A SERP-overlap SEO skill that turns seed keywords into validated hub-and-spoke content architectures, link matrices, briefs, and interactive cluster maps.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install seo-cluster
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install seo-cluster using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Semantic Topic Clustering?
Semantic Topic Clustering is an Openclaw Skills workflow for planning SEO content architecture based on how Google actually ranks keywords. Instead of grouping terms by text similarity, it compares shared top-10 organic SERP results to determine which keywords belong on the same page, in the same cluster, or in separate content areas.
The skill expands a seed keyword, classifies search intent, selects pillar and spoke pages, checks for cannibalization, creates bidirectional internal-link structures, and generates both machine-readable plans and interactive visualizations. When the claude-blog skill is available, it can also execute the plan and create content; otherwise, it produces detailed briefs for editorial or AI-assisted writing.
Semantic Topic Clustering Use Cases
- Build a complete topic cluster from a seed keyword or URL.
- Design pillar pages and supporting spoke content using hub-and-spoke SEO architecture.
- Group keywords according to shared Google SERP results rather than superficial wording similarity.
- Import an existing SEO strategy and enrich it with SERP-overlap analysis.
- Identify duplicate primary keywords and reduce content cannibalization before publishing.
- Generate internal-link matrices that connect pillars, spokes, and adjacent clusters.
- Produce content briefs when claude-blog is not installed.
- Execute prioritized content creation with claude-blog, starting with the pillar page.
- Regenerate an interactive cluster map as planned posts become published.
- Create a post-execution scorecard covering coverage, links, orphan pages, images, and content gaps.
How Semantic Topic Clustering Works
- Accept a seed or strategy: Run a fresh plan with a seed keyword or import the latest SEO plan with
--from strategy. - Expand the keyword set: Use related searches, People Also Ask questions, long-tail modifiers, question variants, and commercial terms to create approximately 30–50 unique keywords.
- Collect SERP evidence: Search candidate keyword pairs and compare their top-10 organic URLs. When available, DataForSEO MCP can provide advanced SERP data after a cost check.
- Cluster by overlap: Treat 7–10 shared results as the same target page, 4–6 as the same cluster, 2–3 as an interlinking relationship, and 0–1 as separate coverage.
- Classify intent: Label keywords as informational, commercial, transactional, or navigational. Exclude navigational terms and flag ambiguous intent for review.
- Create the architecture: Select the strongest pillar keyword, organize two to five subtopic clusters, assign two to four spokes per cluster, choose page templates, and set word-count targets.
- Check cannibalization: Merge keywords with very high SERP overlap and ensure every post has a distinct primary keyword.
- Build the link matrix: Connect every spoke to the pillar and the pillar to every spoke, then add sibling and cross-cluster links while preventing orphan pages.
- Generate the cluster map: Populate the HTML visualization with pillar, cluster, post, link, status, and summary metadata.
- Execute or brief: Use claude-blog for prioritized content creation with resume support, or write individual markdown briefs when the writing skill is unavailable.
- Measure completion: Generate a scorecard and optionally regenerate the map to reflect written-versus-planned post status.
Semantic Topic Clustering Setup
No separate package installation command is defined in the skill specification. Make the skill available to the agent, then run it from the directory where plans and generated files should be stored.
Basic planning
/seo cluster plan <seed-keyword>
Import an existing SEO strategy
/seo cluster plan --from strategy
The strategy import searches the current directory for files matching patterns such as *SEO*Plan*, *strategy*, and *content-strategy*. If no strategy is found, provide a seed keyword or run /seo plan first.
Execute a plan
/seo cluster execute
For automatic article creation, ensure the claude-blog skill is installed at ~/.claude/skills/blog/SKILL.md. Without it, execution creates content briefs in cluster-briefs/ instead.
Regenerate the visualization
/seo cluster map
Optional integrations include DataForSEO MCP for SERP collection, plus related SEO skills for content quality, schema markup, reporting, and strategy import. DataForSEO cost checks run before API batches; blocked or unavailable API requests fall back to WebSearch when possible.
Semantic Topic Clustering Data Schema & Taxonomy
The skill writes all artifacts to the current working directory.
| File or directory | Purpose |
|---|---|
cluster-plan.json |
Complete machine-readable plan containing keywords, intent, architecture, posts, links, and execution metadata. |
cluster-plan.md |
Human-readable summary of the planned pillar, clusters, posts, and linking strategy. |
cluster-map.html |
Interactive SVG-based visualization of the cluster structure and link relationships. |
cluster-briefs/ |
Individual markdown briefs generated when claude-blog is unavailable. |
cluster-scorecard.md |
Post-execution quality report with coverage and implementation metrics. |
The interactive map uses a CLUSTER_DATA object with this structure:
{
pillar: { title, keyword, volume, template, wordCount, url },
clusters: [
{
name,
color,
posts: [{ title, keyword, volume, template, wordCount, url, status }]
}
],
links: [{ from, to, type }],
meta: { totalPosts, totalClusters, totalLinks, estimatedWords }
}
The internal-link matrix is represented as a JSON adjacency list. Link records include from, to, type, and anchor. Supported relationship types include mandatory pillar links, within-cluster spoke links, and cross-cluster links.
The metadata taxonomy includes:
- Intent: informational, commercial, transactional, or navigational.
- Page role: pillar or spoke.
- Template: ultimate-guide, how-to, listicle, explainer, comparison, review, best-of, or landing-page.
- Status: planned or written.
- Architecture metrics: total posts, clusters, links, estimated words, search volume, and target URLs.
- Quality metrics: coverage, link density, orphan pages, cannibalization conflicts, image count, pillar-link completion, cross-link completion, and content gaps.
Recommended targets are 2,500–4,000 words for a pillar page, 1,200–1,800 words for each spoke, at least three incoming links per post, zero orphan pages, and no duplicate primary keywords.
Semantic Topic Clustering Advanced Features
- SERP-overlap clustering uses shared organic results as the primary relevance signal, avoiding unreliable text-only keyword grouping.
- Intent pre-grouping reduces pairwise comparisons and improves efficiency for larger keyword sets.
- DataForSEO MCP support can replace WebSearch with
serp_organic_live_advancedafter per-batch cost approval. - Automatic fallback to WebSearch or intent-only clustering supports resilience when SERP data is unavailable.
- Strategy import reuses existing
/seo planoutput while adding fresh SERP validation and enrichment. - Cannibalization detection merges keywords with seven or more shared results into one target page.
- Bidirectional pillar-to-spoke linking, sibling links, cross-cluster links, anchor guidance, and orphan-page checks enforce a robust internal-link architecture.
- Interactive HTML visualization makes cluster relationships, post status, link counts, and estimated word volume easy to inspect.
- claude-blog integration creates content in priority order, beginning with the pillar and then ranking spokes by volume.
- Resume support scans existing output and avoids rewriting posts that have already been completed.
- Backward-link placeholder injection adds newly published URLs to earlier content after each post is written.
- Automatic brief generation preserves the full editorial plan when automated writing is unavailable.
- Cluster scorecards measure coverage, link density, orphan pages, cannibalization, image usage, cross-links, and content gaps.
- Cross-skill workflows connect SEO planning, E-E-A-T validation, schema markup, DataForSEO collection, and PDF reporting.
SKILL.md
Loading
Related Openclaw Skills
TicketSource Connector
Use TicketSource through a secure OOMOL connector to discover events, performance dates, and venues without handling API credentials.
Software Delivery Team
A six-role AI delivery team that turns software specifications into tested, reviewed, security-checked, and release-ready code.
DEX Token Analysis
An AI-assisted DEX token analysis skill for holder clustering, advanced trader tracking, liquidity-pool research, and structured crypto market insights.
Financial Literacy Q&A
An AI-powered financial literacy and A-share analysis skill for answering questions, personalizing education, evaluating risk, and supporting investment workflows.
Brand Mention Scanner
Brand Mention Scanner measures how strongly a brand is represented across the platforms that AI systems use for entity recognition, citations, and recommendations.
ComPDF PDF to Markdown
Convert PDF files into accurate, structured Markdown through the official ComPDF Server API.
Featured*
