Anthropic Adds Guardrail to Stop Claude Code Cloning Existing Games
Anthropic has added a new instruction to Claude Code that stops the AI tool from recreating a specific existing game, while still supporting original games and genre-based projects.
The Claude Code game rule is a new instruction from Anthropic that changes how Claude Code, its AI coding tool, handles requests to recreate a specific existing video game. The change was confirmed on October 30, 2026, after users spotted a new rule in the tool’s server-delivered settings. The instruction tells Claude Code to build original games, games in a genre and games from the user’s own designs and assets, but not to build a recreation of a specific existing game, even if it is renamed or described as original. Anthropic technical staff member Addy Osmani confirmed the update, saying the company added two sentences to the prompt for its Opus 5.5 model to steer it away from recreating someone else’s specific game, because using Claude to infringe intellectual property is against its usage policy.
The rule follows a wave of AI-assisted game clones that spread online after the Opus 5.5 model was released. Reports said browser-playable versions of well-known titles began circulating, with Claude credited for helping to rebuild them. Anthropic said the change is narrow: it targets copies of one particular title, not game development in general. Original games, genre games and projects built from a user’s own designs and assets are meant to keep working. The instruction was first spotted by a user posting as Chasm around October 27, and Anthropic confirmed it three days later.
Claude Code game rule and the limits of AI guardrails
The Claude Code game rule raises a practical question for developers: how reliably can a short instruction stop a powerful AI model from copying a game? Anthropic has announced an intended behaviour, not demonstrated its effectiveness, and some developers have already reported refusals affecting legitimate work, including reverse engineering and security testing. Anthropic acknowledged that legitimate projects may sometimes be rejected and said it will continue adjusting the restriction to strike the right balance. The episode shows how AI companies are now drawing lines in real time, responding to misuse with prompt-level guardrails rather than waiting for a full product update. For game studios, the rule is a welcome signal that AI tools should respect creative work. For developers, it is a reminder that the tools they rely on can change quickly, with new limits arriving without warning and without a setting to switch them off.
Source: Polygon
