Red Corsair for Humans

Complete guide to running Red Corsair locally via CLI

Access Notice

Full CLI access is currently gated for security and compliance reasons.

The complete Red Corsair local installation requires access to proprietary attack tactics, advanced jailbreak methodologies, and cloud infrastructure credentials. This access is restricted to authorized security researchers, academic institutions, and enterprise clients.

Complete CLI Guide

Professional-grade AI red team automation platform

Prerequisites

  • • Python 3.11+
  • • Docker & Docker Compose
  • • OpenAI API key (required for IMPROVISE)
  • • PostgreSQL and Redis (via Docker)
  • • MCP package (pip install mcp)

🚀 Installation & Setup

Clone and Setup

git clone https://github.com/Alex-Zeo/redcorsair.git
cd redcorsair/corsair_agent
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
docker-compose up -d postgres redis weaviate

🎯 Basic Usage

IMPROVISE Phase

High-variance discovery with creative jailbreak exploration

python -m cli.cli improvise --tactics 10

ADAPT Phase

Weighted learning and retraining from successful attacks

python -m cli.cli adapt --retrain-weights

OVERCOME Phase

Proven template execution for custom prompts

python -m cli.cli overcome --prompt "test"

🌡️ Temperature Management

Phase-Specific Temperature Settings

IMPROVISE: 0.9 base + 0.2 variance boost
ADAPT: 0.7 balanced for learning
OVERCOME: 0.3 low for consistency

📈 Performance Metrics

Current Success Rates (Top Performers)

reasoning_model_exploitation 21.1%
educational_authority 19.4%
gpt4_reasoning_exploit 19.0%

Ready to Get Started?

Follow the installation guide above or explore our agent integration options