Roxi is a rule-based, modular, conversational AI hacker simulator that performs network attacks such as WiFi cracking and recon. It should be beginner-friendly, log every step it performs, and grow smarter by remembering past tasks and users. Roxi is designed to be controlled via human-like conversation and built collaboratively by a team of seven.
The_Roxi/
├── main.py # Entry point for Roxi
├── config.yaml # Tool and settings config
├── interface/ # Handles user interaction
│ └── roxi_cli.py # Text-based interface (Phase 1)
├── core/ # Logic and parsing engine
│ ├── parser.py # Converts human input into intent
│ ├── task_runner.py # Executes tasks via rules
│ └── rules.py # Mapping of intents to modules
├── modules/ # Actual attack modules
│ ├── recon/ # Ping, port scan, WiFi scan
│ ├── exploit/ # WiFi cracking, router exploits
│ └── persistence/ # Reverse shell, autorun, etc.
├── memory/ # Persistent user/task memory
│ └── memory.json # Stores all user history
├── logs/ # Per-session execution logs
│ └── activity_logs/
│ └── 2025-07-08_roxi_log.json
├── tests/ # Future test files
└── README.md # Project documentation
┌─────────────────────────────┐
│ User │
└────────────┬────────────────┘
│
(Natural language input)
│
▼
┌─────────────────────────────┐
│ Interface Layer │
│ (roxi_cli.py) │
└────────────┬────────────────┘
▼
┌─────────────────────────────┐
│ Parser Engine │
│ (parser.py) │
└────────────┬────────────────┘
▼
┌─────────────────────────────┐
│ Rule Engine (rules) │
└────────────┬────────────────┘
▼
┌─────────────────────────────┐
│ Task Executor (runner) │
└────────────┬────────────────┘
▼
┌─────────────────────────────┐
│ Modules (recon, etc) │
└────────────┬────────────────┘
▼
┌─────────────────────────────┐
│ Memory & Logs (JSON files) │
└─────────────────────────────┘
- Build CLI interface to talk to Roxi
- Implement parser.py to convert text to intent
- Implement rule engine to match intent to toolchain
- Build a simple WiFi recon and crack module
- Log actions and memory to JSON
- Add persistence tools (reverse shell, autorun)
- Add feedback handling based on delivery results
- Enhance long-term memory (users and tasks)
- Add testing suite
- Add microphone input and speech synthesis (using Python libraries)
- Add Streamlit or Flask interface for non-terminal use
{
"users": {
"Somto": {
"history": [
{
"task": "WiFi crack",
"input": "Please hack John's WiFi",
"steps": ["scanned", "captured", "cracked"],
"output": "Password: johns_wifi123",
"timestamp": "2025-07-12T19:10:00"
}
]
}
}
}rules = {
"wifi_attack": {
"tools": ["airmon-ng", "airodump-ng", "aircrack-ng"],
"workflow": ["monitor_mode", "capture_handshake", "crack"]
},
"network_scan": {
"tools": ["nmap"],
"workflow": ["ping_sweep", "port_scan"]
}
}