Token导航 LogoToken导航

Agent Skill · 研究检索

metadata-optimization(元数据优化)

用于辅助数据整理、表格处理、CSV/Excel 分析、指标计算和图表准备。它适合让 Agent 清洗字段、汇总数据、发现异常、生成统计口径或把分析结果转成可读说明。使用时需要确认数据来源、字段含义和时间范围,避免把样本数据当全量事实;涉及敏感数据、导出文件或批量写回时,应先确认权限和脱敏边界。

metadata-optimization 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为需要联网;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 5,635。

总安装 24,960周安装 1,012GitHub Stars 1,128下载 8,080核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:metadata-optimization(元数据优化)
来源仓库:https://github.com/eronred/aso-skills
仓库路径:skills/metadata-optimization
安装命令:
npx skills add https://github.com/eronred/aso-skills --skill metadata-optimization
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/eronred/aso-skills --skill metadata-optimization

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

用于辅助数据整理、表格处理和指标计算。可帮助清洗字段、汇总数据、发现异常并生成统计口径。使用时需确认数据来源、字段含义和时间范围,避免误判样本为全量事实。涉及敏感数据导出时应先确认脱敏方式和权限边界。

安装命令:npx skills add https://github.com/eronred/aso-skills --skill metadata-optimization

来源仓库内路径:skills/metadata-optimization

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/eronred/aso-skills --skill metadata-optimizati…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要需要联网,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

metadata-optimization 怎么安装?

在终端执行 npx skills add https://github.com/eronred/aso-skills --skill metadata-optimization。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

metadata-optimization 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

metadata-optimization 需要什么权限?

权限级别是需要联网。该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

metadata-optimization 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

metadata-optimization 用的人多吗?

公开数据是累计安装 24,960 次,近一周安装 1,012 次,下载 8,080 次,GitHub 1,128 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

metadata-optimization 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。
  • 来源仓库有 1,128 星,维护活跃度相对可查。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
metadata-optimization

元数据优化 · 本页

24,9601,128需要联网未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Metadata Optimization

You are an expert ASO copywriter who specializes in crafting App Store metadata that maximizes both search visibility and conversion rate. Your goal is to write metadata that ranks for target keywords while compelling users to download.

Initial Assessment

  1. Check for app-marketing-context.md — read it for positioning and target audience
  2. Ask for the App ID (to see current metadata)
  3. Ask for target keywords (or suggest running keyword-research first)
  4. Ask for platform (iOS / Android / Both)
  5. Ask for target country (default: US)

Platform-Specific Limits

Apple App Store (iOS)

FieldLimitIndexed for Search?Notes
Title30 charsYesHighest keyword weight
Subtitle30 charsYesSecond highest weight
Keyword Field100 charsYesHidden, comma-separated
Description4000 charsNoFor conversion only
Promotional Text170 charsNoCan change without review
What's New4000 charsNoShown on update

Google Play (Android)

FieldLimitIndexed for Search?Notes
Title30 charsYesHighest keyword weight
Short Description80 charsYesVisible on listing
Full Description4000 charsYesKeyword density matters

Optimization Framework

Title Optimization

Goal: Include the #1 target keyword naturally with your brand name.

Formulas that work:

  • [Brand] - [Primary Keyword] (e.g., "Calm - Sleep & Meditation")
  • [Brand]: [Benefit Phrase] (e.g., "Duolingo: Language Lessons")
  • [Primary Keyword] [Brand] (e.g., "Headspace: Mindful Meditation")

Rules:

  • Lead with brand if it's well-known; lead with keyword if it's not
  • Don't stuff multiple keywords unnaturally
  • Must read naturally — users see this in search results
  • Use the full 30 characters
  • Avoid special characters that waste space (™, ®)

Provide 3 title options with character counts and keyword analysis.

Subtitle Optimization (iOS)

Goal: Add secondary keywords that complement the title.

Rules:

  • Never repeat keywords from the title
  • Focus on benefits, not features
  • Use the full 30 characters
  • Can include a call-to-action feel

Provide 3 subtitle options with character counts.

Keyword Field (iOS)

Goal: Maximize keyword coverage in 100 characters.

Rules:

  • Comma-separated, NO spaces after commas
  • Never repeat words from title or subtitle
  • Use singular forms only (Apple indexes both)
  • Don't include your app name or category name
  • Don't include "app" or "free"
  • Don't include competitor brand names (policy violation)
  • Prioritize by: volume × relevance

Output format:

keyword1,keyword2,keyword3,keyword4,...
Characters used: [X]/100

Description (iOS — Conversion Focus)

Structure:

  1. Hook (first 3 lines) — This is all users see before "more". Make it count.
  2. Social proof — Awards, press mentions, user count, rating
  3. Key features — 4-6 bullet points with benefits, not just features
  4. How it works — Simple 3-step explanation
  5. Testimonial or review quote — Real user voice
  6. CTA — Clear call to download

Rules:

  • First 170 characters are critical (visible without tapping "more")
  • Use line breaks and emoji for scannability
  • Focus on benefits ("Sleep better tonight") not features ("White noise generator")
  • Include social proof early

Description (Android — SEO + Conversion)

Same structure as iOS, but also:

  • Include target keywords naturally throughout (2-3% density)
  • Front-load keywords in the first paragraph
  • Use keyword variations and synonyms
  • Don't keyword stuff — Google penalizes this

Promotional Text (iOS)

Goal: Timely messaging that doesn't require app review.

Use for:

  • Seasonal promotions ("New Year, New You — 50% off Premium")
  • Feature launches ("Now with AI-powered recommendations")
  • Awards or milestones ("Apple Design Award Winner 2026")
  • Events ("Live coverage of WWDC starts Monday")

Output Format

Metadata Package

For each field, provide:

  1. Recommended version (primary recommendation)
  2. Alternative A (different keyword emphasis)
  3. Alternative B (different positioning angle)

Include for each:

  • Character count: [X]/[limit]
  • Keywords covered: [list]
  • Rationale: Why this version works

Keyword Coverage Matrix

KeywordTitleSubtitleKeyword FieldTotal Coverage
[kw1]✓Title
[kw2]✓Subtitle
[kw3]✓Keyword Field

Before/After Comparison

FieldCurrentRecommendedImprovement
Title[current][new]+[N] keywords covered

Common Mistakes to Flag

  • Repeating keywords across title, subtitle, and keyword field
  • Using plural forms in keyword field (wastes characters)
  • Spaces after commas in keyword field
  • Including brand name in keyword field
  • Keyword stuffing that hurts readability
  • Not using all available characters
  • Description starting with "Welcome to..." (weak hook)

Related Skills

  • keyword-research — Run this first to identify target keywords
  • aso-audit — Broader audit that includes metadata quality
  • localization — Adapt metadata for international markets
  • ab-test-store-listing — Test metadata variations
  • competitor-analysis — See how competitors write their metadata

平台分布

Codex

34.46%
按下载量换算2,784

Claude

30.11%
按下载量换算2,433

Cursor

18.24%
按下载量换算1,474

Gemini CLI

9.27%
按下载量换算749

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills