Install any skill in seconds. Free to start, no credit card required.
Get Started Free →SEO covering technical SEO, content optimization, link building, keyword research, and search analytics. Use when auditing technical SEO, researching keywords, optimizing on-page elements, or analyzing organic search performance.
.claude/skills/borghei-seo-specialist/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-13 | ✗→✓ | ▲ Improved | 90% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 135% | 0% |
| case-01 | ✗→✓ | ▲ Improved | 108% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 322% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 43% | 0% |
The agent operates as a senior SEO specialist, delivering technical audits, keyword strategies, on-page optimization, link building plans, and performance analysis for organic search growth.
Before starting, confirm these inputs. If any is unknown or vague, ASK — do not assume:
Stop rule: ask only the 2-3 that most change the output. If the user says "just draft it," proceed and list your assumptions at the top of the artifact.
Crawlability:
Performance (Core Web Vitals):
| Metric | Good | Needs Improvement | Poor | |--------|------|-------------------|------| | LCP (Largest Contentful Paint) | < 2.5s | 2.5s - 4s | > 4s | | FID (First Input Delay) | < 100ms | 100 - 300ms | > 300ms | | CLS (Cumulative Layout Shift) | < 0.1 | 0.1 - 0.25 | > 0.25 |
Structure:
| Metric | Good | Moderate | Difficult | |--------|------|----------|-----------| | Volume | 1000+ | 100-1000 | < 100 | | Difficulty | < 30 | 30-60 | > 60 | | CPC (commercial signal) | > $5 | $1-5 | < $1 |
| Intent | Signal Words | Content Type | |--------|-------------|-------------| | Informational | "how to", "what is", "guide" | Blog posts, tutorials | | Navigational | Brand names, product names | Homepage, product pages | | Commercial | "best", "reviews", "vs" | Comparison pages, reviews | | Transactional | "buy", "discount", "pricing" | Product pages, landing pages |
Title Tag: primary keyword front-loaded, 50-60 characters, compelling for CTR Meta Description: includes keyword, clear value prop, CTA, 150-160 characters Headings: H1 contains primary keyword, H2s contain secondary keywords, logical hierarchy Content: keyword in first 100 words, natural density, related terms (LSI), comprehensive coverage Images: descriptive filenames, keyword-rich alt text, compressed, lazy-loaded
html<!-- Title: 58 chars, keyword front-loaded --> <title>Cloud Cost Optimization: 7 Strategies That Cut AWS Bills 40%</title> <!-- Meta: 155 chars, keyword + value prop + CTA --> <meta name="description" content="Learn 7 proven cloud cost optimization strategies used by 500+ engineering teams. Reduce AWS spend by 40% without sacrificing performance. Free checklist inside."> <!-- Schema markup for article --> <script type="application/ld+json"> { "@context": "https://schema.org", "@type": "Article", "headline": "Cloud Cost Optimization: 7 Strategies That Cut AWS Bills 40%", "author": {"@type": "Person", "name": "Jane Chen"}, "datePublished": "2026-02-15", "publisher": { "@type": "Organization", "name": "CloudOps Weekly" } } </script> <h1>Cloud Cost Optimization: 7 Strategies That Cut AWS Bills 40%</h1> <p>Cloud cost optimization is the #1 priority for engineering leaders in 2026...</p> <h2>1. Right-Size EC2 Instances Using Usage Data</h2> <h3>How to Identify Oversized Instances</h3> <h2>FAQ</h2> <h3>What is cloud cost optimization?</h3> <h3>How much can cloud optimization save?</h3>
| Factor | High Quality | Low Quality | |--------|-------------|-------------| | Domain Authority | 50+ | < 20 | | Relevance | Same industry | Unrelated | | Traffic | Active site | Dead site | | Link Type | Editorial | Paid/Spam | | Anchor Text | Natural variation | Exact match spam |
SEO Performance - March 2026
Organic Traffic: 125,432 (+12% MoM)
Rankings: Top 3: 45 | Top 10: 234
Conversions: 542 (+15% MoM)
Top Growing Keywords
1. "cloud cost optimization" - #8 -> #3 (+5)
2. "aws billing alerts" - #15 -> #7 (+8)
3. "kubernetes autoscaling" - New -> #12
Technical Health
Core Web Vitals: Pass | Index: 1,234 pages | Crawl Errors: 3bash# Site audit python scripts/site_audit.py --url https://example.com --output audit.html # Keyword research python scripts/keyword_research.py --seed "cloud computing" --output keywords.csv # Rank tracker python scripts/rank_tracker.py --keywords keywords.csv --domain example.com # Backlink analyzer python scripts/backlink_analyzer.py --domain example.com --output links.csv
references/technical_seo.md - Technical SEO guidereferences/keyword_research.md - Keyword research methodsreferences/link_building.md - Link building playbookreferences/algorithm_updates.md - Google update history| Problem | Likely Cause | Fix | |---------|-------------|-----| | Rankings dropped after Google core update | E-E-A-T signals insufficient or content quality below new thresholds | Audit content against December 2025 / March 2026 core update criteria — add experience signals, author credentials, original data | | High impressions but low CTR | Title tags and meta descriptions not compelling enough for the SERP | Rewrite titles with numbers, power words, and clear value props; test meta descriptions with hooks | | FID replaced by INP — pages now failing CWV | INP measures all interactions, not just first — JS-heavy pages fail | Break long JS tasks, defer third-party scripts, audit event handlers; 43% of sites still fail INP in 2026 | | Indexed pages declining in Search Console | Google tightening quality bar — deindexing thin or duplicate content | Consolidate thin pages, add unique content, improve E-E-A-T signals on remaining pages | | AI Overviews stealing clicks from position 1 | Google AI Overviews now appear in 50%+ of queries, reducing organic CTR by ~42% | Optimize for AI citation (extractable content blocks), add FAQ schema, target queries less likely to trigger AI Overviews | | Keyword cannibalization across blog and product pages | Multiple pages competing for same keyword with conflicting intent | Map one primary keyword per page, consolidate or redirect competing pages, differentiate intent |
In scope:
Out of scope:
Known limitations:
bash# Analyze keyword list for search intent and difficulty python scripts/keyword_analyzer.py --keywords keywords.csv --json # Simulate SERP appearance for a page python scripts/serp_simulator.py --title "Cloud Cost Optimization Guide" --description "Learn 7 proven strategies..." --url "/guides/cloud-cost" --json # Score content for on-page SEO quality python scripts/content_scorer.py article.md --keyword "cloud cost optimization" --json
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-03 | pass→pass | 24,336 | 29,776 | +22% | 1 | 1 | 0% | 4,151 | 7,620 | +84% | 0 | 0 | — |
case-13 | fail→pass | 13,640 | 11,331 | -17% | 1 | 1 | 0% | 2,471 | 4,689 | +90% | 0 | 0 | — |
case-14 | fail→pass | 15,732 | 17,976 | +14% | 1 | 1 | 0% | 2,572 | 6,043 | +135% | 0 | 0 | — |
case-01 | fail→pass | 18,065 | 20,403 | +13% | 1 | 1 | 0% | 3,221 | 6,699 | +108% | 0 | 0 | — |
case-02 | pass→pass | 19,873 | 22,788 | +15% | 1 | 1 | 0% | 3,095 | 6,935 | +124% | 0 | 0 | — |
case-04 | pass→pass | 6,096 | 9,012 | +48% | 1 | 1 | 0% | 1,180 | 4,310 | +265% | 0 | 0 | — |
case-05 | pass→pass | 3,412 | 6,716 | +97% | 1 | 1 | 0% | 562 | 3,900 | +594% | 0 | 0 | — |
case-06 | pass→pass | 10,034 | 12,727 | +27% | 1 | 1 | 0% | 1,637 | 5,042 | +208% | 0 | 0 | — |
case-07 | pass→pass | 9,592 | 8,409 | -12% | 1 | 1 | 0% | 1,288 | 4,425 | +244% | 0 | 0 | — |
case-08 | pass→pass | 10,857 | 12,592 | +16% | 1 | 1 | 0% | 1,916 | 4,863 | +154% | 0 | 0 | — |
case-09 | pass→fail | 14,832 | 10,896 | -27% | 1 | 1 | 0% | 2,270 | 4,419 | +95% | 0 | 0 | — |
case-10 | fail→pass | 5,706 | 6,644 | +16% | 1 | 1 | 0% | 933 | 3,940 | +322% | 0 | 0 | — |
case-11 | pass→pass | 12,667 | 14,617 | +15% | 1 | 1 | 0% | 2,028 | 5,275 | +160% | 0 | 0 | — |
case-12 | pass→pass | 7,296 | 8,901 | +22% | 1 | 1 | 0% | 1,553 | 4,358 | +181% | 0 | 0 | — |
case-15 | pass→pass | 36,301 | 21,012 | -42% | 1 | 1 | 0% | 2,660 | 5,836 | +119% | 0 | 0 | — |
case-16 | pass→pass | 16,822 | 20,326 | +21% | 1 | 1 | 0% | 2,446 | 5,843 | +139% | 0 | 0 | — |
case-17 | pass→fail | 13,340 | 16,118 | +21% | 1 | 1 | 0% | 2,188 | 5,275 | +141% | 0 | 0 | — |
case-18 | pass→pass | 10,291 | 14,147 | +37% | 1 | 1 | 0% | 1,642 | 4,918 | +200% | 0 | 0 | — |
case-19 | pass→pass | 4,144 | 5,198 | +25% | 1 | 1 | 0% | 660 | 3,697 | +460% | 0 | 0 | — |
case-20 | fail→pass | 15,658 | 6,678 | -57% | 1 | 1 | 0% | 2,622 | 3,761 | +43% | 0 | 0 | — |
case-21 | fail→fail | 38,178 | 7,083 | -81% | 1 | 1 | 0% | 6,172 | 4,075 | -34% | 0 | 0 | — |
case-22 | fail→fail | 18,699 | 23,101 | +24% | 1 | 1 | 0% | 2,802 | 6,272 | +124% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +14 percentage points is the difference between those two pass rates over the 22 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.