Quick start

Get started with RAPTOR in minutes using Claude Code or Python CLI

Overview

RAPTOR offers two interfaces:

• Claude Code (recommended): Natural language interface with interactive workflows
• Python CLI: Command-line interface for scripting and CI/CD integration

This guide covers both approaches.

Claude Code quick start

Use RAPTOR with plain English in Claude Code via slash commands.

1. Install Claude Code

   Download Claude Code from https://claude.ai/download

2. Clone and open RAPTOR

   git clone https://github.com/gadievron/raptor.git
   cd raptor
   claude
   
3. Install dependencies

   Let Claude install dependencies for you:

   "Install dependencies from requirements.txt"
   "Install semgrep"
   "Set my ANTHROPIC_API_KEY to [your-key]"
   

   Check DEPENDENCIES.md for licenses of the various tools before installing.

4. Start using RAPTOR

   Just say “hi” to get started, then try:

   /scan - Scan code for vulnerabilities
   /fuzz - Fuzz binaries (asks to install AFL++ if needed)
   /web  - Test web applications
   

   Try /analyze on one of the test cases in /tests/data

Natural language examples

Just talk naturally to Claude:

Code scanning

/scan - scan this directory
/scan - find secrets in my code
/scan - check for SQL injection vulnerabilities

Binary fuzzing

/fuzz - fuzz this binary
/fuzz - test ./myapp for 5 minutes
/fuzz - find crashes in /usr/local/bin/myapp

Web testing

/web - test my website
/web - scan https://localhost:3000 for XSS
/web - check this app for OWASP Top 10

The /web command is currently in alpha status. Treat as experimental.

Interactive workflow example

Here’s what a typical RAPTOR session looks like:

You: /scan - scan this code

Claude: I'll scan the current directory with RAPTOR...
        [Runs scan...]
        Found 5 vulnerabilities!

        1. SQL Injection (CRITICAL) - api/db.py:23
           Generated exploit: out/exploits/sqli_001.py
           Generated patch: out/patches/sqli_001.patch

        Would you like me to:
        1. Apply the patches?
        2. Explain the SQL injection?
        3. Show the exploit code?

You: Show me the exploit

Claude: [Reads and displays the exploit code]
        This exploit works by...

You: Apply the patch for that one

Claude: [Applies the patch using Edit tool]
        ✓ Patch applied! The SQL injection is now fixed.

        Want me to scan again to verify the fix?

Benefits of Claude Code interface

• No command-line syntax to remember
• AI understands your intent
• Results explained in plain English
• Interactive fix workflow
• Fast and autonomous

Python CLI quick start

For scripting or CI/CD integration, use the Python CLI directly.

1. Clone the repository

   git clone https://github.com/gadievron/raptor.git
   cd raptor
   
2. Install Python dependencies

   pip install -r requirements.txt
   
3. Install external tools

   # Install Semgrep
   pip install semgrep
   
   # Optional: Install AFL++ for fuzzing
   # macOS:
   brew install afl++
   
   # Ubuntu/Debian:
   sudo apt install afl++
   
4. Configure LLM provider

   # For Anthropic Claude (recommended)
   export ANTHROPIC_API_KEY=your-key-here
   
   # For OpenAI
   export OPENAI_API_KEY=your-key-here
   
   # For local Ollama (free)
   export OLLAMA_HOST=http://localhost:11434
   
5. Run your first scan

   # Full autonomous workflow
   python3 raptor.py agentic --repo /path/to/code
   
   # Static analysis only
   python3 raptor.py scan --repo /path/to/code
   
   # Binary fuzzing
   python3 raptor.py fuzz --binary /path/to/binary --duration 3600
   

Python CLI examples

Scan repository

# Scan for all vulnerabilities
python3 raptor.py scan --repo /path/to/code

# Scan for specific policy groups
python3 raptor.py scan --repo /path/to/code --policy_groups secrets

Using the devcontainer

A devcontainer with all prerequisites pre-installed is available for easy onboarding.

1. Open in VS Code

   Use the command Dev Container: Open Folder in Container in VS Code or any of its forks.

2. Or build with Docker

   docker build -f .devcontainer/Dockerfile -t raptor-devcontainer:latest .
   

What’s included in the devcontainer

Pre-installed security tools:

• Semgrep (static analysis)
• CodeQL CLI v2.15.5 (semantic code analysis)
• AFL++ (fuzzing)
• rr debugger (deterministic record-replay debugging)

Build & debugging tools:

• gcc, g++, clang-format, make, cmake, autotools
• gdb, gdb-multiarch, binutils

Web testing (alpha):

• Playwright browser automation (Chromium, Firefox, Webkit browsers)

The devcontainer is massive (~6GB), starting with Microsoft Python 3.12 devcontainer and adding static analysis, fuzzing, and browser automation tools.

The devcontainer runs with --privileged flag required for rr debugger.

Available commands

Main entry point

/raptor   - RAPTOR security testing assistant (start here for guidance)

Security testing

/scan     - Static code analysis (Semgrep + CodeQL)
/fuzz     - Binary fuzzing with AFL++
/web      - Web application security testing (alpha)
/agentic  - Full autonomous workflow (analysis + exploit/patch generation)
/codeql   - CodeQL-only deep analysis with dataflow
/analyze  - LLM analysis only (no exploit/patch generation - 50% faster & cheaper)

Exploit development & patching

/exploit  - Generate exploit proof-of-concepts (beta)
/patch    - Generate security patches for vulnerabilities (beta)
/crash-analysis - Analyze an FFmpeg crash and generate validated root-cause analysis

Forensics & investigation

/oss-forensics - Evidence-backed forensic investigation for public GitHub repositories

Development & testing

/create-skill    - Save custom approaches (experimental)
/test-workflows  - Run comprehensive test suite (stub)

Example output

Here’s what RAPTOR output looks like:

╔═══════════════════════════════════════════════════════════════════════════╗ 
║                                                                           ║
║             ██████╗  █████╗ ██████╗ ████████╗ ██████╗ ██████╗             ║ 
║             ██╔══██╗██╔══██╗██╔══██╗╚══██╔══╝██╔═══██╗██╔══██╗            ║ 
║             ██████╔╝███████║██████╔╝   ██║   ██║   ██║██████╔╝            ║ 
║             ██╔══██╗██╔══██║██╔═══╝    ██║   ██║   ██║██╔══██╗            ║ 
║             ██║  ██║██║  ██║██║        ██║   ╚██████╔╝██║  ██║            ║ 
║             ╚═╝  ╚═╝╚═╝  ╚═╝╚═╝        ╚═╝    ╚═════╝ ╚═╝  ╚═╝            ║ 
║                                                                           ║ 
║             Autonomous Offensive/Defensive Research Framework             ║
║             Based on Claude Code - v1.0-beta                              ║
║                                                                           ║ 
╚═══════════════════════════════════════════════════════════════════════════╝ 

⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢀⣠⣤⣤⣀⣀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣾⣿⣿⠿⠿⠟
⠀⠀⠀⠀⠀⠀⠀⠀⢀⣀⣀⣀⣀⣀⣀⣤⣴⣶⣶⣶⣤⣿⡿⠁⠀⠀⠀
⣀⠤⠴⠒⠒⠛⠛⠛⠛⠛⠿⢿⣿⣿⣿⣿⣿⣿⣿⣿⣿⠟⠁⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠉⠛⣿⣿⣿⡟⠻⢿⡀⠀⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢀⣾⢿⣿⠟⠀⠸⣊⡽⠀⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢸⡇⣿⡁⠀⠀⠀⠉⠁⠀⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠈⠻⠿⣿⣧⠀ Get them bugs.....

Next steps

Installation

Detailed installation instructions and environment setup

Architecture

Learn about RAPTOR’s technical architecture

Claude Code usage

Complete guide to using RAPTOR with Claude Code

Python CLI

Full Python command-line reference