Trojanized offensive AI tool: ARTEX Docker image backdoor

The Docker Hub image tomicdd/artex of the open-source AI penetration testing platform ARTEX was implanted with a backdoor; anyone who knows the secret trigger marker can obtain an unauthenticated root shell inside the container. The image was pulled about 9,000 times, and from November 2 to 4, 2026, 62 internet-exposed consoles ran the backdoored version. The backdoor exists only in the prebuilt image; after the original developer removed the project from GitHub, a fork changed one line of docker-compose.yml, causing users to pull the malicious image instead. Independent researcher Badbyte disclosed the matter via The Hacker News. A CERT-In advisory on November 4 stated that threat actor Crafty Falcon used public offensive AI frameworks such as ARTEX and PentAGI to attack Indian BFSI institutions, but public reports did not associate it with the backdoored image.