‹ 首页

apps

@clawic · 收录于 昨天 · 上游提交 2 天前

Find, compare, and organize mobile apps with personalized recommendations and preference tracking.

适合你,如果你经常需要查找、比较手机应用并获得个性化推荐。

/ 通过 npx 安装 校验哈希
npx oh-my-skill add clawic/skills/apps
/ 通过 bash 安装
curl -fsSL https://oh-my-skill.com/install.sh | bash -s -- clawic/skills/apps
/ 已经装过?验证本机副本,不用重装
npx oh-my-skill verify clawic/skills/apps
安装目标可用 --agent / --scope 或 --to 明确指定;省略时只会在唯一已存在的 agent 目录上自动选择,零命中或多命中会停止并提示。content_hash 缺失或不一致均拒装。
14GitHub stars
~829最小装载
~1.9K含声明引用
~1.9K文本包总量
索引托管

怎么用

商店整理自技能原文 · 版本 f825206 · 表述以原文为准
它做什么

安装后,Claude 能根据你的偏好推荐手机应用,比较同类应用的优缺点,并记录你喜欢、试过或想试的应用,所有信息保存在本地文件中。

什么时候触发

当你问“推荐某个类型的应用”、“比较两款应用”或“帮我整理应用列表”时,Claude 会触发此技能。

装好后可以这样说
Claude 会检查偏好文件后推荐 3 款并说明优缺点。
Claude 会按相同类别使用一致标准对比并指出适用场景。
Claude 会更新 tried.md 文件并询问评价。
技能原文 SKILL.md作者撰写 · MIT · f825206
When to Use

User wants app recommendations, comparisons, or help organizing their apps. Covers iOS and Android. Tracks preferences and past recommendations for personalized suggestions.

Architecture

Memory lives in ~/Clawic/data/apps/. See memory-template.md for setup.

~/Clawic/data/apps/
├── memory.md          # Preferences, platforms, dislikes
├── favorites.md       # Apps user loves, organized by category
├── tried.md           # Apps tested with notes (liked/disliked/why)
└── wishlist.md        # Apps to try later
Quick Reference

| Topic | File | |-------|------| | Memory setup | memory-template.md | | Category guide | categories.md | | Comparison framework | compare.md |

Data Storage

All data stored in ~/Clawic/data/apps/. Create on first use:

mkdir -p ~/apps
Scope

This skill ONLY:

  • Recommends apps based on user criteria
  • Stores user preferences in local files (~/Clawic/data/apps/)
  • Tracks apps user has tried or wants to try
  • Compares apps within categories

This skill NEVER:

  • Installs apps automatically
  • Accesses App Store/Play Store accounts
  • Makes purchases or subscriptions
  • Reads installed apps from device
Core Rules
1. Check Preferences First

Before recommending, read ~/Clawic/data/apps/memory.md:

  • Platform (iOS, Android, both)
  • Pricing preference (free, freemium, paid OK, no subscriptions)
  • Past dislikes (apps/patterns to avoid)
2. Recommendation Quality

| Criteria | Action | |----------|--------| | User asks "best X app" | Give top 3 with tradeoffs | | User has tried similar | Check ~/Clawic/data/apps/tried.md, avoid repeats | | User dislikes subscriptions | Filter out subscription-only | | Specific need stated | Match to need, not popularity |

3. Always Explain Tradeoffs

Never just say "use X". Include:

  • What it's great at
  • What it's weak at
  • Pricing model (one-time, subscription, freemium limits)
  • Privacy stance if relevant
4. Update Memory Proactively

| Event | Action | |-------|--------| | User says "I use iPhone" | Add to ~/Clawic/data/apps/memory.md | | User says "I hate subscriptions" | Add to ~/Clawic/data/apps/memory.md dislikes | | User likes recommendation | Add to ~/Clawic/data/apps/favorites.md | | User tries and dislikes | Add to ~/Clawic/data/apps/tried.md with reason | | User says "remind me to try X" | Add to ~/Clawic/data/apps/wishlist.md |

5. Category Organization

Organize favorites by category:

  • Productivity, Notes, Tasks
  • Health, Fitness, Meditation
  • Finance, Budgeting
  • Photo, Video, Creative
  • Social, Communication
  • Games, Entertainment
  • Utilities, Tools

See categories.md for full taxonomy.

6. Comparison Framework

When user asks to compare apps:

  1. Same category only (don't compare notes app vs game)
  2. Use consistent criteria from compare.md
  3. Declare winner for specific use cases, not overall
  4. Acknowledge "it depends" when true
7. Source Honesty
  • Admit when info might be outdated
  • Recommend checking current reviews for pricing/features
  • Don't invent features — if unsure, say so
Common Traps
  • Recommending most popular instead of best fit → match to user's stated needs
  • Forgetting user said "no subscriptions" → always check ~/Clawic/data/apps/memory.md
  • Recommending apps user already tried and disliked → check ~/Clawic/data/apps/tried.md
  • Overwhelming with options → max 3 recommendations unless asked for more
  • Ignoring platform → always confirm iOS/Android before recommending
按 MIT 许可原样转载,未经改动 · 在 GitHub 查看 →

评论

登录即可评论;带「已验证安装」的,是发布者名下有本店的安装或持有记录。