Anthropic confirms new Claude Code limits to prevent game cloning — users report Opus 5.5 is also actively blocking reverse engineering tasks
Anthropic employee confirms that Claude Code now has orders not to reproduce game IP — company "iterating to ensure [...] the right balance"
Anthropic has introduced new safeguards within Claude Code, specifically for Opus 5.5, to prevent users from reproducing copyrighted video games or reverse-engineering software. Anthropic spokesperson Addy Osmani confirmed via tweet the addition of two sentences to Claude Code's prompt, which instruct the AI to refrain from creating specific game replicas and disallowing reproduction of existing intellectual property.
These rules still permit users to build original, genre-based, or custom-designed games. This development has sparked criticism from the software development community, with users reporting instances where Claude Code has denied assistance for game creation and reverse engineering tasks. Users have taken to social media and GitHub to express their concerns, alleging that Claude Code has begun blocking their requests for game-related work and reverse engineering.
Some have theorized that these restrictions might be limited to the Ollama AI harness, while others suggest that Anthropic could be working on cybersecurity-related rules. Redditors have also reported account suspensions after using Opus 5.5 for reverse engineering, further highlighting the controversy surrounding these new restrictions.
Anthropic maintains a Cyber Verification Program, accepting only pre-approved entities to access the company's most advanced AI models.
Written by urgent.news from Tom's Hardware's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
