mcpField
K
keyword-research

keyword-research

github.com/majesticlabs-devMarketing & growth5051 installsUpdated 7w ago
Repository

About keyword-research

keyword-research is a skill for content teams that need to find and prioritize topics using free tools and structured research methods. It combines topic clustering, search-result review, and audience-question mining.

Key features

  1. 01Six Circles topic clustering
  2. 02Free-tool research workflows
  3. 03Manual search-result competition analysis
  4. 04Keyword scoring and prioritization
  5. 05Reddit and community question mining
  6. 06Structured content-plan output

Use cases

  1. 01Planning topic clusters from a seed keyword
  2. 02Finding audience questions for article ideas
  3. 03Comparing competition across target keywords
  4. 04Prioritizing topics for a content queue

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/majesticlabs-dev/majestic-marketplace --skill keyword-research

After installing, type/keyword-researchin Claude Code to confirm it loaded.

Details

From the project README

Keyword Research Without Expensive Tools

Find content topics and prioritize them using free tools and systematic frameworks.

Required Input

Ask the user for:

  1. Niche/Industry - What space are you in?
  2. Target Audience - Who are you writing for?
  3. Business Goal - Traffic, leads, sales, authority?
  4. Existing Content - Do you have a site/blog already?

The 6 Circles Method

Generate 13 content ideas from a single seed keyword using nested topic clusters.

Structure
                    ┌─────────────────────┐
                    │   PRIMARY KEYWORD   │  ← 1 pillar topic
                    │   (Largest Circle)  │
                    └─────────────────────┘
                              │
            ┌─────────────────┼─────────────────┐
            ▼                 ▼                 ▼
     ┌───────────┐     ┌───────────┐     ┌───────────┐
     │ Sub-topic │     │ Sub-topic │     │ Sub-topic │  ← 3 supporting themes
     │     A     │     │     B     │     │     C     │
     └───────────┘     └───────────┘     └───────────┘
            │                 │                 │
      ┌─────┼─────┐     ┌─────┼─────┐     ┌─────┼─────┐
      ▼     ▼     ▼     ▼     ▼     ▼     ▼     ▼     ▼
     [1]   [2]   [3]   [4]   [5]   [6]   [7]   [8]   [9]  ← 9 content pieces
Process
  1. Identify Primary Keyword - High-traffic, high-intent topic in your niche
  2. Brainstorm 3 Sub-topics - Related themes that support the primary
  3. Generate 3 Pieces per Sub-topic - Specific articles addressing aspects of each
Content Types for Supporting Pieces
TypeExample
How-to Guide"How to [achieve outcome] with [method]"
Comparison"[Option A] vs [Option B]: Which is better for [use case]"
List Post"[Number] [adjective] ways to [achieve result]"
Case Study"How [persona] achieved [result] using [approach]"
Trend Analysis"[Topic] trends in [year]: What's changing"
Expert Roundup"[Number] experts share their [topic] strategies"
Beginner Guide"[Topic] for beginners: Everything you need to know"
Tool Review"Best [category] tools for [audience/use case]"
Problem-Solution"Why [problem happens] and how to fix it"

Free Tool Arsenal

Demand Research
ToolUse ForHow to Access
Google Keyword PlannerVolume ranges, competitionads.google.com (free account, no spend)
Google TrendsSeasonality, rising topicstrends.google.com
Ahrefs Free Generator150 suggestions with difficultyahrefs.com/keyword-generator
Ubersuggest3 searches/day with metricsneilpatel.com/ubersuggest
Intent Discovery
ToolUse ForHow to Access
Google AutocompleteReal-time user queriesType in Google search bar
People Also AskQuestion-based contentScroll down in Google SERP
Answer the PublicVisual question mapsanswerthepublic.com (limited free)
AlsoAskedPAA clusteringalsoasked.com (limited free)
Community Mining
SourceWhat to Extract
RedditPain points, questions, terminology
QuoraQuestion patterns, answer gaps
Industry ForumsNiche-specific problems
YouTube CommentsObjections, confusion points
Amazon ReviewsCustomer language, unmet needs

Research Protocol

Step 1: Seed Generation (50-100 keywords)
## Seed Keywords for [Niche]

### From Google Autocomplete
- [keyword] + a-z variations
- "how to [keyword]"
- "best [keyword] for"
- "[keyword] vs"

### From People Also Ask
- [question 1]
- [question 2]
- [question 3]

### From Reddit r/[subreddit]
- Top posts this month: [themes]
- Common questions: [list]
- Pain points mentioned: [list]

### From Answer the Public
- Questions: [list]
- Prepositions: [list]
- Comparisons: [list]
Step 2: Competition Assessment (Manual SERP Analysis)

For each target keyword, analyze top 10 results:

SignalWhat to Look ForOpportunity If...
Content DepthWord count, comprehensivenessTop results are thin (<1000 words)
FreshnessPublication date, last updateTop results are 2+ years old
Format MatchDoes format match intent?Results don't match searcher need
Authority GapDomain strength of rankersSmall sites ranking (not all big brands)
Content GapsMissing subtopics, unanswered questionsPAA questions not covered
Step 3: Prioritization Matrix

Score each keyword 1-10:

KeywordDemandCompetitionIntent FitOpportunity
[keyword][1-10][1-10, lower=easier][1-10]Demand - Competition

Prioritize: Opportunity score >3, Intent fit >6

Sweet Spot Targets
Site StageMonthly Search VolumeCompetition Level
New site (0-6 months)100-500Low
Growing site (6-18 months)500-2,000Low-Medium
Established site (18+ months)1,000-10,000Medium

Output Format

# Keyword Research: [Niche/Topic]

## 6 Circles Content Plan

### Primary Keyword: [keyword]
- Monthly searches: [range]
- Competition: [low/medium/high]
- Intent: [informational/commercial/transactional]

### Sub-topic A: [theme]
1. [Article idea] - [target keyword] - [volume]
2. [Article idea] - [target keyword] - [volume]
3. [Article idea] - [target keyword] - [volume]

### Sub-topic B: [theme]
4. [Article idea] - [target keyword] - [volume]
5. [Article idea] - [target keyword] - [volume]
6. [Article idea] - [target keyword] - [volume]

### Sub-topic C: [theme]
7. [Article idea] - [target keyword] - [volume]
8. [Article idea] - [target keyword] - [volume]
9. [Article idea] - [target keyword] - [volume]

## Prioritized Content Queue

| Priority | Topic | Keyword | Volume | Difficulty | Opportunity |
|----------|-------|---------|--------|------------|-------------|
| 1 | [topic] | [keyword] | [vol] | [1-10] | [score] |
| 2 | [topic] | [keyword] | [vol] | [1-10] | [score] |
| 3 | [topic] | [keyword] | [vol] | [1-10] | [score] |

## Quick Wins (Low Competition, Decent Volume)
- [keyword 1] - [rationale]
- [keyword 2] - [rationale]
- [keyword 3] - [rationale]

## Content Gaps Found
- [gap 1] - competitors missing [what]
- [gap 2] - outdated content on [topic]
- [gap 3] - no good [format] exists for [query]

## Reddit/Community Insights
- Top pain point: [description]
- Common question: [question]
- Terminology used: [terms]

## Next Steps
1. [ ] Create pillar content for [primary keyword]
2. [ ] Write [specific article] targeting [quick win keyword]
3. [ ] Research [sub-topic] deeper for more ideas

Reddit Mining Protocol

Search Queries
site:reddit.com "[niche]" + "how do I"
site:reddit.com "[niche]" + "help"
site:reddit.com "[niche]" + "struggling with"
site:reddit.com "[niche]" + "recommend"
site:reddit.com "[niche]" + "vs"
What to Extract
  • Questions asked repeatedly → Content opportunities
  • Complaints about existing solutions → Angle differentiation
  • Terminology and slang → Keyword variations
  • Recommended resources → Competitor analysis
  • Upvote patterns → Topic demand signals

Integration with Other Skills

After This SkillUse For
content-plannerDetailed outlines for prioritized topics
bofu-keywordsBottom-of-funnel conversion keywords
content-calendarScheduling the content queue
content-writerWriting the actual content

What This Skill Does NOT Do

  • Provide exact search volume (use Keyword Planner for ranges)
  • Guarantee rankings
  • Replace comprehensive SEO audits
  • Analyze existing site performance (use Google Search Console)

Comments

Markdown supported