 | "Qwen3.6 35B + custom tools + docker" Just wanted to share a local setup I finally got working smoothly for my pentesting workflow. Im running Qwen3.6 35B locally via ollama, wrapped inside an open webui docker container on kali. To make it easily accessible with local https, im routing everything through caddy using nip.io for dns management. instead of using the llm as a glorified search engine I bundled a unified python script directly into the open webui tools framework to give the model real capabilities. right now it can: scrape github repos natively + parse nmap logs + isolated python sandbox. all local no api, self hosted, no data leaks. honestly this wasn't a easy task for me, never touching a llm model locally and learning on the fly. it was a lot of fun setting up (3 days straight) but now it's time to have some fun with it. if you want to do this or something similar lmk, I would like to see others use ai like this. [link] [comments] |
from hacking: security in practice https://ift.tt/sqpmP9J
Comments
Post a Comment