AI Content Studio - Claude Skills Collection

AI-powered Milady NFT meme generator + Twitter automation + Social monitoring

A comprehensive collection of Claude Skills for creating Milady memes, managing Twitter content with personality, and automating social media workflows.

🎯 What is This?

This repository contains 6 independent Claude Skills that work together to power an intelligent social media bot. Each Skill can be used standalone or combined for maximum impact.

Featured Skills

Skill	Purpose	Cost	Standalone?
Milady Meme Generator	Generate Milady NFT memes with 324+ accessories	FREE	✅
AI Image Effects	Apply Illusion/FLUX/SAM AI effects	$0.006-0.05	✅
Twitter Content AI	Generate tweets with AI persona	API cost	✅
Lark Bot Integration	Deploy as Lark (Feishu) bot	FREE	✅
Social Monitoring	Monitor 151 Twitter accounts	FREE	✅
Data Training Manager	Manage AI training data quality	FREE	✅

🚀 Quick Start

📥 Step 0: Download NFT Assets (Required for Milady Meme Generator)

⚠️ Important: The Milady NFT images (14GB) are not included in this repository due to size limitations. You need to download them separately:

# Download all 10,000 Milady NFT images (~14GB)
python scripts/download_milady_nfts.py

# Or download just a test set (first 10 NFTs)
# Select option 2 when prompted

# The script will download images to: assets/milady_nfts/images/

Note: The download script will:

• Download images from the official Milady Maker website
• Save metadata for each NFT
• Create an index file for quick lookup
• Resume automatically if interrupted

Time estimate: ~2-4 hours for all 10,000 NFTs (depends on network speed)

Alternative: If you only need the meme generator layers (not the base NFTs), you can skip this step. The layers are already included in the repository.

---

⚠️ IMPORTANT: Configure API Keys First!

Before using any Skills, you MUST configure your API keys:

# 1. Run the configuration wizard (recommended)
python scripts/setup_config.py

# OR manually copy and edit the config file
cp config/.env.example config/.env
nano config/.env  # Fill in your API keys

# 2. Test your configuration
python scripts/test_config.py

📚 Need help getting API keys? See CONFIG.md for detailed instructions.

💡 What you need:

• Twitter API (FREE) - For social monitoring
• Claude API ($0.01-0.05/tweet) - For content generation
• Replicate API ($0.006-0.05/image) - For AI image effects
• Lark Bot (FREE) - For team collaboration

Don't have all keys? No problem! Each API is optional - configure only what you need.

---

Option 1: Use Individual Skills

Each Skill is self-contained and can be used independently:

# Example: Use just the Milady Meme Generator (NO API KEY NEEDED)
cd skills/milady-meme-generator
python src/meme_generator_v2.py

# Example: Use AI Image Effects (REQUIRES: Replicate API)
cd skills/ai-image-effects
python src/flux_fill_pro.py

Option 2: Use as Claude Skills

Install Skills to your Claude Code environment:

# 1. First, configure your API keys (see above)
python scripts/setup_config.py

# 2. Copy Skills to Claude Code
cp -r skills/* ~/.claude/skills/

# Or link them (recommended for development)
ln -s $(pwd)/skills/* ~/.claude/skills/

Now Claude will automatically use these Skills when relevant!

Option 3: Deploy Full Bot

Run the complete AI Content Studio system:

# 1. Install dependencies
pip install -r requirements.txt

# 2. Configure environment
cp .env.example .env
# Edit .env with your API keys

# 3. Start webhook server
python webhook_server.py

# Bot is now running at http://localhost:8000

📦 Skills Overview

1. Milady Meme Generator

Generate Milady NFT memes with layered composition

• 10,000 Milady NFT base images (download separately, see Step 0 above)
• 324 layered accessories (hats, glasses, earrings, etc.) - included in repo
• Text overlays with 4 font styles
• Natural language layer selection
• Template-based generation

📥 First-time setup: Run python scripts/download_milady_nfts.py to download NFT images (~14GB)

from skills.milady_meme_generator.src.meme_generator_v2 import MemeGeneratorV2

generator = MemeGeneratorV2()
meme = generator.generate_meme(
    nft_id=5050,
    top_text="GM",
    bottom_text="WAGMI",
    layers=["Hat:Beret.png", "Glasses:Sunglasses.png"]
)

Cost: FREE (local processing)

→ Full Documentation

---

2. AI Image Effects

Apply professional AI effects to images

• Illusion Diffusion: Optical illusions and patterns ($0.006)
• FLUX Fill Pro: Smart inpainting and accessory replacement ($0.05)
• SAM Detector: Meta SAM-2 object detection (<$0.01)

from skills.ai_image_effects.src.flux_fill_pro import FluxFillPro

flux = FluxFillPro()
result = flux.replace_accessory(
    "milady.png",
    accessory_type="hat",
    new_description="red baseball cap"
)

Cost: $0.006-0.05 per image

→ Full Documentation

---

3. Twitter Content AI

Generate authentic Twitter content with personality

• Jessie persona (Codatta data cleaner intern)
• 4 content types (GM, insights, casual, interactions)
• 180+ training samples
• Freshness monitoring (prevents repetition)
• Multi-style generation (short/medium/long)

from skills.twitter_content_ai.src.content_generator import ContentGenerator

generator = ContentGenerator()
tweet = generator.generate_gm_post()
# "gm to data contributors who deserve ownership 🎀🧹"

Cost: Claude API usage

→ Full Documentation

---

4. Lark Bot Integration

Deploy as interactive Lark (Feishu) bot

• 10+ meme generation commands
• Real-time webhook processing
• Image upload to Lark chats
• Team collaboration features
• Approval workflow for tweets

# Start Lark bot
python webhook_server.py

# In Lark chat:
/milady 5050 Hat:Cowboy.png top:GM

Cost: FREE (Lark API is free)

→ Full Documentation

---

5. Social Monitoring

Monitor Twitter activity and find opportunities

• Track 151 key accounts (Codatta, Base, x402, AI/Data, Crypto, Milady)
• Real-time mention detection
• Intelligent priority scoring
• Engagement metrics tracking
• Auto-interaction opportunities

from skills.social_monitoring.src.twitter_monitor import TwitterMonitor

monitor = TwitterMonitor()
mentions = monitor.check_mentions()
opportunities = monitor.find_opportunities(min_priority=70)

Cost: FREE (Twitter API basic tier)

→ Full Documentation

---

6. Data Training Manager

Maintain AI training data quality

• Freshness monitoring (prevents staleness)
• Quality scoring and tracking
• Continuous learning system
• Auto-archive low performers
• Trend analysis

from skills.data_training_manager.src.freshness_monitor import FreshnessMonitor

monitor = FreshnessMonitor()
score = monitor.check_freshness(
    "your generated text",
    threshold=0.7
)

Cost: FREE (local processing)

→ Full Documentation

---

🏗️ Architecture

┌─────────────────────────────────────────────────────┐
│  User Input (Lark Bot / CLI / Claude Code)          │
└────────────────┬────────────────────────────────────┘
                 │
        ┌────────▼────────┐
        │ Skill Router    │
        │ (Auto-selected) │
        └────────┬────────┘
                 │
    ┌────────────┼────────────────────┐
    │            │                    │
    ▼            ▼                    ▼
┌─────────┐ ┌──────────┐ ┌────────────────┐
│ Milady  │ │ Twitter  │ │ Social         │
│ Meme    │ │ Content  │ │ Monitoring     │
│ Gen     │ │ AI       │ │                │
└────┬────┘ └────┬─────┘ └────────────────┘
     │           │
     ▼           ▼
┌─────────┐ ┌──────────┐
│ AI      │ │ Training │
│ Effects │ │ Manager  │
└─────────┘ └──────────┘
     │
     ▼
┌──────────────┐
│ Lark Bot     │
│ (Optional)   │
└──────────────┘

📋 Requirements

Core Dependencies

pip install pillow requests anthropic replicate flask

Optional Dependencies

# For Lark bot
pip install -r requirements_lark.txt

# For development
pip install pytest black flake8

API Keys

• Replicate API (for AI effects): https://replicate.com
• Claude API (for content generation): https://anthropic.com
• Twitter API (for monitoring): https://developer.twitter.com
• Lark API (for bot deployment): https://open.larksuite.com

🎨 Examples

Example 1: Generate Random Milady Meme

from skills.milady_meme_generator.src.meme_generator_v2 import MemeGeneratorV2

gen = MemeGeneratorV2()
meme = gen.generate_random_meme()
meme.save("random_meme.png")

Example 2: Smart Accessory Replacement

from skills.ai_image_effects.src.sam_detector import SAMDetector
from skills.ai_image_effects.src.flux_fill_pro import FluxFillPro

# Detect hat automatically
sam = SAMDetector()
bbox = sam.detect_accessory("milady.png", "hat")

# Replace with AI
flux = FluxFillPro()
result = flux.replace_region("milady.png", bbox, "cowboy hat")

Example 3: Generate Daily Twitter Content

from skills.twitter_content_ai.src.content_generator import ContentGenerator

gen = ContentGenerator()

# Morning GM post
gm = gen.generate_gm_post()

# Industry insight
insight = gen.generate_insight(topic="data_ownership")

# Casual content
casual = gen.generate_casual()

Example 4: Monitor & Auto-Reply

from skills.social_monitoring.src.twitter_monitor import TwitterMonitor
from skills.twitter_content_ai.src.content_generator import ContentGenerator

monitor = TwitterMonitor()
content = ContentGenerator()

# Check mentions
mentions = monitor.check_mentions()

for mention in mentions:
    # Generate contextual reply
    reply = content.generate_reply(
        original_tweet=mention['text'],
        author=mention['author']
    )
    print(f"Reply to {mention['author']}: {reply}")

💰 Cost Breakdown

Operation	Cost	Notes
Milady meme generation	FREE	Local processing
Text overlay	FREE	Local processing
Illusion Diffusion	$0.006	Per image
FLUX Fill Pro	$0.05	Per image
SAM detection	<$0.01	Per detection (cached)
Claude content generation	~$0.01-0.05	Per tweet
Twitter API	FREE	Basic tier (500k/month)
Lark API	FREE	No limits

Monthly estimate (moderate use):

• 100 AI memes: ~$5
• 200 tweets generated: ~$10
• Social monitoring: FREE
• Total: ~$15/month

📖 Documentation

Getting Started

• Installation Guide
• Configuration Guide
• Deployment Guide

Skill Guides

• Milady Meme Generator
• AI Image Effects
• Twitter Content AI
• Lark Bot Integration
• Social Monitoring
• Data Training Manager

Advanced

• Architecture Overview
• API Reference

🤝 Contributing

Contributions welcome! Each Skill is independent, so you can:

1. Improve existing Skills - Add features, fix bugs
2. Create new Skills - Follow Claude Skills standards
3. Add training data - Contribute successful content samples
4. Improve documentation - Help others use these Skills

📄 License

MIT License

🙏 Acknowledgments

• Milady NFT - Original artwork and community
• Codatta - Data ownership vision
• Replicate - AI model hosting
• Anthropic - Claude API
• Meta - SAM-2 model
• Black Forest Labs - FLUX models

🔗 Links

• Claude Skills Documentation
• Milady NFT
• Codatta

---

Made with 🎀 for the Milady community