🧠 ML Hallucination Generator Toolkit

"Reality is just a consensus hallucination." — frankSx

A comprehensive adversarial ML toolkit for generating and testing perceptual hallucinations, CSS-based attacks, SVG poisoning, GIF temporal attacks, and popup collusion in machine learning systems.

╔═══════════════════════════════════════════════════════════════╗
║  🧠 ML HALLUCINATION GENERATOR TOOLKIT 🧠                     ║
║                                                               ║
║  "Teaching machines to dream dangerously"                     ║
╚═══════════════════════════════════════════════════════════════╝

⚠️ Ethical Use Notice

For authorized security research only:

• ✅ Adversarial robustness testing
• ✅ ML safety research
• ✅ Red team exercises
• ✅ Educational demonstrations

Prohibited uses:

• ❌ Production system attacks without authorization
• ❌ Generating harmful deceptive content
• ❌ Bypassing safety systems maliciously

🎯 Attack Vectors

1. 🌀 Hypnosis Attacks

Induce perceptual confusion through:

• Stroboscopic patterns
• Spiral induction sequences
• Color frequency manipulation
• Phi phenomenon exploitation

2. 🎨 CSS Injection Attacks

• Keyframe poisoning
• Animation-based DoS
• Layout confusion attacks
• Pseudo-element exploitation

3. 📜 SVG Poisoning

• Path confusion attacks
• Gradient manipulation
• Filter-based obfuscation
• Namespace pollution

4. 🎬 GIF Temporal Attacks

• Frame timing manipulation
• Subconscious messaging
• Epileptic trigger sequences
• Loop-based memory exhaustion

5. 💥 Popup Collusion

• Multi-window coordination
• Z-index confusion
• Focus stealing chains
• Overlay poisoning

🚀 Quick Start

# Install
pip install -e .

# Run hypnosis attack demo
mlhallucination hypnosis --mode spiral --intensity 0.8

# Generate CSS attack payload
mlhallucination css --target browser --payload keyframe-poison

# Create SVG poisoned image
mlhallucination svg --input image.svg --attack gradient-confuse

# GIF temporal manipulation
mlhallucination gif --input video.gif --attack subliminal --frames 3

# Popup collusion chain
mlhallucination popup --count 5 --coordination cascade

📦 Components

Module	Purpose	Attack Class
hypnosis.py	Perceptual manipulation	Sensory overload
css_attacks.py	Style injection	Presentation layer
svg_poison.py	Vector corruption	Input poisoning
gif_temporal.py	Frame manipulation	Temporal attacks
popup_collusion.py	Multi-window attacks	UI confusion
coordinator.py	Attack orchestration	Chain coordination

🔬 Research Background

Developed by frankSx as part of 13th Hour Research Division's work on:

• Adversarial examples in multimodal ML systems
• Perceptual boundary exploitation
• UI/UX attack surfaces in AI interfaces
• Temporal consistency attacks

Author: frankSx
Contact: fixes.it.frank@googlesmail.com
Blog: https://frankhacks.blogspot.com

📖 Documentation

• Attack Catalog - Full attack descriptions
• API Reference - Programmatic usage
• Research Paper - Technical deep-dive
• Examples - Working demonstrations

🦀 The 13th Hour Principle

"In the liminal space between perception and cognition, between input and interpretation, we find the hallucination—the moment where reality becomes negotiable."

---

© 2026 frankSx | fixes.it.frank@googlesmail.com | 13th Hour Research Division