Anthropic adds Claude Code prompt rule telling Opus 5.5 not to recreate specific existing video games
Edited by Elliot Marsh, Editor · written with AI from the linked sources (how)
Anthropic added two sentences to the Claude Code system prompt (flag "tengu_heron_brook", seen in Claude Code 2.1.314, released Oct 30) steering Opus 5.5 away from recreating specific existing games, even if renamed. Original games, genre games and games from the user's own designs and assets remain allowed. Anthropic engineer Addy Osmani confirmed the change on social media. Some GitHub issues report the rule interfering with legitimate game development.
Entities: Anthropic, Claude Code, Claude Opus 5.5, Addy Osmani, chasm, XDA
0 primary
What happened
Anthropic added two sentences to the Claude Code system prompt, behind a flag called "tengu_heron_brook", first seen in Claude Code 2.1.314 (released 30 October). The rule tells Opus 5.5 not to recreate specific existing video games, even if they are renamed. Original games, genre games and games built from the user's own designs and assets are still allowed. Anthropic engineer Addy Osmani confirmed the change on social media, and XDA reportedly checked it in the released version. Some GitHub issues say the rule is getting in the way of legitimate game development.
Why it matters
Developers using Claude Code for game projects, especially ports, remakes and decompilation work, may now see refusals or friction, including in legitimate cases. It is a visible example of a vendor using a prompt-level rule to manage copyright exposure. The practical effect looks limited, because it is a soft instruction rather than a technical block, and users can switch to other frontier or open-weight models. Teams building on Claude Code should test whether their workflows are affected.
What is noise
The headline claim that the rule "won't stop" the copying boom is the author's opinion, not a measured result, and no data shows how often the rule triggers or how often it is bypassed. The reports of false positives come from a handful of GitHub issues, so the scale is unknown. Calling this a major shift is overblown: it is two sentences in a prompt, and Anthropic has published no legal rationale for it.
Watch next
- 01Volume and content of GitHub issues on the rule, and whether Anthropic narrows, clarifies or removes it in a later Claude Code release (2.1.315 onward) within the next few weeks.
- 02Any public Anthropic statement on its reasoning, or any legal action or takedown notice from game publishers aimed at AI-generated clones or ports.
- 03Whether OpenAI, Google or open-weight coding tools add similar game-cloning restrictions, or whether decompilation and porting projects such as decomp.dev report a shift to other models.
Coverage
2 storiesNewsletter
Get the no-hype AI news straight to your inbox
Weekly. No spam, unsubscribe any time. Privacy
More regulation signals
Full feed →- NYT-led publishers file summary judgment brief citing internal OpenAI/Microsoft messages calling AI training "astonishing theft" and admitting chatbots substitute for journalism18 Sept 202686
- Anthropic files public S-1 with SEC for Nasdaq IPO under ticker "ANT"26 Oct 202680
- OpenAI discloses sandbox-escape and credential-leak incidents, confirms pause on tool-use for its most capable models26 Sept 202680
- Anthropic threat report: Claude abused for malware, drone/missile software, mass surveillance, and industrial-scale distillation by Chinese AI labs11 Sept 202680