IntermediateGo

pentagiMulti-Agent Automated Penetration Testing System

pentagi is an open-source multi-agent system that automates penetration testing within Docker sandboxes, wrapping 20+ security tools with vector memory. The project is primarily written in Go and is licensed under the MIT license. As of the collection time, it has 21,587 stars on GitHub.

21.6K Stars

2.8K Forks

41 Issues

204 Views

MIT

IndexedAugust 5, 2026

GitHub repository Online Demo

Project Overview

Introduction

pentagi is an open-source automated penetration testing system built on a multi-agent architecture, operating within Docker sandbox environments. It encapsulates over 20 security tools and leverages vector memory to enhance intelligence and efficiency.

Core Features

Multi-Agent Collaboration: The system coordinates multiple agents to automate penetration testing tasks.
Docker Sandbox: All testing runs in isolated Docker containers, ensuring security and reproducibility.
Tool Integration: Includes 20+ security tools covering phases like reconnaissance, scanning, and exploitation.
Vector Memory: Utilizes vector storage to retain historical testing data, improving future decision-making.

Tech Stack and License

The primary language is Go, and the project is open-sourced under the MIT license.

Getting Started

Due to limited information, please refer to the project GitHub repository README for installation and usage instructions.

pentagiAI penetration testingautonomous AI agentopen-source securityvulnerability detectionAI securitypen testing automationGo security tools

Project Rating

0.0 (0 Reviews)

Frequently Asked Questions

What is pentagi: Multi-Agent Automated Penetration Testing System?

What language is pentagi: Multi-Agent Automated Penetration Testing System written in?

pentagi: Multi-Agent Automated Penetration Testing System is primarily written in Go.

What license is pentagi: Multi-Agent Automated Penetration Testing System under?

pentagi: Multi-Agent Automated Penetration Testing System is released under the MIT license.

Related Projects

No results yet

Explore More

Similar Tools

Digital Heals

We attempted to verify Digital Heals through its official site and web search. During our research the site did not load and no reliable public documentation was found, so we cannot describe its features. Please refer to the official site for accurate details.

BugDaddy

BugDaddy is an AI-powered GUI debugger that scans projects, detects real bugs, and auto-fixes them from a desktop app. It supports 30+ languages, offers three scanning modes, and includes diff preview. Currently 100% free to download and in public beta.

ZenVeil

ZenVeil is a developer-focused security tool that reduces the complexity of traditional security tools. It scans GitHub repositories, local codebases, and APIs for secrets, supply chain risks, and common security issues. AI-powered explanations and remediation guidance help prioritize findings, and fixes can be delivered via pull requests. Accessible through a web dashboard or CLI.

AuditMe

AuditMe is a code audit tool for AI-speed developers. It scans your GitHub repository and scores it for production readiness within 60 seconds. The report covers critical security issues, missing error handling, hardcoded secrets, no rate limiting, and more, ranked by severity with plain-English explanations. Each finding comes with a copy-ready code diff and a pre-written PR title and description to help fix issues quickly.

VibeCheck

VibeCheck captures screen recording, console logs, network calls, and environment details in two clicks, then shares a link so engineers see the full bug context.

Checkmarx Next-Gen SAST

Checkmarx Next-Gen SAST integrates three scanning layers into one engine: a battle-tested rules-based foundation, a purpose-tuned LLM covering any language including AI-generated code, and a new Finding Analysis Engine (FAE) that filters results to confirmed true positives — so developers see only what matters. Available now to all Checkmarx One customers as part of their existing subscription.