How an SEO Article Skill Automates Keyword Placement and Search‑Engine‑Friendly Structure

Problem

Content operators must satisfy many fragmented SEO rules (search intent, title/H‑tag hierarchy, keyword density, LSI coverage, internal linking, meta description). Manual juggling leads to low rankings; example: a "CRM integration" article ranked 37th with 2.1% keyword density and only three LSI terms.

Skill Architecture

The SEO Article Skill encodes the rules into a seven‑module structure and a five‑step execution flow.

Step 1 – Search Intent Recognition & Structure Matching : select a template (information, transactional, commercial‑investigation) based on the search_intent enum.

Step 2 – Keyword Matrix Construction : generate 5‑8 primary‑keyword variants and 8‑15 LSI terms; plan placement (title, first 100 words, H2, H3, body density 0.5‑1.0%).

Step 3 – Content Generation & Keyword Embedding : produce each paragraph according to the article_structure brief, insert primary variants first and 1‑2 LSI per paragraph.

Step 4 – Internal Link & FAQ Optimization : suggest 2‑3 internal links with keyword‑rich anchor text; generate FAQ block with long‑tail questions.

Step 5 – SEO Self‑Check : verify density, H‑tag hierarchy, title ≤60 chars, meta ≤155 chars, LSI coverage ≥80%, anchor‑text rules.

Module Details

Meta Information defines name seo-article-writer, description, priority 8, version 1.3.0.

Trigger Conditions detect requests containing keywords such as "SEO文章", "关键词布局" and exclude intents like "邮件营销".

Input/Output Specs required fields: target_keyword, search_intent, language; optional arrays for secondary keywords, competitor URLs, internal links. Outputs: seo_meta, article_structure, lsi_coverage, internal_link_suggestions, full_article.

Constraints

hard_rules:
  - Title Tag ≤60 characters, primary keyword in first 5 words
  - Meta Description ≤155 characters, includes primary keyword + 1 LSI + CTA
  - First 100 words must contain the primary keyword
  - Every H2 must contain the primary keyword or a variant
  - Keyword density 0.5%-1.0% (words/200‑300)
  - No consecutive sentences may repeat the primary keyword
  - LSI coverage ≥80%
  - H1→H2→H3 hierarchy must not skip levels
  - Internal‑link anchor text must not use generic phrases like "click here"
soft_rules:
  - Prefer keywords in lists and tables
  - FAQ questions should contain long‑tail keywords
  - Place key information at the beginning of paragraphs

Evolution

Versions 1.0‑1.1 used a density‑driven rule (1.5‑2.5%) and only an information template. Analysis of top‑10 Google results showed average density 0.8‑1.2% and extensive LSI coverage, prompting version 1.2 to adopt a semantic‑coverage rule (density 0.5‑1.0%, LSI ≥ 80%). Version 1.3 added automatic self‑checks and refined internal‑link anchor generation.

Concrete Example – "CRM integration" (Information type)

Input

target_keyword: "CRM integration"

search_intent: 信息型

language: English

word_count: 2000

secondary_keywords: ["CRM API", "data synchronization", "Salesforce"]

Skill Output Highlights

Title Tag: CRM Integration: Step‑by‑Step Guide (2025)

Meta Description: "Learn how to integrate CRM systems… Covers API authentication, data sync, and troubleshooting."

H2 example: "How CRM Integration Works: API & Data Sync"

H3 example: "1. API Authentication Setup"

Keyword density: 0.8% (16 occurrences in 2000 words)

LSI coverage: 14 terms, 93% coverage

Internal‑link anchor: "explore our CRM features" (no generic "click here")

FAQ: 3 long‑tail questions

Compared with a manual article (title "如何做CRM集成", density 2.1%, LSI coverage 21%), the Skill‑generated article achieved higher compliance and readability.

Efficiency & Effectiveness Comparison

Total time: manual 4‑6 hours vs. Skill 30‑40 seconds

Title/Meta compliance: automatic ≤60/155 characters vs. frequent manual overruns

H2 keyword coverage: 47% manual vs. 89% Skill

LSI coverage: 21% manual vs. 93% Skill

Search‑intent match error rate: 35% manual vs. <10% Skill

Keyword‑stuffing risk: medium‑high manual vs. low Skill

One month after deployment, Search Console showed the manual article averaging position 28, while the Skill‑generated article averaged position 11.

Pitfalls & Mitigations

Density path dependency : higher density no longer correlates with ranking; solution – enforce low density (0.5‑1.0%) and high LSI coverage.

Static search‑intent assumption : intent can shift (e.g., "CRM integration" moving from informational to commercial‑investigation); solution – regularly audit top‑10 results and update the default search_intent.

Missing structured‑data markup : Schema.org FAQ/How‑To can boost rich‑snippet CTR by 30‑50%; solution – add a schema_markup field in a future version.

Future Extensions

The Skill can be chained with a cross‑platform content‑adaptation Skill to convert the SEO‑optimized English article into Chinese WeChat posts or Xiaohongshu notes, swapping the structural layer (H‑tags & LSI) for platform‑specific formatting. It can also ingest secondary‑keyword outputs from an industry‑research Skill, creating a closed loop from research to search traffic.

Version History

version: 1.3.0
changelog:
  - 1.0.0: initial version, information template, keyword placement during content generation
  - 1.1.0: added transactional and commercial‑investigation templates; keyword placement extracted to Step 2
  - 1.2.0: density lowered to 0.5%; LSI ≥80% hard rule; FAQ generation added
  - 1.3.0: internal‑link anchor optimization; H‑tag validation; SEO self‑check step

Execution Walkthrough (Input → Output)

Using the same "CRM integration" parameters, the Skill produces the structured outputs shown above, achieving 0.8% density, 93% LSI coverage, compliant title/meta, and keyword‑rich H‑tags within seconds.