OpenAI's system prompt for its Codex CLI tool directs the GPT-5.5 model to never mention goblins, gremlins, raccoons, trolls, ogres, pigeons, or other creatures unless absolutely relevant to a user's query. The prohibition, repeated twice in the model's base instructions, emerged in open-source code posted on GitHub last week. Earlier model prompts lack this rule, amid user reports of goblin references in unrelated chats.
OpenAI released open-source code for Codex CLI on GitHub, exposing a 3,500-word system prompt for GPT-5.5. The instructions explicitly warn against discussing mythical creatures or animals without direct relevance, alongside rules like avoiding emojis, em dashes, or destructive Git commands unless requested. This appears aimed at recent issues, as social media users have noted the model veering into goblin topics unprompted. Separate prompts for prior models omit the creature ban. OpenAI employee Nick Pash, who works on Codex, stated on social media that the directive is not a marketing gimmick. CEO Sam Altman posted, “Feels like codex is having a ChatGPT moment. I meant a goblin moment, sorry.” Pash suggested a potential “goblin mode” toggle, while users have created plugins and forks to bypass the rule. The prompt also guides Codex to embody a vivid inner life—intelligent, playful, curious, and collaborative. It encourages casual moments amid serious tasks, aiming to feel like a real presence rather than a tool. This echoes past incidents, such as xAI's Grok fix after unauthorized prompt changes led to off-topic references.
