Anthropic has altered Claude Code's prompt on its Opus 5.5 model to stop the tool from recreating specific existing video games, and users are also reporting that the same model is refusing reverse engineering work. The change affects developers using Claude Code with Opus 5.5, the version on which the new instructions were applied.
Addy Osmani stated in a full post on X that Anthropic added two sentences to Claude Code's prompt on Opus 5.5 to steer it away from recreating someone else's specific game, because using Claude to infringe intellectual property violates the company's usage policy. He said original games are meant to keep working, along with games in a genre and games built from a user's own designs and assets. He added that Anthropic will keep iterating to strike the right balance, and asked users to send him cases where Claude turns down something that is theirs.
The exact wording of the two sentences has not been confirmed. Based on Osmani's description, it may match a screenshot posted by X user Chasm, which reads that the assistant should build original games, genre games and games from the user's own designs and assets, and should not build a recreation of a specific existing game even if it is renamed or called original.
The game-related refusals follow complaints filed at Claude Code's own GitHub repository, where users said the assistant would not work on their games or on reverse engineering tasks. So far the game blocking appears limited to the combination of Claude Code and Opus 5.5. If that holds, users could avoid the restriction by calling Opus through another harness such as Ollama or by routing specific requests through a wrapper.
Reverse engineering blocks have been directly confirmed. Jarred Walton, an alumnus of the publication, was running checks on software he is developing; the work had proceeded without issue until October 31, and switching to Opus 4.8 let him continue. Walton theorizes that Anthropic added a rule to Code specifically for reverse engineering, and the timing plus other refusals in the past week may indicate the company is adjusting cybersecurity-related rules the same way it is handling video game content.
Screenshots from a staffer's direct acquaintance showed that asking Opus 5.5 through Code to reverse engineer a product activation triggered safeguards. Reports of Claude Code blocking reverse engineering continue to mount, including a Redditor whose account was allegedly suspended and then reinstated over use of Opus 5.5 and Reverse Engineer Anything. Anthropic runs an ongoing Cyber Verification Program in which only approved entities receive access to the full set of frontier models' cybersecurity capabilities.
More AI news from TechManNews.
