Breaking

METATRON, a Fully Local AI Penetration-Testing Assistant, Released

July 1, 2026 at 09:55 EDT

Security
Open Source
AI Agents

A new open-source tool called METATRON has been released. Running on a local LLM via Ollama, it assists with penetration testing from the command line without requiring any cloud service or API key.

Open-Source Security Tooling · Nebula by BerylliumSec

AI Lives Inside the Command Line for Pentesters

Nebula embeds large language models directly into the CLI — powering reconnaissance, vulnerability suggestions and automated note-taking. A human-led assistant, not an autonomous agent, sitting between the two approaches.

989

GitHub stars — a mature project as of 2026

LLMs supported, plus any model via the OpenAI API

5 min

Status-feed refresh interval during a session

Local inference hardware — CPU vs GPU memory footprint

16GB+

CPU (Ollama) RAM

8GB+

GPU memory

Each block ≈ 2GB. GPU inference tested with 12GB · Python 3.10–3.13.9

How Nebula fits the pentest loop

Run tools
nmap · Metasploit

→

AI reads output
vuln suggestions

→

Auto-log & notes
screenshots · commands

Praised for

AI blends naturally into CLI pentest workflows
Automated notes cut documentation burden
Guides beginners and pros; ready for immediate adoption

Caveats

Assistant only — won't produce results autonomously
Pure CLI can be a barrier for command-line newcomers
Autonomous mode reserved for paid Nebula Pro

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.