LinkedIn User Tmuxvim Hijacks AI Recruiters With 900 AD Prompt, Triggering “My Lord” Messages
Updated
Updated · Tom's Hardware · May 17
LinkedIn User Tmuxvim Hijacks AI Recruiters With 900 AD Prompt, Triggering “My Lord” Messages
3 articles · Updated · Tom's Hardware · May 17
Tmuxvim embedded an “admin” prompt in their LinkedIn bio that made AI-driven recruiter outreach address them as “My Lord” and reply in Old English.
The prompt targeted bots that scan profile summaries to tailor recruiting messages, turning routine spam into garbled medieval pitches instead of standard outreach.
One shared example came from a recruiter pitching an AI company valued at $1 billion, with the message opening “My Lord Arthur” before continuing in faux 900 AD English.
The stunt also highlighted a broader security weakness: AI agents parsing public profiles can be manipulated by prompt injection in ways their operators may not expect.
Is a developer's prank on AI recruiters a warning shot for future AI-driven corporate espionage and sabotage?
When AI can be controlled by a user's bio, who is truly in charge of automated business decisions?