seo-mcp-tools-expert（SEO MCP tools expert 搜索）

SEO MCP Tools Expert

This skill teaches how to use seo-mcp MCP tools effectively. It does NOT teach SEO concepts — see the other skills for that.

Tool Categories

A. Page Analysis (10 tools — always available, no API key needed)

B. Performance (2 tools — free, optional API key for higher limits)

C. Research (4 tools — require DataForSEO or Moz API key)

D. Google Search Console (3 tools — require OAuth2 setup)

E. Generation (3 tools — always available)

F. Site Crawler (3 tools — always available, no API key needed)

System (1 tool)

Common Workflows

Workflow 1: Full Page Audit

Use when: User says "audit this page", "check my SEO", or provides a URL for analysis.

Step 1: analyze_page(url, { includeContent: true, followRedirects: true })
  → Get overall score and identify major issues

Step 2: analyze_headings(url, { targetKeyword: "..." })
  → Validate heading structure (only if Step 1 shows heading issues)

Step 3: analyze_images(url, { checkFileSize: true })
  → Check image optimization (only if Step 1 shows image issues)

Step 4: check_core_web_vitals(url, { strategy: "mobile" })
  → Get performance metrics

Step 5: extract_schema(url, { validateGoogle: true })
  → Check structured data

Step 6: generate_meta_suggestions(url, { targetKeyword: "..." })
  → Suggest improvements

Key: Start with analyze_page — it's the broadest tool. Only drill down with specific tools if issues are found.

Workflow 2: Full Site Crawl

Use when: User wants to audit the entire site, find all issues, or detect duplicates/redirect chains.

Step 1: crawl_site(url, { maxPages: 100, maxDepth: 10 })
  → Returns crawl ID immediately. Crawl runs in background.

Step 2: crawl_status(crawlId)
  → Poll progress. Repeat until state is "completed".

Step 3: crawl_results(crawlId, { filter: "summary" })
  → Get aggregated issues by type/severity, status code distribution

Step 4: crawl_results(crawlId, { filter: "issues", severity: "high" })
  → Get high/critical issues with page URLs

Step 5: crawl_results(crawlId, { filter: "duplicates" })
  → Find duplicate titles, descriptions, H1s, exact content duplicates

Step 6: crawl_results(crawlId, { filter: "redirects" })
  → Find redirect chains (>2 hops) and loops

Key: The crawler analyzes each page for SEO issues automatically. Use crawl_results with different filters to focus on specific problem areas.

Workflow 3: Technical Site Audit

Use when: User wants to audit the entire site's technical health (without full crawl).

Step 1: analyze_robots_txt(domain)
  → Check crawlability rules

Step 2: analyze_sitemap(domain)
  → Validate sitemap structure

Step 3: analyze_page(homepage, { followRedirects: true })
  → Check homepage technical elements

Step 4: audit_security_headers(homepage)
  → Check HSTS, CSP, mixed content

Step 5: check_core_web_vitals(url, { strategy: "mobile", categories: ["performance", "seo"] })
  → Get Core Web Vitals for key pages

Step 5: check_mobile_friendly(url)
  → Verify mobile usability

Step 6 (if GSC available): gsc_index_coverage(siteUrl)
  → Check real indexing data

Workflow 3: Content Optimization

Use when: User wants to optimize content for a keyword.

Step 1: analyze_page(url, { includeContent: true })
  → Analyze current on-page elements

Step 2: analyze_headings(url, { targetKeyword: "keyword" })
  → Check keyword in headings

Step 3: generate_meta_suggestions(url, { targetKeyword: "keyword", secondaryKeywords: [...] })
  → Get improved meta tags

Step 4 (if API available): research_keywords(["keyword"])
  → Get search volume and related keywords

Step 5 (if API available): analyze_serp("keyword")
  → Understand what Google shows for this query

Workflow 4: Competitive Analysis

Use when: User wants to compare their site against competitors.

Step 1: analyze_page(competitorUrl) — for each competitor
  → Compare on-page elements

Step 2 (if API available): analyze_domain_authority([userDomain, ...competitorDomains])
  → Compare authority metrics

Step 3 (if API available): analyze_backlinks(competitorDomain)
  → Analyze competitor backlink profile

Step 4 (if API available): analyze_serp("target keyword")
  → See who ranks and what features they trigger

Workflow 5: Schema Implementation

Use when: User wants to add or fix structured data.

Step 1: extract_schema(url, { validateGoogle: true })
  → Check existing schema (if any)

Step 2: generate_schema({ type: "Article", data: {...} })
  → Generate correct JSON-LD

Step 3: extract_schema(url)  (after implementation)
  → Validate the new schema

Tool Parameter Reference

analyze_page

analyze_headings

analyze_images

analyze_internal_links

check_core_web_vitals

research_keywords

generate_schema

audit_security_headers

analyze_url_structure

validate_hreflang

crawl_site

crawl_status

crawl_results

detect_orphan_pages

compare_crawls

list_crawls

No parameters. Returns all crawl sessions (in-memory + saved to disk).

audit_accessibility

Common Mistakes

1. Not starting with analyze_page

Always start with analyze_page — it's the broadest tool and helps you decide which specific tools to use next. Don't jump straight to analyze_images or analyze_headings without context.

2. Using research tools without API keys

research_keywords, analyze_serp, analyze_backlinks, and analyze_domain_authority require DataForSEO or Moz API keys. If the user hasn't configured these, the tools will return an error with setup instructions.

How to check: Try the tool — if it fails with an API key error, inform the user and suggest alternatives (manual research, free tools like Google Keyword Planner).

3. Forgetting to specify strategy for Core Web Vitals

check_core_web_vitals defaults to mobile. Always run mobile first (Google uses mobile-first indexing), then optionally run desktop for comparison.

4. Not using targetKeyword with analyze_headings

The keyword check is optional but very useful. Always ask the user for their target keyword before running heading analysis.

5. Using renderJs unnecessarily

renderJs: true requires puppeteer and is much slower. Only use it for single-page applications (React, Vue, Angular) where content is rendered client-side. For most sites, the default cheerio parsing is sufficient and faster.

API Key Configuration

Always available (no keys needed)

All 11 analysis tools, 2 performance tools, 3 generation tools, 6 crawler tools, 1 documentation tool = 23 tools

Optional: PageSpeed API key

Increases rate limits from 25 to 400 requests per 100 seconds. Set: PAGESPEED_API_KEY in environment.

Optional: DataForSEO

Enables keyword research, SERP analysis, backlink analysis, domain authority. Set: DATAFORSEO_LOGIN and DATAFORSEO_PASSWORD in environment.

Optional: Moz

Alternative for backlinks and domain authority. Set: MOZ_ACCESS_ID and MOZ_SECRET_KEY in environment.

Optional: Google Search Console

Enables GSC performance, index coverage, and sitemap tools. Set: GSC_CLIENT_ID, GSC_CLIENT_SECRET, GSC_REFRESH_TOKEN in environment.

Related Skills

• seo-on-page-optimization — what on-page elements to check
• seo-technical-audit — technical audit methodology
• seo-content-strategy — keyword research interpretation
• seo-schema-structured-data — schema generation details
• seo-off-page-backlinks — backlink analysis interpretation
• seo-local-seo — local SEO optimization

See ANALYSIS_GUIDE.md for detailed analysis tool usage. See PERFORMANCE_GUIDE.md for performance tool details. See GENERATION_GUIDE.md for generation tool details. See AUDIT_WORKFLOW_GUIDE.md for complete audit workflow.