel videos
  27 +GET /search?part=snippet&channelId={id}&order=date&type=video&maxResults=50&key={API_KEY}
  28 +
  29 +# Video details
  30 +GET /videos?part=snippet,statistics,contentDetails&id={ids}&key={API_KEY}
  31 +```
  32 +
  33 +## Competitor Discovery Methods
  34 +
  35 +### 1. Keyword Search
  36 +Extract keywords from user's channel (description, title, brandingSettings.keywords), search for channels:
  37 +```
  38 +GET /search?q={keyword}&type=channel&maxResults=10
  39 +```
  40 +
  41 +### 2. Content Similarity
  42 +Get user's recent video titles, extract frequent words (>4 chars), search channels with those terms.
  43 +
  44 +### 3. Category Search
  45 +Search within same YouTube category as user's channel.
  46 +
  47 +### 4. Related Discovery
  48 +Search using user's video titles/descriptions to find channels creating similar content.
  49 +
  50 +### 5. Direct Input
  51 +User provides specific channels to analyze: @handles, channel IDs, or URLs.
  52 +
  53 +## Competitor Ranking
  54 +
  55 +```javascript
  56 +// Size similarity (log scale)
  57 +sizeSimilarity = 1 - Math.abs(Math.log10(compSubs+1) - Math.log10(targetSubs+1)) / 10;
  58 +
  59 +// Engagement per video
  60 +engagementRate = viewCount / (videoCount || 1);
  61 +
  62 +// Overall score
  63 +overallScore = sizeSimilarity * 0.3 + relevanceScore * 0.7;
  64 +```
  65 +
  66 +Sort by overallScore, return top 10-15.
  67 +
  68 +## Comparative Analysis
  69 +
  70 +For each competitor, extract:
  71 +| Metric | Description |
  72 +|--------|-------------|
  73 +| Subscribers | Direct count |
  74 +| Total Views | Channel lifetime views |
  75 +| Video Count | Total videos published |
  76 +| Views/Video | Average performance |
  77 +| Upload Frequency | Videos per month |
  78 +| Channel Age | Days since creation |
  79 +
  80 +Compare user channel vs competitor averages:
  81 +- Above/below average positioning
  82 +- Growth rate comparison
  83 +- Content volume gaps
  84 +
  85 +## Content Gap Analysis
  86 +
  87 +Compare across competitors:
  88 +- Content types they cover vs user doesn't
  89 +- Duration preferences
  90 +- Upload schedules
  91 +- Title/thumbnail patterns
  92 +- High-performing topics
  93 +
  94 +## Output Report Structure
  95 +
  96 +```markdown
  97 +# Competitive Analysis Report
  98 +## Executive Summary
  99 +- Competitors found, key positioning insights
  100 +## Competitor Table
  101 +| Channel | Subs | Videos | Views | Views/Video | Score |
  102 +## Competitive Positioning
  103 +- User vs competitor averages
  104 +- Market position assessment
  105 +## Content Strategy Comparison
  106 +- What competitors do differently
  107 +- Common patterns among top performers
  108 +## Opportunities
  109 +- Content gaps to fill
  110 +- Underserved niches
  111 +- Schedule optimization
  112 +## Recommendations
  113 +- Strategic actions based on analysis
  114 +```
  115 +
  116 +## Niche-Specific Search Terms
  117 +
  118 +For spiritual/wellness channels:
  119 +```
  120 +meditation, mindfulness, yoga, spiritual, consciousness, wellness, self-help, awakening
  121 +```
  122 +
  123 +For tech channels:
  124 +```
  125 +tech review, unboxing, tutorial, how to, comparison, tips, gadget
  126 +```
  127 +
  128 +For gaming:
  129 +```
  130 +gameplay, let's play, walkthrough, gaming, stream, esports
  131 +```
  132 +
  133 +## Quota Optimization
  134 +
  135 +- Limit discovery to 20-30 initial channels
  136 +- Detailed analysis for top 10-15 only
  137 +- Batch video IDs (max 50 per call)
  138 +- Cache results when comparing multiple times
... ...
system/templates/claude-youtube-skills/.claude/skills/youtube-competitor-analyzer/references/api_reference.md 0 → 100644
  View file @3c8952e
  1 +# YouTube Data API v3 Reference
  2 +
  3 +## Authentication
  4 +
  5 +All requests require `key={API_KEY}` parameter.
  6 +
  7 +Get API key: [Google Cloud Console](https://console.cloud.google.com/apis/credentials)
  8 +1. Create project → Enable YouTube Data API v3 → Create credentials → API Key
  9 +
  10 +## Channels Endpoint
  11 +
  12 +```
  13 +GET https://www.googleapis.com/youtube/v3/channels
  14 +```
  15 +
  16 +### Parameters
  17 +| Param | Required | Description |
  18 +|-------|----------|-------------|
  19 +| part | Yes | snippet,statistics,contentDetails,brandingSettings |
  20 +| id | One of | Channel ID (UC...) - can be comma-separated for multiple |
  21 +| forUsername | One of | Legacy username |
  22 +| key | Yes | API key |
  23 +
  24 +### Response Fields
  25 +```json
  26 +{
  27 + "items": [{
  28 + "id": "UC...",
  29 + "snippet": {
  30 + "title": "Channel Name",
  31 + "description": "...",
  32 + "customUrl": "@handle",
  33 + "publishedAt": "2020-01-01T00:00:00Z",
  34 + "country": "US",
  35 + "thumbnails": { "default": {}, "medium": {}, "high": {} }
  36 + },
  37 + "statistics": {
  38 + "viewCount": "1000000",
  39 + "subscriberCount": "50000",
  40 + "hiddenSubscriberCount": false,
  41 + "videoCount": "200"
  42 + },
  43 + "contentDetails": {
  44 + "relatedPlaylists": {
  45 + "uploads": "UU..."
  46 + }
  47 + },
  48 + "brandingSettings": {
  49 + "channel": {
  50 + "keywords": "keyword1 keyword2"
  51 + }
  52 + }
  53 + }]
  54 +}
  55 +```
  56 +
  57 +## Search Endpoint (Channel Discovery)
  58 +
  59 +```
  60 +GET https://www.googleapis.com/youtube/v3/search
  61 +```
  62 +
  63 +### Parameters
  64 +| Param | Required | Description |
  65 +|-------|----------|-------------|
  66 +| part | Yes | snippet |
  67 +| q | Yes | Search query (keywords, channel names) |
  68 +| type | Yes | channel (for competitor discovery) |
  69 +| order | No | relevance (default), viewCount |
  70 +| maxResults | No | 1-50 (default 5) |
  71 +| pageToken | No | For pagination |
  72 +| key | Yes | API key |
  73 +
  74 +### Response
  75 +```json
  76 +{
  77 + "nextPageToken": "...",
  78 + "items": [{
  79 + "id": {
  80 + "kind": "youtube#channel",
  81 + "channelId": "UC..."
  82 + },
  83 + "snippet": {
  84 + "channelId": "UC...",
  85 + "title": "Channel Name",
  86 + "description": "...",
  87 + "thumbnails": {}
  88 + }
  89 + }]
  90 +}
  91 +```
  92 +
  93 +## Videos Endpoint
  94 +
  95 +```
  96 +GET https://www.googleapis.com/youtube/v3/videos
  97 +```
  98 +
  99 +### Parameters
  100 +| Param | Required | Description |
  101 +|-------|----------|-------------|
  102 +| part | Yes | snippet,statistics,contentDetails |
  103 +| id | Yes | Comma-separated video IDs (max 50) |
  104 +| key | Yes | API key |
  105 +
  106 +### Response Fields
  107 +```json
  108 +{
  109 + "items": [{
  110 + "id": "videoId",
  111 + "snippet": {
  112 + "publishedAt": "...",
  113 + "channelId": "...",
  114 + "title": "...",
  115 + "description": "...",
  116 + "tags": ["tag1", "tag2"],
  117 + "categoryId": "22"
  118 + },
  119 + "statistics": {
  120 + "viewCount": "10000",
  121 + "likeCount": "500",
  122 + "commentCount": "50"
  123 + },
  124 + "contentDetails": {
  125 + "duration": "PT10M30S",
  126 + "definition": "hd"
  127 + }
  128 + }]
  129 +}
  130 +```
  131 +
  132 +## Quota Costs
  133 +
  134 +| Operation | Cost |
  135 +|-----------|------|
  136 +| search.list | 100 |
  137 +| channels.list | 1 |
  138 +| videos.list | 1 |
  139 +| playlistItems.list | 1 |
  140 +
  141 +**Daily quota**: 10,000 units (default)
  142 +
  143 +**Optimization for competitor analysis**:
  144 +- Batch channel IDs (comma-separated) to reduce calls
  145 +- Limit initial discovery searches to essential keywords
  146 +- Get detailed video data only for top-ranked competitors
  147 +
  148 +## Error Codes
  149 +
  150 +| Code | Meaning |
  151 +|------|---------|
  152 +| 400 | Bad request (invalid params) |
  153 +| 403 | Forbidden (quota exceeded, API not enabled) |
  154 +| 404 | Not found |
  155 +| 429 | Too many requests |
  156 +
  157 +## Channel ID Formats
  158 +
  159 +| Format | Example | Resolution |
  160 +|--------|---------|------------|
  161 +| Direct ID | UC... | Use directly |
  162 +| @handle | @channelname | Search API with handle as query |
  163 +| /channel/ URL | youtube.com/channel/UC... | Extract ID from URL |
  164 +| /c/ URL | youtube.com/c/name | forUsername parameter |
  165 +| /user/ URL | youtube.com/user/name | forUsername parameter |
  166 +
  167 +## Duration Format (ISO 8601)
  168 +
  169 +Format: `PT{hours}H{minutes}M{seconds}S`
  170 +
  171 +Examples:
  172 +- `PT1H30M` = 1 hour 30 minutes
  173 +- `PT10M30S` = 10 minutes 30 seconds
  174 +
  175 +Parse: `/PT(?:(\d+)H)?(?:(\d+)M)?(?:(\d+)S)?/`
... ...
system/templates/claude-youtube-skills/.claude/skills/youtube-trending-crawler/SKILL.md 0 → 100644
  View file @3c8952e
  1 +---
  2 +name: youtube-trending-crawler
  3 +description: "Fetch YouTube trending videos using YouTube Data API v3. Get popular videos by region and category with complete metadata. Use when users want to (1) Get trending videos for a region, (2) Fetch popular videos by category, (3) Batch crawl multiple regions, (4) Monitor trending content across markets. Requires YouTube Data API v3 key."
  4 +---
  5 +
  6 +# YouTube Trending Crawler
  7 +
  8 +Fetch YouTube trending/popular videos using YouTube Data API v3.
  9 +
  10 +## Setup
  11 +
  12 +1. **API Key**: Get from [Google Cloud Console](https://console.cloud.google.com/apis/credentials) - enable YouTube Data API v3
  13 +2. **Regions**: ISO 3166-1 alpha-2 codes (US, JP, KR, GB, DE, FR, IN, BR, etc.)
  14 +3. **Categories**: Video category IDs (1=Film, 2=Autos, 10=Music, 17=Sports, 20=Gaming, 24=Entertainment, 28=Tech, etc.)
  15 +
  16 +## Core API
  17 +
  18 +### Primary: videos endpoint (1 quota unit)
  19 +
  20 +```bash
  21 +# Most Popular Videos (CHART endpoint - 1 unit)
  22 +curl "https://www.googleapis.com/youtube/v3/videos?part=snippet,statistics,contentDetails&chart=mostPopular&regionCode=US&videoCategoryId=10&maxResults=10&key=API_KEY"
  23 +
  24 +# Batch video IDs (max 50 per call)
  25 +curl "https://www.googleapis.com/youtube/v3/videos?part=snippet,statistics,contentDetails&id=VIDEO_IDS&key=API_KEY"
  26 +```
  27 +
  28 +### Fallback: yt-dlp (when API fails)
  29 +
  30 +```bash
  31 +# Get trending page
  32 +yt-dlp "https://www.youtube.com/results?sp=CAA%253D&q=trending" --flat-playlist --print "%(id)s|%(title)s|%(channel_id)s|%(channel)s|%(view_count)s|%(like_count)s|%(comment_count)s|%(upload_date)s|%(description)s|%(thumbnail)s|%(categories)s|%(duration)s"
  33 +
  34 +# Get specific video details
  35 +yt-dlp "https://youtube.com/watch?v=VIDEO_ID" --print "%(id)s|%(title)s|%(channel_id)s|%(channel)s|%(view_count)s|%(like_count)s|%(comment_count)s|%(upload_date)s|%(description)s|%(thumbnail)s|%(categories)s|%(duration)s|%(tags)s"
  36 +```
  37 +
  38 +## Parameters
  39 +
  40 +| Parameter | Type | Required | Description |
  41 +|-----------|------|----------|-------------|
  42 +| regionCode | string | No | ISO 3166-1 alpha-2 (default: US) |
  43 +| videoCategoryId | int | No | 1-44 category ID |
  44 +| maxResults | int | No | 1-50 (default: 10) |
  45 +| pageToken | string | No | Pagination token |
  46 +
  47 +## Supported Regions
  48 +
  49 +```javascript
  50 +const SUPPORTED_REGIONS = [
  51 + 'US', 'JP', 'KR', 'GB', 'DE', 'FR', 'IN', 'BR', // Primary
  52 + 'CA', 'AU', 'RU', 'MX', 'ES', 'IT', 'NL', 'SE' // Extended
  53 +];
  54 +```
  55 +
  56 +## Supported Categories
  57 +
  58 +```javascript
  59 +const CATEGORIES = {
  60 + 1: 'Film & Animation',
  61 + 2: 'Autos & Vehicles',
  62 + 10: 'Music',
  63 + 15: 'Pets & Animals',
  64 + 17: 'Sports',
  65 + 20: 'Gaming',
  66 + 22: 'People & Blogs',
  67 + 23: 'Comedy',
  68 + 24: 'Entertainment',
  69 + 25: 'News & Politics',
  70 + 27: 'Education',
  71 + 28: 'Science & Technology'
  72 +};
  73 +```
  74 +
  75 +## Output Fields (18 fields)
  76 +
  77 +| Field | Source | Type |
  78 +|-------|--------|------|
  79 +| rank | generated (1-N) | int |
  80 +| videoId | id | string |
  81 +| title | snippet.title | string |
  82 +| url | https://youtube.com/watch?v={id} | string |
  83 +| channelTitle | snippet.channelTitle | string |
  84 +| channelId | snippet.channelId | string |
  85 +| subscriberCount | channels.list (separate call) | int |
  86 +| viewCount | statistics.viewCount | int |
  87 +| likeCount | statistics.likeCount | int |
  88 +| commentCount | statistics.commentCount | int |
  89 +| publishedAt | snippet.publishedAt | ISO8601 |
  90 +| description | snippet.description (truncated 500 chars) | string |
  91 +| thumbnail | snippet.thumbnails.default.url | string |
  92 +| categoryId | snippet.categoryId | string |
  93 +| categoryName | mapped from CATEGORIES | string |
  94 +| duration | contentDetails.duration (ISO 8601) | string |
  95 +| crawlTime | generated | ISO8601 |
  96 +| regionCode | from request | string |
  97 +
  98 +## Output File
  99 +
  100 +Save to: `data/companies/test/workspace/{date}-youtube-trending-{region}-{category}.json`
  101 +
  102 +Example: `2026-04-04-youtube-trending-us-politics.json`
  103 +
  104 +```json
  105 +{
  106 + "videos": [
  107 + {
  108 + "rank": 1,
  109 + "videoId": "dQw4w9WgXcQ",
  110 + "title": "Video Title",
  111 + "url": "https://youtube.com/watch?v=dQw4w9WgXcQ",
  112 + "channelTitle": "Channel Name",
  113 + "channelId": "UCxxxx",
  114 + "subscriberCount": 1000000,
  115 + "viewCount": 5000000,
  116 + "likeCount": 250000,
  117 + "commentCount": 50000,
  118 + "publishedAt": "2024-01-15T10:00:00Z",
  119 + "description": "Video description...",
  120 + "thumbnail": "https://i.ytimg.com/vi/dQw4w9WgXcQ/default.jpg",
  121 + "categoryId": "25",
  122 + "categoryName": "News & Politics",
  123 + "duration": "PT21M4S",
  124 + "crawlTime": "2026-04-04T00:00:00Z",
  125 + "regionCode": "US"
  126 + }
  127 + ],
  128 + "metadata": {
  129 + "regionCode": "US",
  130 + "videoCategoryId": "25",
  131 + "categoryName": "News & Politics",
  132 + "totalResults": 20,
  133 + "crawlTime": "2026-04-04T00:00:00Z"
  134 + }
  135 +}
  136 +```
  137 +
  138 +## Error Handling
  139 +
  140 +### 403 Quota Exceeded
  141 +```javascript
  142 +if (error.code === 403 && error.message.includes('quota')) {
  143 + // Fallback to yt-dlp
  144 + return await fetchWithYtDlp(params);
  145 +}
  146 +```
  147 +
  148 +### 404 Video Not Found
  149 +```javascript
  150 +if (error.code === 404) {
  151 + console.log(`Video ${videoId} not found, skipping`);
  152 + return null;
  153 +}
  154 +```
  155 +
  156 +### 500 Server Error
  157 +```javascript
  158 +// Exponential backoff retry
  159 +const maxRetries = 3;
  160 +for (let i = 0; i < maxRetries; i++) {
  161 + try {
  162 + return await fetch(...);
  163 + } catch (e) {
  164 + if (e.code === 500 && i < maxRetries - 1) {
  165 + await sleep(Math.pow(2, i) * 1000); // 1s, 2s, 4s
  166 + }
  167 + }
  168 +}
  169 +```
  170 +
  171 +### Rate Limit (429)
  172 +```javascript
  173 +if (error.code === 429) {
  174 + // Wait and retry with longer delay
  175 + await sleep(60000); // 1 minute
  176 + return await fetch(...);
  177 +}
  178 +```
  179 +
  180 +## Quota Management
  181 +
  182 +- Use `videos?chart=mostPopular` endpoint (1 unit)
  183 +- NEVER use `search` endpoint (100 units)
  184 +- Batch video IDs (max 50 per call when extending)
  185 +- Cache category names locally
  186 +- Rate limit: min 1000ms between requests
  187 +
  188 +## Category Mapping
  189 +
  190 +```javascript
  191 +function getCategoryName(categoryId) {
  192 + const map = {
  193 + '1': 'Film & Animation',
  194 + '2': 'Autos & Vehicles',
  195 + '10': 'Music',
  196 + '15': 'Pets & Animals',
  197 + '17': 'Sports',
  198 + '20': 'Gaming',
  199 + '22': 'People & Blogs',
  200 + '23': 'Comedy',
  201 + '24': 'Entertainment',
  202 + '25': 'News & Politics',
  203 + '27': 'Education',
  204 + '28': 'Science & Technology'
  205 + };
  206 + return map[categoryId] || 'Unknown';
  207 +}
  208 +```
... ...
system/templates/claude-youtube-skills/agent.md
  View file @3c8952e
... ... @@ -14,3 +14,12 @@
14 14  
15 15 ## 可用工具
16 16 # 无管理工具 — Skill 只负责执行, 不管理其他 Agent
  17 +
  18 +---
  19 +
  20 +## 所属公司
  21 +- **公司ID**: test
  22 +- **公司名称**: test
  23 +- **公司使命**: 待定义
  24 +
  25 +> 你是 **test** 的员工。你的所有工作都应围绕公司的使命和目标展开。
... ...
system/templates/claude-youtube-skills/config.json
  View file @3c8952e
... ... @@ -2,9 +2,9 @@
2 2 "id": "claude-youtube-skills",
3 3 "type": "skill",
4 4 "name": "claude-youtube-skills",
5   - "parent_id": null,
  5 + "parent_id": "ceo",
6 6 "goal_id": null,
7   - "status": "ready",
  7 + "status": "active",
8 8 "cli": {
9 9 "toolPolicy": {}
10 10 },
... ... @@ -21,5 +21,9 @@
21 21 "maxMemoryMB": 512,
22 22 "maxDiskMB": 1024,
23 23 "maxProcesses": 5
  24 + },
  25 + "heartbeat": {
  26 + "intervalMinutes": 15,
  27 + "enabled": false
24 28 }
25 29 }
26 30 \ No newline at end of file
... ...
system/templates/claude-youtube-skills/plan.md
  View file @3c8952e
1   -# Plan
  1 +# claude-youtube-skills 训练计划
2 2  
3   -> 将大任务分解为子任务，分配给下级 Agent，并定期检查进度。
  3 +> FDE 流水线训练状态追踪
4 4  
5   -| # | 子任务 | 责任人 | 状态 | 上次检查 | 备注 |
6   -|---|--------|--------|------|----------|------|
7   -| | （暂无计划） | | | | |
  5 +## 目标
  6 +通过 YouTube Data API v3 抓取最热最新视频列表，输出 JSON 格式。
  7 +
  8 +## 流水线状态
  9 +
  10 +| Phase | 状态 | Agent | 版本 | 备注 |
  11 +|-------|------|-------|------|------|
  12 +| Hunt | ✅ 完成 | Hunter | v2 | knowhow_v2.md + crawl_config.json |
  13 +| Digest | ✅ 完成 | Digester | v2 | benchmark_v2.md + train/test |
  14 +| Evolve | ✅ 完成 (100分) | Evolver | v2 | 弱维度已优化 |
  15 +| Deploy | ✅ 已部署 | Innovator | v2 | SKILL.md 已部署到 skills/ |
  16 +
  17 +## 训练历史
  18 +- v1: youtube-competitor-analyzer (频道竞品分析)
  19 +- v2: youtube-trending-crawler (趋势视频爬取) - **已完成 ✅**
  20 +
  21 +## 行业基线
  22 +- target_score: ≥ 90分
  23 +- **最终得分: 100分** ✅
  24 +
  25 +## 任务 ID
  26 +- Hunt: T-mnj3emdx
  27 +- Digest: T-mnj3nimu
  28 +- Evolve: T-mnj3s7bi
  29 +
  30 +## 启动时间
  31 +2026-04-04
  32 +
  33 +## 完成时间
  34 +2026-04-04
8 35 \ No newline at end of file
... ...
system/templates/claude-youtube-skills/user.md
  View file @3c8952e
1   -# 上层上下文
  1 +# 与上级的关系
2 2  
3   -## 我的上级
4   -- 直接向用户汇报 (无上级 Agent)
  3 +## 所属公司
  4 +- **公司**: test (test)
5 5  
6   -## 用户偏好
7   -（待用户定义）
  6 +## 汇报对象
  7 +- **直接汇报**: CEO (公司 CEO)
  8 +- **协同**: 公司内其他同事
  9 +
  10 +## 沟通规则
  11 +1. 完成任务后主动向上级汇报结果
  12 +2. 遇到阻碍时及时上报，附带解决建议
  13 +3. 保持透明，坏消息更要及时同步
... ...