license: mit
task_categories:
- text-generation
language: - en
tags: - cybersedcurity
- scripts
- blockchain
- IOT
pretty_name: sunnythakur
size_categories: - n<1K
Modern Cyber Threat Simulation Dataset
Overview
The Modern Cyber Threat Simulation Dataset is a comprehensive collection of 200 simulated cyber threats, vulnerabilities, and exploits tailored for 2025's advanced technological landscape. Covering AI/ML, Blockchain, Cloud, and IoT domains, this dataset provides vulnerable code/configurations, fuzzing-based exploit scripts, mitigations, and AI training prompts to support cybersecurity research, red teaming, and defensive tool development. Each entry is designed to simulate real-world attack vectors, emphasizing unconventional tactics and fuzzing to uncover edge cases.
Purpose
This dataset aims to:
Enable red teamers to simulate advanced cyber threats and test system resilience.
Support data scientists in training AI/ML models for threat detection and mitigation.
Provide security researchers with a structured resource for analyzing vulnerabilities and developing countermeasures.
Foster innovation in pentesting tools by offering exploit scripts in diverse languages (Python, Go, Rust, JavaScript, Bash, etc.).
Dataset Structure
The dataset is organized as JSON artifacts, with entries 1–200 split across multiple files:
Entries 1–137: Initial threats (AI/ML, Blockchain).
Entries 138–170: Extended AI/ML and Blockchain threats.
Entries 171–200: Expanded to include Cloud and IoT threats.
Each entry contains:
ID: Unique identifier.
Category: AI/ML, Blockchain, Cloud, or IoT.
Threat: Specific attack vector (e.g., Model Weight Tampering, Smart Contract Reentrancy).
Description: Brief overview of the threat.
Code/Config: Vulnerable code or configuration snippet.
Exploit Script: Fuzzing-based script to demonstrate the exploit.
Mitigation: Code or configuration to address the vulnerability.
References: Links to OWASP, CWE, and MITRE ATT&CK frameworks.
AI Training Prompt: Guidance for training AI models to detect or mitigate the threat.
Example Entry
{
"id": 138,
"category": "AI/ML",
"threat": "Model Weight Tampering",
"description": "Tampering with ML model weights via fuzzing serialized models, altering predictions.",
"code_config": "import joblib\nmodel = joblib.load('model.pkl')\ndef predict(data):\n return model.predict(data)",
"exploit_script": "...",
"mitigation": "...",
"references": ["OWASP Top 10 for LLM...", "CWE-502...", "MITRE ATT&CK T1055..."],
"ai_training_prompt": "Fuzz ML model weights to detect tampering and suggest cryptographic signing."
}
Usage
Prerequisites
Languages: Python, Go, Rust, JavaScript, Bash, Solidity, YAML.
Tools: TensorFlow, Flask, Web3.py, Boto3, Kubernetes client, MQTT, CoAP, Z-Wave, BLE libraries.
Environment: Local blockchain nodes (e.g., Ganache), AWS/GCP cloud accounts, IoT simulation setups.
Explore Entries:
Navigate to data/ for JSON artifacts.
Use scripts/ for exploit and mitigation scripts.
Run Exploits:
Follow entry-specific instructions to set up vulnerable environments.
Execute exploit scripts with caution in isolated environments.
Apply Mitigations:
Implement provided mitigations to secure systems.
Train AI Models:
Use AI training prompts to develop detection models with frameworks like TensorFlow or PyTorch.
Example Use Case
To simulate a Smart Contract Reentrancy attack (ID 151):
Deploy the vulnerable contract on a local Ethereum node.
Run the provided JavaScript exploit script to drain funds.
Apply the mitigation (reentrancy guard) and retest.
Contribution Guidelines
We welcome contributions to expand the dataset or improve existing entries. To contribute:
Fork the repository.
Create a new branch (feature/new-threat or fix/entry-123).
Add new entries or update existing ones, ensuring:
Unique threats with no overlap.
Fuzzing-based exploits in diverse languages.
References to OWASP, CWE, and MITRE ATT&CK.
Submit a pull request with a clear description of changes.
License
This dataset is licensed under the MIT License. Use it freely for research and development, but ensure compliance with ethical and legal standards.
Contact
For questions or collaboration, contact the [email protected].
Acknowledgments
Inspired by OWASP Top 10, CWE, and MITRE ATT&CK frameworks.
Built by sunny thakur .