CyberStrikeAI

CyberStrikeAI is an AI-native security testing platform that orchestrates a vast library of security tools through an Intelligent Agent and the MCP protocol. It combines role-based testing, a skills system with specialized prompts, and a comprehensive lifecycle for conversations, tool invocations, and result visualization. Users can interact via a chat-style interface to initiate security assessments, assign predefined roles (e.g., Penetration Testing, Web App Scanning, API Security), and leverage 100+ integrated tools spanning network scanning, vulnerability assessment, subdomain enumeration, container and cloud security, and more. The system supports knowledge retrieval from a knowledge base, an attack-chain graph for analysis, and an auditable history with replay capabilities, making it suitable for collaborative security testing and compliance reporting.

Prerequisites:

• Docker installed on your host
• Access to a compatible OpenAI-compatible model (or Claude/DeepSeek) via an API key
• Optional: a GPU-enabled environment for faster AI inference

1. Prepare the environment

• Ensure Docker is running and you have network access to pull the container image.

2. Pull or build the CyberStrikeAI image

• If a prebuilt image is provided by your team: docker pull cyberstrikeai-image
• If you build locally, ensure the Dockerfile for CyberStrikeAI is available and build: docker build -t cyberstrikeai-image .

3. Run the MCP server

• Start the server using the provided container image: docker run -i cyberstrikeai-image

4. Initial configuration

• After startup, access the UI at http://localhost:8080 (or the port you exposed).
• Configure your OpenAI-compatible API key and base URL in Settings, for example: openai: api_key: "sk-your-key" base_url: "https://api.openai.com/v1" # or another compatible provider model: "gpt-4o"

5. Optional local validation (alternative to Docker)

• If you have Go and Python toolchains locally and prefer a manual run, follow the project’s Quick Start from the repository: git clone https://github.com/Ed1s0nZ/CyberStrikeAI.git cd CyberStrikeAI-main chmod +x run.sh && ./run.sh

Prerequisites (inline):

• Go 1.21+ and Python 3.10+ if building or running components directly
• macOS: brew install nmap sqlmap nuclei httpx gobuster feroxbuster subfinder amass
• Ubuntu/Debian: sudo apt-get install nmap sqlmap nuclei httpx gobuster feroxbuster

Notes:

• The run.sh script will validate environments, create a Python venv, install dependencies, download Go dependencies, build, and launch the server.
• If you use the Docker approach, ensure the container exposes port 8080 (or the port you configure) for UI access.

Tips and common issues:

• Ensure your OPENAI_API_KEY is valid and has access to the chosen model; if using a different provider (e.g., Claude or DeepSeek), adjust the base_url and model settings accordingly.
• If a required tool is missing in the container, the AI will attempt fallbacks or report tool unavailable; you can install missing tools into the container or adjust the toolset via Roles and Skills configurations.
• The system uses SQLite for conversation history and an audit trail; make sure the container has write permissions to its data directory or configure external persistence if needed.
• For CI/CD or deployment in cloud environments, map persistent volumes for logs, knowledge base, and SQLite databases to preserve state across restarts.