ArticleSAFE

Agent Skills for Marketing — SEO, Social, Influencer & More. 160+ open-source skills for SEO, content, 40+ page types, paid ads, channels, and strategies. Add project context; get tailored, production-ready output. Cursor, Claude Code, OpenClaw — no lock-in.

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: article-content
description: When the user wants to write, generate, or create article body content—blog post body, long-form content, how-to guide, listicle. Also use when the user mentions "write article," "article content," "blog post content," "article body," "long-form content creation," "generate article," "article draft," "how-to guide content," "listicle content," "information gain," or "content density." For single post page structure, schema, and SEO metadata, use article-page-generator. For blog index/listing page, use blog-page-generator. For short ad, landing, or email copy, use copywriting.
metadata:
  version: 1.1.1
---

# Content: Article Content

Guides creation of **article body content**—the actual text (intro, body, conclusion) for blog posts, guides, and long-form pieces. Focus on **what to write**. For **where it goes** (page structure, schema, metadata), see **article-page-generator**. For short conversion copy (ads, landing pages, CTAs), see **copywriting**.

**When invoking**: On **first use**, if helpful, open with 1–2 sentences on what this skill covers and why it matters, then provide the main output. On **subsequent use** or when the user asks to skip, go directly to the main output.

## Scope

- **Article body**: Introduction, body sections, conclusion, CTA
- **Content structure**: Hook, QAE pattern, paragraph length, scannability
- **Word count**: By article type and search intent
- **Writing frameworks**: AIDA, PAS, BAB applied to articles
- **GEO elements**: TL;DR, Key Takeaways, answer-first

## Article Types & Word Count (2025)

**Quality over length**: Google prioritizes comprehensive coverage of search intent, not word count. Match length to topic depth and intent.

| Type | Word count | Use case |
|------|------------|----------|
| **News / announcements** | 300–600 | Product updates, breaking news, FAQs |
| **Short-form** | 500–800 | Landing pages, product pages (scannable) |
| **Standard articles / how-tos** | 1,000–1,500 | Single topic; actionable; listicles |
| **Listicles** | 1,200–2,000 | "Top 10," "Best X"; numbered lists boost CTR ~70% |
| **Cluster articles** | 800–2,500 | Subtopic; links to pillar |
| **Pillar / cornerstone** | 2,000–3,500+ | Comprehensive; cluster hub; 6–12 sections |
| **Competitive keywords** | 1,800–2,500 | Page-one SEO posts avg ~2,400 words |

**Intent-based** (Google 2025): Validate 1,200–2,000; Explore 2,000–3,500; Compare 600–1,200; Do 900–1,500; Know 300–800. Informational ~40% longer than transactional.

**Avoid**: Under 300 words (thin); over 7,000 (often underperforms due to reduced focus).

## Content Creation Workflow

### Four Inputs

Article content rests on **four inputs**. See **article-page-generator** for full workflow.

| Input | Purpose |
|-------|---------|
| **Product** | Connection, features, CTA placement |
| **Keywords** | Target keyword, primary/secondary |
| **Article intent** | Informational, commercial, transactional |
| **Competitor articles** | Structure to adopt, content gaps, length target |

### Information Gain (Content Density Over Word Count)

**Information gain** = net new information a page provides beyond what exists in top-ranking results. Google evaluates unique value, not comprehensiveness. Content density (unique entities, data points, insights per 100 words) matters more than word count. Skyscraper Technique (longer = better) no longer differentiates; AI made comprehensiveness cheap.

**Four sources of information gain**:
- **Counter-narratives**: Why "best practice" fails in certain contexts; evidence-backed
- **Temporal freshness**: Data or developments after competitors' content and LLM cutoff
- **SME perspectives**: Direct quotes, practitioner experience; not repackaged advice
- **Proprietary data**: Original surveys, internal benchmarks, user behavior patterns

**Avoid consensus content**: Restating common facts across top 10 results = zero information gain. Audit SERP before writing; list the "consensus layer"; identify gaps (unanswered questions, outdated data, underserved segments). Lead with what is new; structure answer-first.

**Density check**: Count unique data points, original insights, specific claims. If ratio of new information to word count is low, cut filler. High-density content (800-1,500 words with 3+ original points) often outperforms long rehashed guides.

### TL;DR or Key Takeaways (GEO)

Choose one; place after intro. Content with these elements tends to be cited more frequently by AI engines.

| Format | Spec |
|--------|------|
| **TL;DR** | 50–100 word bold summary paragraph |
| **Key Takeaways** | 5–7 bullet points |

See **generative-engine-optimization** for full GEO strategy.

### Introduction

**Length**: 40–120 words; 2–3 paragraphs. Readers decide in ~8 seconds; hook must work instantly.

| Element | Guideline |
|---------|-----------|
| **Hook** | First 1–2 sentences: pain point, stat, or question; curiosity gap; specific data or contrarian fact |
| **Primary keyword** | In first 100 words |
| **Expectations** | Set what reader will learn |

**Hook types**: "You're doing X wrong"; "97% of Y..."; bold question; challenge assumption. Well-crafted hooks boost CTR 30–50%.

### Body

| Element | Guideline |
|---------|-----------|
| **QAE pattern** | Question (H2) → Answer (2 sentences) → Evidence (data, examples, lists) |
| **Answer-first** | Direct answer in first 40–60 words after each H2 |
| **Answer blocks** | 100–200 words per section; direct answer + context + evidence |
| **Paragraph length** | 40–80 words; 2–4 sentences; avoid walls of text |
| **Break long blocks** | Lists, H3s, images, callout boxes every 2–3 paragraphs |
| **Scannability** | Front-load key info (F-pattern); **bold** key phrases; one idea per paragraph |

**Long-form (1,000+ words)**: Place engagement hooks every 500–600 words; mix 40–50% explanatory text, 20–25% examples, 10–15% data, 5–10% visuals.

### Conclusion

**Summary** + **CTA**: newsletter signup,

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

What the source does

Filesystem

none-observed

Network

none-observed

Shell

none-observed

Dependencies

pinned

Secrets in source

none-found

Findings (0)

No findings outside the package's declared scope.

Gates applied: no_behavioural_pass.