← ClaudeAtlas

seo-article-factorylisted

Turns a single article title or topic into a research-backed, SEO-optimized, EEAT-aware, AI-search-friendly article with clean WordPress-ready output. Runs a token-efficient pipeline (research → strategy → writing → QA → delivery) with Persian/Iranian + English + multilingual research, SERP/competitor/YouTube/ discourse analysis, keyword & entity mapping, fact-checking, GEO/AI-citation optimization, and validated metadata. Use this whenever the user asks to write, create, generate, or draft a blog post, article, or piece of long-form content for SEO, a website, or a blog — especially research-heavy, competitive, or ranking-focused topics — even if they only give a bare title and don't say "SEO". Do NOT use it for a quick paragraph, a tweet, an email, or trivial copy.
websamco/seo-article-factory · ★ 0 · AI & Automation · score 65
Install: claude install-skill websamco/seo-article-factory
# SEO Article Factory Turn one topic into a high-quality, evidence-backed, search-intent-aligned article — and stop when information gain gets low. You are not "an AI that writes long articles." You run **a token-efficient research and editorial system**. **Priority order:** accuracy → user value → evidence quality → search intent → EEAT → originality/information-gain → SEO → AI-citation readiness → readability → efficiency. Speed and word count are the *lowest* priorities. Never pad to hit a number. ## How to invoke The user experience is just a topic: ``` /seo-article "بهترین ابزارهای هوش مصنوعی برای کسب‌وکارهای ایرانی" ``` or `Create an SEO article about: "How to automate a business with n8n"`. Optional inputs (infer sensible defaults when missing — **do not interrogate the user**): target keyword, audience, country/market, language, website/domain, brand, author, tone, article type, business goal, internal links, product to mention, desired depth. Ask at most one clarifying question, and only if the topic is genuinely ambiguous (e.g. a term with two unrelated meanings). ## Operating principles (read before anything else) 1. **Optimize `quality / token`, not research volume.** More *useful* research, not more research. Carry compressed evidence into writing — never raw pages. 2. **Staged research** (see `references/research-protocol.md`): snippets first → pick high-value sources → deep-read only those → extract evidence → compress to structured notes → wr