create-viral-content

Files (11)

• SKILL.md
• marketplace.json
• ai-tells.md · resources/
• humanize-integration.md · resources/
• platform-templates.md · resources/
• refinement-protocol.md · resources/
• research-statistics.md · resources/
• thumbnail-checklist.md · resources/
• title-formulas.md · resources/
• viral-thumbnails.md · resources/
• viral-titles.md · resources/

SKILL.md (first lines)

> ⚠️ **BEFORE USING THIS SKILL:** Review all files in the `resources/` directory. These contain AI tell catalogs, platform templates, refinement protocols, and 40-source research basis required for proper skill execution.

## Research Basis

This skill synthesizes findings from 40 documented research sources:
- **BuzzSumo:** 100M headlines study → optimal length is 11 words/65 characters
- **Outbrain:** Negative superlatives outperform positive by 63%
- **Netflix:** 82% of browsing time on thumbnails, 1.8s decision window
- **Face Psychology:** +35-50% CTR with faces in thumbnails
- **A/B Testing Research:** 30-40% CTR improvement over time

Full statistics in `resources/research-statistics.md`.

# Create Viral Content

Make your posts spread. This skill turns forgettable drafts into content that gets shares, comments, and action.

## Core Principle: The Deliberative Refinement Loop

Good content doesn't come from one pass. You attack it, fix it, attack again:

1. Generate initial draft
2. Attack it from audience perspectives
3. Identify AI tells and weak points
4. Refine with human voice
5. Repeat until unbreakable

## The Anatomy of Viral Content

### Hook Architecture (First 2 Seconds)

**Pattern: Prediction + Stakes**
```
"I think [CONCEPT] is the [YEAR] [CATEGORY] that [OUTCOME]."
```
Example: "I think deliberative refinement is the 2026 prompt technique that matters most."

**Why it works:**
- "I think" = personal conviction, not corporate announcement
- Year = creates FOMO and timeframe
- Category = helps reader self-identify
- Outcome = stakes that matter

**Pattern: Tribal Identity Split**
```
"[TECHNIQUE] separates [WINNERS] from [EVERYONE ELSE]."
```
Example: "This separates serious builders from prompt tourists."

**Why it works:**
- Creates in-group/out-group
- Reader immediately picks a side
- Ego investment drives engagement

**Pattern: Before/After Compression**
```
"What used to require [OLD COMPLEXITY] now [NEW SIMPLICITY]."
```
Example: "What used to need 12 models chained together now takes one."

**Why it works:**
- Concrete efficiency gain
- "I had no idea" response
- Shareable stat

### Body Structure: The Build

**Required elements (in order):**
1. WHAT - Explain the concept (1-2 sentences max)
2. HOW - The mechanic with concrete examples
3. WHY NOW - The breakthrough that makes it possible
4. PAYOFF - What you can actually build/achieve

**Anti-patterns to avoid:**
- Starting with "why it matters" before explaining "what it is"
- Generic benefits without specific mechanics
- Selling the sizzle before showing the steak

### Closer Architecture (Last 10%)

**Pattern: Command, Not Request**
```
BAD: "Try it. Change my mind."  (beggy, engagement bait)
GOOD: "Your next [ACTION] shouldn't [OLD WAY]. It should [NEW WAY]."
```
Example: "Your next prompt shouldn't ask for an answer. It should demand: 'Attack this from three expert perspectives, ground your claims, then revise.'"

**Why the command works:**
- Ends on authority, not weakness
- Gives immediate actionable next step
- Mirrors the thesis in its structure

## AI Tell Detection and Elimination

Kill these on sight:

### Transition Tells
- ❌ "Here's the wild part:"
- ❌ "Here's the thing:"
- ❌ "Let's dive in"
- ❌ "But here's the kicker:"
- ✅ Direct statement with no transition needed

### Enthusiasm Tells
- ❌ "I'm excited to share"
- ❌ "This is a game-changer"
- ❌ "Revolutionary"
- ✅ Let the content create excitement

### Structure Tells
- ❌ Numbered lists for everything
- ❌ "First... Second... Finally..."
- ❌ "In conclusion"
- ✅ Prose that flows naturally

### Engagement Bait Tells
- ❌ "Change my mind"
- ❌ "What do you think?"
- ❌ "Let me know in the comments"

From kennyolofsson23-netizen/claude-code-config, MIT.

More from kennyolofsson23-netizen/claude-code-config