Sensei is a CLI tool powered by Google Gemini, designed to assist penetration testers, CTF players, and cybersecurity students directly from their terminal.

• Context Aware: Understands standard tools (Nmap, Metasploit, Burp).
• Stealth & Ethics: Warns about OpSec failures.
• Terminal Beauty: Rich markdown rendering.
• Project Agnostic: Works on any Linux distro (Kali, Blackfin, Ubuntu).

git clone https://github.com/mjcc30/sensei.git
cd sensei
uv sync
uv run main.py ask how do I scan for SMB vulnerabilities?

Set your API Key once:

export GEMINI_API_KEY="your_key_here"

Quick Question:

# No quotes needed!
uv run main.py ask how do I use nmap stealthily?

Interactive Mode:

uv run app/main.py chat

Sensei exposes its tools (like Nmap) via a standard MCP Server. This allows other AI assistants (Claude Desktop, Gemini Advanced) to use Sensei's capabilities.

Run the Server:

uv run app/server.py

(This starts a stdio JSON-RPC server)

Claude Desktop Configuration: Add this to your claude_desktop_config.json:

{
  "mcpServers": {
    "sensei": {
      "command": "uv",
      "args": ["run", "/path/to/sensei/app/server.py"]
    }
  }
}

Sensei uses a default persona ("Helpful Mentor"). You can customize or override these prompts (e.g., for specialized research or Red Teaming) by creating a prompts.yaml file.

1. Copy the example:

   cp prompts.example.yaml prompts.yaml

2. Edit prompts.yaml to define your own System Prompts for each agent (Novice, Researcher, Master).
3. Note: prompts.yaml is ignored by Git to protect your custom methodology or sensitive instructions.

Sensei uses a Multi-Agent Orchestrator:

1. Router: Analyzes your query (Simple vs Complex).
2. Specialists: Delegates to the best agent (Novice, Researcher, or Master).
3. A2A Bus: Asynchronous internal communication.