Content
# reverse-seo-traffic
> Claude Code skill — Reverse-engineer competitor SEO and grow organic traffic on your own sites, product pages, and blogs.
Based on the "Semrush MCP + Claude Code" technique demonstrated by SEO creators using natural-language competitive intelligence plus AI-driven page generation. This skill packages the full workflow as a reusable, phase-gated process: competitor discovery → organic mining → opportunity scoring → semantic cluster extraction → page architecture → multi-stack build → indexing → tracking.
---
## What this skill does
**Objective:** Drive organic traffic to **your own** web assets — product pages, blog posts, marketing sites, free-tool landing pages — using competitive reverse-engineering plus AI-assisted content generation.
**Not for:** client work, agency deliverables, third-party domains.
**Stack-aware build templates included:**
- Next.js (App Router)
- Astro (Content Collections + i18n)
- WordPress (ACF / Yoast / RankMath)
- HTML estático (pure HTML + serverless functions)
**Bilingual SEO support:** PT-BR + EN with proper `hreflang` configuration.
**Primary data source:** Semrush MCP connector for Claude Code. Manual fallback via `WebFetch` + SERP analysis if Semrush MCP is unavailable.
---
## The method (8 phases)
| Phase | What happens | Output |
|-------|-------------|--------|
| **0 — Preflight** | Verify Semrush MCP connected, target domain identified, stack chosen, language defined, goal stated | Go/no-go |
| **1 — Competitor discovery** | Natural-language prompt finds 3-5 real competitors ranking in your niche. Uses Semrush + WebFetch on G2/Capterra/Reddit/SERP | Ranked competitor list |
| **2 — Organic mining** | Semrush organic research per competitor. Pull top 20 pages by traffic. Flag free-tool / calculator / generator pages | Raw opportunity dataset |
| **3 — Opportunity scoring** | Score each candidate: 30% traffic + 20% KD inverse + 25% relevance + 15% build effort + 10% lead-gen potential. Pick top 3 | Top 3 ranked picks |
| **4 — Semantic cluster** | Deep dive chosen keyword. Extract all semantically related variants, PAA questions, top 5 SERP H1/H2 structures, common page sections | Hub + spokes cluster map |
| **5 — Page architecture** | Design page bundling all cluster keywords: H1 → spokes → free tool → testimonials → FAQ → schema → internal links → CTA | Page blueprint |
| **6 — Build** | Generate code in your stack from `references/` templates. Free tools use Groq Llama or Gemini Flash for cheap inference behind email gate | Production-ready code |
| **7 — Ship + index** | Deploy, verify schemas, submit to GSC + Bing, update sitemap, add internal links from existing high-authority pages | Live indexable page |
| **8 — Track + iterate** | Semrush position tracking, compare to SERP competitors, identify gaps, refresh quarterly | Ranking data |
---
## Why this works
**The mechanic:** competitors with high domain authority already proved which keywords convert. Their top-traffic pages are templatable patterns. You ship the same patterns in **your own** voice + product, with better semantic coverage and cheaper free-tool lead magnets. Google rewards comprehensive content matching searcher intent — semantic clustering signals exactly that.
**The leverage:** free-tool pages with email gates are the highest-ROI page archetype. Real example from the source video: a competitor's "free domain authority checker" was generating 4,000 monthly visits → all email-gated → lead-gen pipeline. Same pattern works in any niche: meta description generator, ROI calculator, template builder, compliance checker, simulator. Backed by cheap LLM inference (Groq, Gemini Flash) the marginal cost per lead is fractions of a cent.
**The differentiator vs generic SEO advice:** this skill assumes you're using Claude Code + Semrush MCP for natural-language competitive intelligence. Instead of manually exporting CSVs from a dashboard, you ask "find my competitors and show me what's ranking" and the model orchestrates the API calls. End-to-end workflow ships pages in hours, not weeks.
---
## Expected results
Built and indexed correctly, expect:
| Metric | Realistic range | Timeline |
|--------|-----------------|----------|
| Time to publish a cluster page | 4-8 hours per page | Day 1 |
| Google indexing after submission | 1-7 days | Week 1 |
| First impressions in Search Console | 50-300/day | Week 2-4 |
| First page-1 rankings (low KD long-tail) | 1-3 keywords | Week 3-6 |
| Free-tool email captures | 5-30/month per tool | Month 2+ |
| Organic visits per page (low competition niche) | 200-2,000/month | Month 3-6 |
| Compounding portfolio effect (10+ pages) | 2k-20k visits/month total | Month 6-12 |
**Caveat:** numbers depend on niche KD, domain authority, content quality, and backlink profile. Skill cannot guarantee rankings — it guarantees a repeatable workflow grounded in proven patterns.
---
## Installation
### Claude Code skill (recommended)
Place the skill in your Claude Code skills directory:
**macOS / Linux:**
```bash
git clone https://github.com/giovani-junior-dev/reverse-seo-traffic.git ~/.claude/skills/reverse-seo-traffic
```
**Windows (PowerShell):**
```powershell
git clone https://github.com/giovani-junior-dev/reverse-seo-traffic.git $env:USERPROFILE\.claude\skills\reverse-seo-traffic
```
Restart Claude Code. The skill auto-loads and triggers on relevant prompts.
### Prerequisites
1. **Claude Code** installed and authenticated.
2. **Semrush MCP connector** configured in Claude Code (recommended). Without it the skill runs in manual mode via `WebFetch` + `WebSearch` — slower but functional.
3. **One of:** Next.js, Astro, WordPress, or HTML static site to deploy pages to.
4. **Optional but high-value:** Groq API key (free tier) OR Gemini API key for powering free-tool pages.
---
## How to use
### Quick start — find opportunities
Open Claude Code in your project directory and say:
```
Find SEO opportunities for mydomain.com
```
The skill activates, runs Phases 1-3 (competitor discovery, organic mining, scoring), and returns the top 3 ranked opportunities with effort estimates. You pick one. Then say:
```
Build opportunity 1
```
The skill executes Phases 4-7 (cluster extraction, page build, ship + indexing instructions). Output is production-ready code in your stack.
### Targeted commands
```
Build a free-tool page for [TARGET_KEYWORD]
Build a cluster page for [TARGET_KEYWORD] in pt-br
Build a vertical guide for [NICHE] in en
Reverse-engineer [competitor.com]
Track ranking for [my-url] vs SERP top 5
```
### Multi-language setup
The skill handles PT-BR + EN clusters with proper `hreflang`. State both languages in your initial prompt:
```
Build cluster page for "[keyword]" in pt-br and en with hreflang
```
Templates in `references/hreflang-multilang.md` cover subdirectory routing, Search Console international targeting, and validation.
### Example real-world session
Tested on `wablastmessage.com` (a Brazilian WhatsApp bulk-messaging SaaS):
- **Phase 1** discovered 5 competitors: SocialHub, ClickMassa, Botconversa, Multichat, Comunicacaoemmassa.
- **Phase 2** mined SocialHub's blog, identified 3 dominant content patterns: vertical guides per industry, "X vs Y" comparatives, "2026 freshness" listicles.
- **Phase 3** scored top 3 picks. Winner: **vertical guides for industries SocialHub had not yet covered** (salões de beleza, e-commerce, advogados, academias, pet shops, escolas, oficinas, autônomos) — 8 templatable pages, low KD, high relevance, lead-gen via gated checklist.
- **Phase 4+** would extract semantic clusters per vertical and generate pages.
---
## Repository structure
```
reverse-seo-traffic/
├── SKILL.md # Skill manifest + 8-phase methodology
├── README.md # This file
└── references/
├── competitor-prompts.md # Semrush MCP prompt cookbook (8 prompts)
├── faq-schema.md # FAQPage / Article / SoftwareApplication JSON-LD
├── free-tool-pattern.md # Email-gated free-tool template + Groq/Gemini
├── hreflang-multilang.md # PT-BR + EN hreflang config
├── page-template-nextjs.md # Next.js App Router page
├── page-template-astro.md # Astro + Content Collections + i18n
├── page-template-wordpress.md # ACF + Yoast/RankMath template
└── page-template-html.md # Pure HTML static page + serverless functions
```
---
## Anti-patterns to avoid
This skill enforces several rules to prevent wasted effort and SEO penalties:
- Copy competitor copy verbatim — duplicate-content penalty. Reuse patterns, not prose.
- Skip schema markup — lose rich snippets and AI Overview eligibility.
- Free tool without email gate — zero lead-gen ROI. Always capture.
- Build without checking keyword difficulty — outranked by domain authority you cannot match.
- Orphan pages with no internal links — slow indexing, weak ranking signals.
- Trust AI output blindly — every generated page needs human review before publish.
- Multi-language sites without hreflang — SERP cannibalization between language variants.
- Work on client domains — this skill is scoped to your own assets only.
---
## Tech stack used by the skill
| Component | Role |
|-----------|------|
| Claude Code | Host for the skill, executes the workflow |
| Semrush MCP | Primary SEO data source (organic research, keyword data, position tracking) |
| `WebFetch` / `WebSearch` | Manual fallback for SERP analysis when Semrush MCP unavailable |
| Groq Llama 3.3 / Gemini 2.0 Flash | Cheap LLM inference behind free-tool pages |
| Resend / Supabase | Email capture and lead storage |
| Google Search Console | Indexing requests and performance tracking |
| Bing Webmaster Tools | Secondary indexing |
| schema.org validator | Pre-deploy schema validation |
---
## Contributing
Issues and pull requests welcome. Areas where contributions are especially valuable:
- Additional stack templates (Remix, SvelteKit, Hugo, Eleventy)
- Free-tool patterns for new niches (legal, healthcare, finance, e-commerce, real-estate)
- Localization for more languages (ES, FR, DE)
- Case studies with measurable before/after data
- Integration with additional SEO data sources (Ahrefs MCP, SE Ranking, Ubersuggest)
---
## License
MIT — use freely on your own projects. Attribution appreciated but not required.
---
## Credits
Method inspired by community SEO workflows demonstrating Semrush MCP + Claude Code integration for competitive reverse-engineering and AI-assisted page generation. Skill packaged and maintained by [@giovani-junior-dev](https://github.com/giovani-junior-dev).
---
## Quick reference card
```
Activate: "find SEO opportunities for [domain]"
Build: "build opportunity [N]" OR "build cluster page for [keyword]"
Track: "track ranking for [url]"
Stack: Next.js / Astro / WordPress / HTML static
Languages: PT-BR + EN with hreflang
Data source: Semrush MCP (or manual via WebFetch)
Time/page: 4-8 hours
First rank: 3-6 weeks (low KD)
```
