Skip to content
Hot eventDeveloping

ARTEX autonomous pentesting platform released as self-hosted fork

1 reports1 reporting sourcesUpdated 1 h ago

Start here

AI digest

BaaRai published an English-translation fork of ARTEX, a self-hosted, LLM-driven penetration testing platform originally by Autumn-27. A planner splits goals into intents and workers execute them with real tools such as nmap, nuclei and sqlmap; a shared asset graph and a per-task exploration graph linked by anchors let each finding carry the captured request and response behind it. Autonomy is bounded by scope and deny gates, human approval rules, token and turn budgets, a target noise governor, and prompt-injection framing of tool output. It ships as a single Go binary with the Next.js UI embedded, requires PostgreSQL, and needs an LLM endpoint (Anthropic, OpenAI-compatible or local Ollama) for exploration. There is no prebuilt Docker image, so the image is always built from the checkout, and the UI listens on 127.0.0.1 by default. The project is under AGPL-3.0 and is described as an offensive-security tool to be used only against systems the user owns or is explicitly authorized to test.

Generated from reports · updated 1 h ago

Report timeline

Follow the reports to see the event from different sides.

November 16
  1. Hacker News · AI
    ARTEX releases self-hosted LLM multi-agent autonomous penetration testing platform

    ARTEX is a self-hosted, LLM-driven penetration testing platform where a planner splits goals into intents and workers execute them with real tools such as nmap, nuclei and sqlmap. It uses a dual-graph design, a shared asset graph and a per-task exploration graph linked by anchors, so every finding can carry the captured request and response that justify it. Autonomy is bounded by scope and deny gates, human approval rules, token and turn budgets, a target noise governor and prompt-injection framing of tool output, and it runs as a single Go binary with the Next.js UI embedded plus PostgreSQL, using Anthropic, OpenAI-compatible or local Ollama endpoints under AGPL-3.0.

Heat for this event

Not enough continuous observations yet to draw a trend.