
content-generator
by az9713
Claude Code plugin implementing 5 proven marketing psychology frameworks for promoting AI products
SKILL.md
name: content-generator description: Generate marketing content applying all 5 frameworks - for GitHub READMEs, landing pages, social media, and more allowed-tools:
- Read
- Write
- WebFetch
- Task invocation: user
Content Generator
Use this skill to generate optimized marketing content that applies all 5 marketing frameworks. Supports multiple content types.
Content Types Supported
- GitHub README - Developer-focused, technical credibility
- Landing Page - Conversion-focused, full persuasion stack
- X/Twitter Thread - Viral-focused, hook-driven
- LinkedIn Post - Professional, thought leadership
- YouTube Script Intro - Hook-focused, retention-optimized
- Substack/Blog Article - Long-form, value-driven
- Product Hunt Launch - Launch-optimized, community-focused
Generation Process
Step 1: Gather Context
Required inputs:
- Product name and description
- Target audience
- Key features/benefits
- Voice profile (if available)
- Competitors (for differentiation)
Step 2: Apply Framework Stack
For each piece of content, apply:
- Awareness Analyzer → Determine messaging level
- NESB Scorer → Ensure headlines hit all 4 quadrants
- Persuasion Auditor → Include key levers
- Copy Optimizer → Apply tactical techniques
- Tribe Builder → Add movement elements
Step 3: Adapt for Platform
Each platform has specific requirements:
| Platform | Tone | Length | Key Elements |
|---|---|---|---|
| GitHub | Technical, direct | Comprehensive | Code examples, badges, install |
| Landing | Persuasive | Scannable | Hero, benefits, social proof, CTA |
| Twitter/X | Punchy, hook-driven | 280 chars/tweet | Thread structure, engagement |
| Professional | 1300 chars ideal | Story, insight, CTA | |
| YouTube | Conversational | 30-60s intro | Hook, promise, curiosity gap |
| Blog | Educational | 1500-3000 words | Value first, soft sell |
| Product Hunt | Community-focused | Structured | Problem, solution, why now |
Step 4: Generate with Voice
Match the user's extracted voice profile.
Step 5: Score and Refine
Run generated content through scorers before presenting.
Output Format
## Generated Content
### Content Type: [Type]
### Target Audience: [Audience]
### Awareness Level: [Level]
---
### Generated Content
[The actual content, properly formatted for the platform]
---
### Framework Application
| Framework | How Applied |
|-----------|-------------|
| Awareness | [How awareness level was matched] |
| NESB | [NEW/EASY/SAFE/BIG elements] |
| Persuasion | [Levers used] |
| Tactics | [Specific techniques] |
| Tribe | [Movement elements] |
### NESB Score
- NEW: X/10
- EASY: X/10
- SAFE: X/10
- BIG: X/10
- **Total: XX/40**
### Suggested Variations
1. [Alternative version 1]
2. [Alternative version 2]
Platform-Specific Templates
See supporting files:
github-readme.md- GitHub README templatelanding-page.md- Landing page sectionssocial-posts.md- Twitter and LinkedIn templatesblog-article.md- Blog post structure
Quick Generation Prompts
GitHub README
"Generate a README for [product] targeting [audience]. Key differentiator: [unique mechanism]. Include: badges, quick start, features, why us."
Landing Page Hero
"Generate hero section for [product]. Promise: [outcome]. Proof: [social proof]. Make it hit NEW, EASY, SAFE, BIG."
Twitter Thread
"Generate a 5-tweet thread about [topic/product]. Hook with [problem], reveal [solution], end with [CTA]."
LinkedIn Post
"Generate a LinkedIn post sharing [insight/story] that positions [product] as the solution. Professional tone, personal story."
Interactive Workflow
When generating content interactively:
- Propose → Generate initial draft
- Score → Show framework scores
- Refine → Iterate based on feedback
- Finalize → Present polished version
- Variations → Offer alternative angles
Voice Matching
Before generating, check if a voice profile exists:
- If yes: Match the extracted voice patterns
- If no: Use platform-appropriate defaults or offer to extract voice first
Examples
Example: GitHub README Hero
Input: AI code review tool for Python developers
Generated:
# CodeReviewAI
**Stop shipping bugs. Start shipping confidence.**
The first AI that actually understands your Python codebase - not just syntax, but intent.
[](link) [](link)
## Why CodeReviewAI?
Traditional linters catch typos. We catch logic errors, security holes, and "that thing that'll break in production at 3am."
- **Context-aware** - Reads your entire codebase, not just the diff
- **Zero config** - Works in 30 seconds, not 30 hours
- **Privacy-first** - Your code never leaves your infrastructure
> "Caught a critical bug in our payment flow that 3 human reviewers missed." - CTO, YC Startup
NESB Score: NEW 8, EASY 9, SAFE 8, BIG 8 = 33/40
Example: Twitter Thread Hook
Input: Launch announcement for productivity app
Generated:
Tweet 1:
I spent 6 months building in silence.
Today I'm launching [Product].
It does one thing: makes your to-do list actually get done.
Here's why existing tools fail you 🧵
Tweet 2:
Every productivity app assumes you have infinite willpower.
"Just add tasks and check them off!"
But willpower is finite. By 2pm, you're drained.
That's why 89% of tasks never get completed.
Tweet 3:
[Product] works differently.
Instead of fighting your brain, it works WITH it...
Score
Total Score
Based on repository quality metrics
SKILL.mdファイルが含まれている
ライセンスが設定されている
100文字以上の説明がある
GitHub Stars 100以上
3ヶ月以内に更新がある
10回以上フォークされている
オープンIssueが50未満
プログラミング言語が設定されている
1つ以上のタグが設定されている
Reviews
Reviews coming soon