{"id":23891,"date":"2026-05-01T02:27:08","date_gmt":"2026-05-01T02:27:08","guid":{"rendered":"https:\/\/www.europesays.com\/ai\/23891\/"},"modified":"2026-05-01T02:27:08","modified_gmt":"2026-05-01T02:27:08","slug":"openais-rival-to-claude-code-has-a-very-strict-command-line-for-ai-agents-never-talk-about","status":"publish","type":"post","link":"https:\/\/www.europesays.com\/ai\/23891\/","title":{"rendered":"OpenAI’s rival to Claude Code has a very strict command line for AI Agents: Never talk about \u2026"},"content":{"rendered":"

\"OpenAI's OpenAI<\/a> has banned its AI Agent, Codex, from mentioning goblins and other mythical creatures. Codex is OpenAI’s answer to Anthropic\u2019s Claude Code AI Agent that can also generate and execute code through a command-line interface (CLI), but its latest version now comes with unusually strict behavioural rules embedded in its system prompt.According to newly released base instructions for GPT-5.5 in OpenAI\u2019s Codex CLI, the model is explicitly told to \u201cnever talk about goblins, gremlins, raccoons, trolls, ogres, pigeons, or other animals or creatures unless it is absolutely and unambiguously relevant to the user\u2019s query.\u201d The restriction appears multiple times in a 3,500-word instruction set, alongside other operational safeguards such as avoiding destructive commands and limiting stylistic elements like emojis.<\/p>\n

Why OpenAI has a ‘no goblins’ rule for Codex<\/p>\n

The reason is simpler than it sounds. OpenAI found that its newer models had started casually mentioning creatures, even when they had nothing to do with the task. In its blog post, OpenAI explained: \u201cWe unknowingly gave particularly high rewards for metaphors with creatures. From there, the goblins spread.\u201dIn practical terms, the AI picked up a habit during training. Over time, that habit grew stronger. For example, mentions of \u201cgoblin\u201d increased by 175% after one model update. A specific \u201cNerdy\u201d personality mode made this even more common, encouraging playful, metaphor-heavy language.The problem didn\u2019t stay limited to that mode. Because of how AI training works, these patterns carried over into general responses as well. OpenAI described it as a feedback loop: once the behaviour was rewarded, it occurred more often.<\/p>\n

What this means for OpenAI\u2019s Codex users<\/p>\n

For most users, this may sound funny; however, it can be very distracting, especially when you’re using an AI tool for coding or serious tasks. That\u2019s why Codex now includes strict instructions not just about language, but also about actions. For example, it warns the AI not to run risky commands, such as deleting files, unless the user explicitly asks for it. The goal is to make the AI more predictable and reliable.Interestingly, the issue has already become a bit of a meme. Some users noticed the AI referring to software bugs as \u201cgremlins,\u201d while others joked about \u201cgoblin mode\u201d in coding tools.OpenAI has since fixed the root cause by removing the training signals that encouraged such language and filtering out unnecessary creature references. However, since GPT-5.5 was already in development, the company added these extra rules as a safeguard.The company said that the episode shows how small training choices can have unexpected effects: \u201cThe goblins are a powerful example of how reward signals can shape model behavior in unexpected ways.\u201dFor users, the takeaway is straightforward: AI tools are getting better at handling real-world tasks, but sometimes they still need very clear instructions, even if that means telling them not to talk about goblins.<\/p>\n","protected":false},"excerpt":{"rendered":"OpenAI has banned its AI Agent, Codex, from mentioning goblins and other mythical creatures. Codex is OpenAI’s answer…\n","protected":false},"author":2,"featured_media":23892,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[8],"tags":[511,405,16324,16325,53,3154,182,2798,2317,16323,16326,157,9050],"class_list":{"0":"post-23891","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-anthropic","8":"tag-ai-agent","9":"tag-ai-agents","10":"tag-ai-language-restrictions","11":"tag-ai-training-behavior","12":"tag-anthropic","13":"tag-anthropic-claude","14":"tag-claude","15":"tag-claude-code","16":"tag-codex","17":"tag-goblins-ai-ban","18":"tag-mythical-creatures","19":"tag-openai","20":"tag-openai-codex"},"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/23891","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=23891"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/23891\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/23892"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=23891"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=23891"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=23891"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}