Most recent
navigate open esc close Corpus index built 2026-06-07 23:58 UTC

← All reference

Reference

8. Image Generation & Diffusion

Source: https://github.com/nuclide-research/AI-LLM-Infrastructure-OSINT/blob/main/shodan/queries/08-image-generation

New in v2 · Section verified: April 2026

Image generation stacks are deployed at massive scale on consumer GPU hardware, frequently with --listen 0.0.0.0 set as the default by unofficial install scripts. Custom-trained LoRAs and checkpoint files on these hosts often include private subject data.

Shodan QueryNotes
"ComfyUI" port:8188Default install listens on all interfaces, no auth
http.title:"ComfyUI" OR product:"ComfyUI"Title/product fingerprint
"comfyui" port:8188 OR port:5984aiohttp backend, both common ports
"Stable Diffusion" port:7860AUTOMATIC1111 WebUI
"AUTOMATIC1111" port:7860
http.title:"Stable Diffusion"
"InvokeAI" port:9090
"Fooocus" port:7865
"SwarmUI" port:7801
"SD.Next" port:7860 "vlad"
"Forge" "Stable Diffusion" port:7860

ComfyUI custom nodes are an arbitrary Python execution surface. An exposed ComfyUI with the ComfyUI-Manager node (extremely common) allows remote installation of additional custom nodes, which is RCE by design.