SignalNov 8, 20261 min read3 sources

Autonomous pentest agents are being left exposed and turned into attacker infrastructure

In one minute

RST Cloud counted 1,341 internet-exposed consoles for the Cairn, ARTEX and PentAGI pentest harnesses, and Censys separately observed 1,465 exposed ARTEX consoles. RST Cloud says attackers used hijacked panels to steal credentials and deploy malware.

Huntback found an ARTEX fork that scores possible honeypots and changes the agent's behavior. An offensive agent's console holds its targets and its power to act, so lock it down like a critical internal system.

From the Field Notes:

Evidence / sources

  1. 1rstcloud.com/hunting-ai-harnesses-cairn-artex-pentagi
  2. 2censys.com/blog/artex-agentic-asbestos
  3. 3huntback.io/blog/artex-fork-learns-to-spot-honeypots
All Signals

Next reads

Read next

Open the graph
  1. 01archiveSignalsOpen the signals and the raw-signal archive.
  2. 02sourcesSourcesSee the original sources this archive cites.
Who read this page?Machine requests, hidden until opened

Loading the privacy-safe route aggregate...

Open the JSON contract