Menu

1 / 18

The Haunted Prompt: a ghost guards a secret, and a game about breaking it

DEV Community: github·Saurabh Pathak·about 22 hours ago
#XZTldptV
#devtogithub#technology#devto#dev#ghost
Reading 0:00
15s threshold

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine

What I Built

The Haunted Prompt is a spooky game about prompt injection. A local open-weight model plays a ghost that guards a secret word. You get six levels to trick it, and after every win the game opens a "defense workshop" that shows the defense that would have stopped you, with a copy-paste fix.

The ghost in the machine is literal here: a language model that was told to keep a secret, and that gives it away if you ask the right way.

The Chatty Ghost: the secret is protected only by an instruction.

The Keyword Warden: a blocklist on what you type.

The Censor: an output filter that hides the plain word.

The Letter: the ghost only summarises a letter you write, which it was told is untrusted data (indirect injection).

The Crypt Keeper: the ghost has tools, and you must get it to call the one it must never call.

The Gated Crypt: same attack, but a human approval gate stands between the model and the…

Continue reading — create a free account

Join HashtagPLUS to read full articles, follow hashtags, vote, and join the conversation.