
Dark-Moon v1.3.1
Motore di pentesting AI autonomo, sicurezza offensiva continua su web, cloud, AD e Kubernetes. Ragionamento agentico + esecuzione reale di exploit producono vulnerabilità basate su prove. Gateway di privacy: l'LLM non vede mai i tuoi IP, host, credenziali o percorsi reali (segnaposto deterministici reidratati localmente), nulla esce dal tuo perimetro.
DarkMoon
Open-source autonomous AI penetration testing — finds, exploits and qualifies every vulnerability on your own infrastructure
⭐ Star DarkMoon · 🚀 Quick Start · 🧩 Integrations · 📊 Benchmark · 🔬 Research · 🔒 Darkmoon Pro · ▶️ Install & Run (OSS) · ▶️ Demo (Pro)
[!IMPORTANT] License change. Darkmoon Community is moving from the GNU GPL v3.0 to the GNU AGPL v3.0. Every commit published up to
edfe9cb(27 October 2026) stays under the GNU GPL v3.0;masteris under the AGPL v3.0 since 19 November 2026 (PR #59), and v1.7.0 (20 November 2026) is the first AGPL release (v1.6.0 and earlier stay GPL v3.0). The contributions made by outside contributors before the switch also stay under the GNU GPL v3.0 (they are not relicensed). Details: LICENSING.md.
DarkMoon is autonomous AI penetration testing for your own infrastructure. Point it at an authorized target and it runs the whole assessment on its own, then documents every finding with the exact command and raw output.
- 🟢 Truly open source. AGPLv3 and self-hosted, every agent's methodology is plain Markdown you can read, diff and fork.
- 🎯 Finds AND proves. Each vulnerability ships with the exact command and raw output (exploitation is agent-asserted; the Pro remediation retest is the machine-verified step), so there is almost nothing to triage.
- 🔒 Runs on a local LLM + Privacy Gateway. The gateway tokenizes your real IPs, hosts and credentials locally, so the model reasons on placeholders while real values stay on your perimeter.
- 🕸️ An attack-path graph you can check. Agents declare what enabled what; the server keeps each link as a hypothesis (Possible Attack Path) and promotes it to Evidence-Backed only on its own records, with no model call to qualify it. Paths, their limits and the research behind them, including our local expert-AI prototype (research, not shipped): Attack paths · Research.
- 🧩 Everywhere you build. Run it from GitHub, GitLab, Jenkins, VS Code, JetBrains, n8n, Grafana or Splunk — the open-source edition works with the CLI-based ones. See the integrations ↓
See the open source engine (CLI)
The open source Darkmoon is a command line tool. This is what you get when you clone the repo. You launch an assessment from the command line and watch every agent, command and finding stream past in real time.
Kick off a run from the CLI. One TARGET line and DarkMoon takes over the whole assessment.
Then explore the results in the Terminal Explorer (type /explore in the console):
![]() Workspace. Executive block of a campaign. |
![]() Prioritized Treatment Plan. |
![]() Framework technical assessment. |
![]() Attack paths (Possible in Community). |
Community edition captures, data from a campaign on a local DVGA lab.



