| name | seo-content |
| description | Keyword research and content optimization for any market. Works without paid API subscriptions using web search. Optionally integrates Ahrefs/Semrush MCP or DataForSEO (seo-research skill) for richer keyword data.
Use when: (1) researching keywords for a topic or page, (2) creating a content brief or editorial plan, (3) optimizing existing page content for SEO+AEO, (4) analyzing SERP competitors for a keyword, (5) clustering keywords by intent, (6) user says "keyword research", "content brief", "optimize content", "seo content", "find keywords", "editorial plan", "подбери ключи", "контент-план", "оптимизируй контент".
Supports any locale including Russia (Yandex-based research) and works even without API keys by using web search as fallback.
|
SEO Content Skill
Keyword research + content strategy + page optimization. Works for any market with WebSearch fallback.
Step 1: Determine Context
Ask user (use AskUserQuestion):
- Topic: What topic, page, or seed keywords to research?
- Locale: Target market? (Russia, USA, global, other)
- Goal: New content, optimize existing page, or editorial plan?
- MCP: Ahrefs or Semrush connected? (optional)
Based on locale, load locale reference:
- Russia →
.agents/skills/_seo-shared/references/locale-ru.md
- English →
.agents/skills/_seo-shared/references/locale-en.md
- Other →
.agents/skills/_seo-shared/references/locale-generic.md
Step 2: Detect Data Sources
Read .agents/skills/_seo-shared/references/mcp-integration.md for detection logic.
Priority order:
- Ahrefs MCP → try
explore_keywords_overview
- Semrush MCP → try
keyword_overview
- DataForSEO (seo-research skill) → check if project has
.agents/skills/seo-research/ AND market is NOT sanctioned (Russia, Belarus, etc.)
- WebSearch → always available
Record available sources. Proceed with the highest-priority available source.
Step 3: Keyword Discovery
Launch keyword-discovery-agent:
Task: Find keywords related to "{topic}" for {locale} market.
Use the best available data source.
IF Ahrefs MCP available:
- Call explore_keywords_overview for seed keywords
- Call explore_matching_terms for expansion
- Extract: keyword, volume, difficulty, CPC
IF Semrush MCP available:
- Call keyword_overview for seed keywords
- Call related_keywords for expansion
- Extract: keyword, volume, difficulty
IF DataForSEO available AND market supported:
- Suggest using /seo-research skill for deep pipeline
- Provide seed keywords for the pipeline
FALLBACK (WebSearch — always works):
- WebSearch: "{topic}" — collect related searches
- WebSearch: "{topic} keywords" — find keyword lists
- WebSearch: "{topic} frequently asked questions"
- WebSearch: "{topic} popular queries {current_year}"
- WebSearch: "people also ask {topic}"
- WebSearch: "site:reddit.com {topic}" — real user questions
For Russia specifically:
- WebSearch: "wordstat {keyword}" — Yandex Wordstat data
- WebSearch: "{keyword} частотность яндекс"
- WebSearch: "site:vc.ru {topic}" — Russian community questions
Estimate volume as HIGH/MEDIUM/LOW:
- HIGH: term appears in autocomplete, many dedicated pages
- MEDIUM: some pages, appears in related searches
- LOW: few results, niche term
Output: JSON array of keywords
[
{"keyword": "...", "volume": N or "HIGH|MEDIUM|LOW", "difficulty": N or null, "source": "ahrefs|semrush|websearch", "intent": "informational|transactional|navigational|commercial"}
]
Aim for 30-100 keywords minimum.
Model: opus. Tools: WebSearch, Ahrefs MCP (optional), Semrush MCP (optional).
Step 4: Cluster by Intent
Launch keyword-cluster-agent:
Task: Group these keywords into semantic clusters by search intent.
Keywords: {JSON array from Step 3}
Rules:
1. Group by user INTENT, not just lexical similarity
2. Each cluster = one potential page/article
3. Assign intent type to each cluster:
- informational: "how to", "what is", "guide", "tutorial"
- transactional: "buy", "price", "order", "discount"
- navigational: brand names, specific product names
- commercial: "best", "top", "review", "comparison", "vs"
4. For each cluster, identify:
- Primary keyword (highest volume)
- Supporting keywords (related terms)
- Estimated total cluster volume
5. Rank clusters by opportunity: volume × (1 / (difficulty + 1))
Output: JSON array of clusters
[
{
"name": "cluster short name",
"intent": "informational|transactional|navigational|commercial",
"primary_keyword": "...",
"keywords": ["kw1", "kw2", ...],
"total_volume": N or "HIGH|MEDIUM|LOW",
"opportunity_rank": 1
}
]
Model: opus. Tools: none (pure reasoning).
Step 5: SERP Analysis
For top 5 clusters by opportunity, launch serp-analysis-agent (parallel):
Task: Analyze search results for "{primary_keyword}" in {locale} market.
Steps:
1. WebSearch for the exact keyword
2. For top 5 organic results:
a. WebFetch each page
b. Extract: title, H1, word count, heading structure (H2-H3), Schema types used, content format
c. Note: unique angle, strengths, weaknesses
3. Identify content gaps — what top results MISS
4. Identify common patterns — what ALL top results DO
5. Determine recommended approach
Output: JSON
{
"keyword": "...",
"serp_landscape": [
{
"position": 1,
"url": "...",
"title": "...",
"word_count": N,
"headings": ["H2: ...", "H2: ...", "H3: ..."],
"schema_types": ["Article", "FAQPage"],
"content_format": "guide|listicle|comparison|tutorial|review",
"unique_angle": "..."
}
],
"common_patterns": ["all include pricing tables", "all have FAQ section"],
"content_gaps": ["none cover X", "outdated data on Y"],
"recommended_approach": "...",
"recommended_word_count": N,
"recommended_schema": "Article|TechArticle|HowTo|FAQPage",
"recommended_format": "guide|listicle|comparison|tutorial"
}
Model: opus. Tools: WebSearch, WebFetch.
Step 6: Generate Output
Based on user's goal from Step 1:
Goal: New Content → Content Brief
Read references/content-brief-template.md for format.
Generate a complete brief using SERP analysis results, keyword clusters, and AEO requirements from .agents/skills/_seo-shared/references/aeo-checklist.md.
Goal: Optimize Existing Page
Launch content-optimizer-agent:
Task: Compare the existing page at {url_or_file} against SERP analysis for "{primary_keyword}".
Steps:
1. Read the existing page (WebFetch URL or Read local file)
2. Compare with SERP winners from serp-analysis results:
- Title: length, keyword presence
- Meta description: length, CTA
- Word count vs SERP average
- Heading structure vs SERP pattern
- Schema.org vs what competitors use
- Content gaps vs what competitors cover
- AEO compliance (autonomous fragments, semantic HTML)
3. Generate specific, actionable changes
Output format:
# Content Optimization: {Page Title}
## Current State
- Word count: X (SERP average: Y)
- SEO Score: X/100
- Keywords targeting: [list]
## vs SERP Winners
| Metric | Your Page | SERP Average | Gap |
|--------|-----------|-------------|-----|
## Specific Changes
### Title
BEFORE: "..."
AFTER: "..."
### Meta Description
BEFORE: "..."
AFTER: "..."
### Content Additions
1. Add section: "[H2]" — competitors cover this, you don't
2. Expand section: "[H2]" — currently X words, recommend Y+
### Schema.org
(generated JSON-LD if missing)
Model: opus. Tools: WebFetch, Read, Grep.
Goal: Editorial Plan
Generate a content calendar mapping keyword clusters to pages.
# Editorial Plan: {Topic Area}
## Keyword Clusters (ranked by opportunity)
### Cluster 1: {Name} — {Intent} — {Volume}
| Keyword | Volume | Difficulty |
|---------|--------|-----------|
| {primary} | ... | ... |
| {supporting} | ... | ... |
### Cluster 2: ...
## Content Calendar
| # | Page Title | Target Cluster | Format | Word Count | Schema | Priority |
|---|------------|---------------|--------|------------|--------|----------|
## Internal Linking Strategy
- {Page A} links to {Page B} because...
- Hub page: {URL} links to all cluster pages
## Quick Wins (low difficulty, high volume)
1. ...
## Long-term Plays (high difficulty, high volume)
1. ...
