1 minute read

objective

create an AI-powered penetration testing automation platform that handles everything from domain discovery to exploit generation, leveraging AI to its fullest potential.

main features of the release

complete workflow

the platform is structured in 5 main phases:

  1. setup - automatic tool installation
  2. domain discovery - domain and subdomain discovery
  3. reconnaissance - http/https technology scanning
  4. scanning - vulnerability analysis (webapp + infrastructure)
  5. ai evaluation - multi-agent intelligent analysis

AI innovations

  • multi-AI crews: system of 3 specialized crews
    • analyst crew (2 agents) for vulnerability analysis
    • formatting crew (1 agent) for output formatting
    • correlation crew (1 agent) for exploit correlations
  • multiple LLM support via Ollama: qwen, deepseek, mixtral, llama, mistral

configuration features

  • YAML-based: pipeline and playbook configurable via YAML
  • modular scans: webapp/infra/full customizable
  • flexible LLM management: different models for specific tasks
  • flexible targets: automatic or predefined domains
  • gpu on-demand: automatic resource management

modern architecture

  • state saving with etcd for automatic recovery
  • distributed storage with minio for json results
  • automated deployment via ansible
  • integrated notifications via discord for monitoring
  • cloud management runpod for GPU resources

added value

this release represents a sophisticated platform that combines:

  • traditional tools of consolidated pentesting
  • modern AI for intelligent automation
  • correlated analysis of vulnerabilities
  • cloud scalability with automatic resource management
  • enterprise configurability for complex environments

the “why not?” philosophy is reflected in the integration of cutting-edge technologies to create a next-generation pentesting tool.

Updated: