Token导航 LogoToken导航

Agent Skill · 研究检索

keyword-research(关键词研究)

keyword-research 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

keyword-research 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,第三方扫描有提醒项,按安装量在研究检索类的 73,675 个 Skill 里排第 9,171。

总安装 14,806周安装 636GitHub Stars 409下载 5,190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:keyword-research(关键词研究)
来源仓库:https://github.com/kostja94/marketing-skills
仓库路径:skills/keyword-research
安装命令:
npx skills add https://github.com/kostja94/marketing-skills --skill keyword-research
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/kostja94/marketing-skills --skill keyword-research

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。

keyword-research 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。适用于关键词挖掘、搜索意图分析和竞品语言策略研究等 SEO 场景。

通过 npx skills add 命令从 GitHub 仓库安装,需确认权限范围和维护状态。使用前建议核对是否会触发联网、命令执行或文件读写操作。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

来源仓库内路径:skills/keyword-research

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/kostja94/marketing-skills --skill keyword-rese…;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

keyword-research 怎么安装?

在终端执行 npx skills add https://github.com/kostja94/marketing-skills --skill keyword-research。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

keyword-research 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。

keyword-research 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。

keyword-research 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。

keyword-research 用的人多吗?

公开数据是累计安装 14,806 次,近一周安装 636 次,下载 5,190 次,GitHub 409 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。

keyword-research 安全吗?

第三方扫描有提醒项,说明存在需要人工确认的行为。安装前请打开源码看清它读了什么、联网到哪里。

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。

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。
  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 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
keyword-research

关键词研究 · 本页

14,806409需要联网未明确有提醒

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

SEO Content: Keyword Research

Guides keyword research for SEO: finding target keywords, assessing difficulty, understanding search intent, and building topical maps. ~95% of keywords get fewer than 10 searches/month; low-volume, high-intent terms often yield faster rankings and conversion.

When invoking: On first use, if helpful, open with 1–2 sentences on what this skill covers and why it matters, then provide the main output. On subsequent use or when the user asks to skip, go directly to the main output.

Initial Assessment

Check for project context first: If .claude/project-context.md or .cursor/project-context.md exists, read it for product, audience, and positioning.

Identify:

  1. Product/service: What you offer
  2. Audience: Who searches for it
  3. Goals: Traffic, conversions, brand
  4. Tool access: Google Keyword Planner, Google Trends, or SEO tools

Discovery Methods

Base Discovery

MethodPurpose
User perspectiveWhat pain points? What would they search? Customer language from product context
Tool expansionRelated keywords, questions, suggestions; Google autocomplete, PAA, Related Searches
Competitor reverseAnalyze competitor titles, H1, URL; identify topics they rank for; find gaps (#4–10 = opportunity) — see competitor-research
Google PAAPeople Also Ask and Related Searches; high-value signals from real user behavior
Extract from articleWhen auditing existing content: extract seed keywords from title, H1, H2s, meta keywords, first 100 words; then search "[primary keyword]" or "[primary keyword] related keywords" for opportunities; use "[primary keyword]" site:competitor.com if competitors known

Google Autocomplete (Long-Tail Discovery)

Google autocomplete reflects real user searches; suggestions only appear if queries have actual traffic. Free; often uncovers low-volume long-tail that keyword tools miss. ~70% of search traffic is long-tail; lower competition, higher conversion.

Alphabet method (seed + space + letter):

  • Type seed keyword + space + each letter: keyword a, keyword b,... keyword z
  • Record relevant suggestions; repeat with numbers 0-9
  • Example: SEO a -> "SEO audit," "SEO agency"; SEO b -> "SEO basics," "SEO best practices"

Position variants (seed in different positions):

  • Prefix: a keyword, b keyword (discover what users add before)
  • Suffix: keyword a, keyword b (most common; alphabet method)
  • Middle: how to keyword a, best keyword for (question + modifier combos)

Question modifiers:

  • how to keyword, what is keyword, why keyword, when to keyword, keyword vs
  • keyword for beginners, keyword for small business, keyword without

Why it works: Keyword tools filter low-volume terms; autocomplete only shows queries with real traffic. Use with PAA and Related Searches for full coverage. Categorize results by intent (informational, commercial, transactional).

Incremental Discovery

  • User feedback: Support, community, reviews, NPS—high-frequency questions = unmet search demand
  • Multi-platform search: Reddit, Quora, X (Twitter), Hacker News—real questions and discussions

Search Intent

IntentContent typeExample
InformationalBlog, guide, FAQ"how to optimize sitemap"
NavigationalBrand page"alignify login"
CommercialComparison, review"SEO tools comparison"
TransactionalProduct, pricing"best SEO tool pricing"

Intent Identification

Modifier words (often signal intent):

IntentModifiers
Informational"how," "what," "why," "guide," "tutorial"
Commercial"best," "compare," "vs," "review," "top"
Transactional"buy," "price," "cheap," "coupon," "free shipping"
LocalLocation names

SERP check: Search the term—knowledge cards/Wiki → informational; product lists/reviews → commercial; brand sites → navigational. Broader terms often show mixed SERP. See serp-features for feature types.

Long-Tail Expansion

  • Google Autocomplete: Alphabet method, position variants, question modifiers; see above. Primary source for long-tail.
  • Intent modifiers: Core + "how," "best," "vs," "compare," "price"
  • Question words: "how to," "what is," "why," "when"
  • Functional modifiers: Core + "-er/-or" (e.g., "image optimizer" for tool-type queries); often higher conversion
  • Clustering: Group by SERP overlap (same top pages), semantic similarity, or intent.

Keyword Clustering & Topical Map

MethodUse
SERP overlapKeywords with overlapping top-ranking pages → same cluster
SemanticGroup by meaning, LSI, related concepts
Intent-basedGroup by intent; separate pages if intent differs within cluster

Pillar–cluster (map keywords to structure):

  • Pillar (Hub): Broad topic page; links to clusters
  • Cluster (Spoke): Focused subtopic; links back to pillar
  • Target long-tail first; then pillar. Interlink clusters within topic.
  • See content-strategy for full pillar-cluster planning and implementation.

Evaluate & Screen

FactorConsider
Search volumeMonthly searches; ~100+/month typical floor; niche can relax
Keyword difficulty (KD)New sites target lower KD
CPCHigher CPC often = stronger commercial intent
SERP featuresFeatured Snippet, PAA, zero-click; SERP features can satisfy intent without click—affects real traffic; see serp-features (Zero-Click section), featured-snippet
Screening order1) Remove irrelevant 2) Filter very low volume 3) Assess achievability 4) Prioritize commercial/transactional

Product Positioning Test (SEO Fit)

Test if positioning is clear enough for search:

  • XXX + Function words: Generator, Creator, Maker, Builder, Changer, Shortener, Scraper, Converter, Downloader, Translator, Extender, Summarizer, Resizer, Remover, Extractor, Recorder, Rewriter, Solver, Calculator; or Platform, Tool, Software, App, Provider, Assistant, Copilot
  • Input + to + Output: e.g., "image to video," "text to speech"—clear input/output signals intent

Agent/Copilot products: Pure native Agent hard to grow via SEO; users rarely search "agent." Release related features first (e.g., CRM, sales bot for sales agent) to build traffic, then funnel to Agent product.

Principles

  • Core rule: Someone must search it—validate with tools; avoid inventing terms
  • Functional keywords: Tool-type (-er/-or) often convert better; users are closer to action
  • Multi-language: Re-research in target language; don't translate existing lists. See translation for translation workflow.

SEO–PPC Keyword Synergy

Keyword research serves both SEO and Google Ads. Align both channels to avoid duplication, cannibalization, and wasted spend.

Data flowUse
keyword-research → google-adsKeyword list, clusters, intent; support terms (login, forum, pricing) → negative keywords for PPC
google-ads → keyword-researchPPC conversion rate, Search Terms report → SEO priority; high-converting PPC terms = worth ranking organically
keyword-research → landing-pageClusters → dedicated LP per intent; PAA questions → FAQ sections
GSC organic rank 4+If you rank well organically, consider reducing/pausing PPC on those terms to avoid cannibalization

PPC data for SEO priority: SEO ROI ≈ (Organic clicks × PPC conversion rate × Customer value) − SEO cost. Use PPC conversion data to validate which keywords to pursue in organic.

Reference: Backlinko – SEO and PPC: 8 Smart Ways to Align

Data Sources

SourceUse
AhrefsKeywords Explorer, Site Explorer
SEMrushKeyword Overview, Organic Research
GSCSearch queries, impressions, clicks
GATraffic by landing page
PostHogFeature/search usage

Report Workflow

  1. Parse — Read Excel/CSV, infer keyword, volume, KD, intent, etc. from headers
  2. Enrich — Web search, visit competitor/product pages; read project-context.md if present
  3. Build — Structure data for report
  4. Generate — Output report in chosen format

Output Format

  • Keyword list with volume, KD, intent
  • Keyword mapping to pages/content
  • Content gaps (competitors rank, you don't)
  • Priority ranking for implementation
  • Topical map (cluster → pillar → page mapping)

Report Structure Reference

SectionContent
Executive SummaryPriorities (top 3)
Keyword OverviewTotal keywords, primary intent, avg KD, content gaps count
Keyword ListKeyword, volume, KD, intent, priority, target page
Keyword MappingPage/URL, target keywords, status
Content GapsKeywords competitors rank for that you don't
Action PlanPriority, action, impact, effort
AppendixSearch intent reference (Informational, Commercial, Transactional, Navigational)

Related Skills

  • seo-strategy: SEO workflow, Product-Led SEO, audit approach; keyword research is Content phase
  • google-ads: Keywords inform Search targeting; PPC data feeds back into SEO priority
  • paid-ads-strategy: When to use paid vs organic; channel selection
  • content-strategy: Keywords inform content plan; topic clusters
  • content-optimization: Keyword placement, density vs stuffing, H2 keywords
  • title-tag, meta-description: Keywords in title, description
  • heading-structure: Keywords in H1, H2
  • link-building: Keywords inform link targets
  • serp-features: SERP features in keyword screening; PAA, Featured Snippet
  • featured-snippet: Snippet-worthy query targeting
  • competitor-research: Competitor keyword/topic analysis; reverse engineering
  • faq-page-generator: PAA questions to FAQ sections; question-based keyword to FAQ content

平台分布

Codex

37.59%
按下载量换算1,951

Claude

28.28%
按下载量换算1,468

Cursor

19.11%
按下载量换算992

Gemini CLI

9.13%
按下载量换算474

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

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。来源安全扫描存在提醒或未通过项,安装前请自行复核。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