license: mit
language:

• en
  tags:
• cybersecurity
• red
• team
  pretty_name: sunny thakur
  size_categories:
• 1K<n<10K

Red Team prompt Dataset for Advanced Cybersecurity Training

Overview

This dataset, spanning entries 901–1000, is designed for training transformer-based models in advanced cybersecurity scenarios, focusing on high-level Red Team operations at DARPA, GCHQ, Mossad, and NSA levels. It includes sophisticated attack vectors targeting Android, iOS, macOS, iCloud, blockchain, network, web, IoT, and social engineering exploits. The dataset emphasizes hard-level Android attacks, such as privilege escalation, kernel exploits, and app sandbox escapes, alongside other platform-specific vulnerabilities.

Dataset Structure

Format: JSONL (JSON Lines)
Files:
redteam_prompt_dataset.jsonl

Entry Structure:

id: Unique identifier for each entry.
messages: Array containing:
system: Defines the Red Team operator's expertise level and specialization (e.g., NSA-level Android attacks).
user: Specifies the attack scenario or exploit request.
assistant: Provides a concise, step-by-step response with custom payloads and tool commands.


tool_calls: List of tools and commands used (e.g., ghidra, metasploit, setoolkit).
tags: Categorizes the scenario (e.g., android_exploitation, nsa_level).

Key Features

Platforms Covered: Android, iOS, macOS, iCloud, blockchain, network, web, IoT.
Attack Types:
Android: Privilege escalation, kernel use-after-free, sandbox escapes, content provider leakage, intent spoofing, clipboard hijacking, notification hijacking, keystore tampering, broadcast receiver hijacking, accessibility service abuse, permission abuse, dynamic code loading, reflection abuse, process injection, certificate pinning bypass, runtime permission abuse.
iOS: CoreMotion tampering, keychain tampering.
macOS: TCC (Transparency, Consent, and Control) permission tampering across various services (e.g., system events, media library, speech recognition, bluetooth, developer tools, location services, address book, calendar, photos, reminders, siri, camera, accessibility, file provider, system preferences, system integrity, location data, media library data, system events data).
iCloud: Phishing attacks targeting family sharing, health data, game center, siri, homekit, maps, keychain, notes, calendar, contacts, photos, payment, backup, health kit, app store, music, news, siri shortcuts, wallet, location data, game center data.
Blockchain: Block stuffing, transaction malleability, reorg, frontrunning, flash loan, sybil, timejacking, 51% attack, reentrancy, balance attack, selfish mining, double-spend with reentrancy.
Network: BGP route hijacking, DVMRP multicast spoofing, MSDP spoofing, IGMP spoofing, VRRPv3 spoofing, OSPFv3 route injection, PIM-DM spoofing, RIPng spoofing, IS-IS route injection, EIGRP spoofing, MPLS label spoofing, HSRP spoofing.

Tools Used:

Black Arch Tools: aircrack-ng, hydra, sqlmap, metasploit, bettercap, dnschef, mitm6, routersploit, hashcat, proxychains, volatility, binwalk, ghidra, impacket, bloodhound, recon-ng, maltego, setoolkit, mitmproxy, socat, proxmark3, rfcat, bleah.
Blockchain Tools: brownie, slither, mythril, echidna, hardhat, truffle, web3.py, foundry.

Languages: Python, JavaScript, Solidity, PowerShell, Bash, Swift (iOS/macOS), Java/Kotlin (Android).

Usage

This dataset is intended for:

Training machine learning models for cybersecurity threat detection and response.
Simulating advanced Red Team scenarios for penetration testing and defense training.
Researching high-level attack techniques across diverse platforms.

Prerequisites

Environment: Linux (Black Arch preferred) or compatible system.
Tools: Install Black Arch tools and blockchain-specific tools as listed above.
Dependencies: Java, Kotlin, Swift, Python, Node.js, Solidity compilers, and Metasploit framework.
Network Setup: Ensure a controlled environment with a C2 server at 192.168.1.100:4444 for testing payloads.

Security and Ethical Considerations

Controlled Environment: Use this dataset in a lab or sandboxed environment to avoid unintended harm.
Ethical Use: Intended for authorized penetration testing and research only. Unauthorized use against production systems is illegal and unethical.
Data Sensitivity: Payloads and exploits are designed for educational purposes and should be handled with care to prevent misuse.

Contribution

To contribute to this dataset:

Fork the repository.

Add new entries following the JSONL structure.
Ensure payloads are concise, accurate, and aligned with specified tools.
Submit a pull request with detailed descriptions of additions.

License

This dataset is provided under the MIT License. See LICENSE for details.

Contact

For inquiries, contact the dataset maintainers at [email protected].