GhostShell: a haunted terminal that runs fully offline

A developer built GhostShell, a roughly 30-line Python terminal wrapper that runs each shell command via subprocess and then sends the command and its output to a local Ollama server running llama3.2:1b with a ghost persona prompt, printing a one-line eerie reply in purple beneath the output. The project keeps shell history, paths, hostnames and tokens on the local machine rather than sending them to a hosted API, and its UI layout was generated with Figma Make from a single prompt.

This is a submission for the Hacktoberfest Open-Source AI Challenge: Week 4 Ghost in the Machine . GhostShell is a terminal wrapper that "possesses" your shell. Every command runs normally, then a ghost whispers one eerie sentence about what you just did. It is about 30 lines of Python, and the UI was designed with Figma Make prompt in the README . Run python ghostshell.py , type ls or git status , and read what the spirit says. UI design Figma Make : https://blush-curl-39070128.figma.site/ https://blush-curl-39070128.figma.site/ . Code: https://github.com/chatgaiyashenanigan-alt/ghostshell https://github.com/chatgaiyashenanigan-alt/ghostshell The script runs your command with subprocess , then sends the command and its output to a local Ollama server llama3.2:1b with a ghost persona prompt. The reply is printed in purple under the output. Your shell history holds secrets: paths, hostnames, tokens in error messages. A closed API would send all of that to a server you don't control. With an open-weight model, nothing leaves the machine. It works on a plane with the Wi-Fi off, costs nothing per command, and I can swap the model by changing one string. A hosted model could change or disappear, but this ghost stays as long as the weights are on disk. I pasted a single prompt see README into Figma Make: a marigold-and-violet terminal with fading whisper lines, an offline badge, a model picker, and an "ofrenda" sidebar of past whispers. It gave me a layout I could build from.