Anthropic has revealed that its Claude OAuth system explicitly blocks the phrase “You are OpenCode” in second‑level system prompts, while allowing other identity statements such as Cursor or Pi. The test script, run with a Claude Pro or Max token, shows that any prompt containing OpenCode fails, confirming targeted censorship.

OpenCode, an open‑source AI coding assistant, competes directly with Anthropic’s Claude Code CLI. By blocking its name, Anthropic signals a defensive stance against rivals that could influence developer adoption. The move also raises questions about how AI platforms enforce brand‑specific restrictions within shared OAuth ecosystems, for developers global today and innovation.

Industry observers will watch whether Anthropic expands its filtering to other competitors and how OpenCode responds—perhaps by rebranding or lobbying for clearer OAuth guidelines. Regulators may also scrutinize such selective blocking as it touches on antitrust concerns and the broader debate over AI platform neutrality in the tech sector today and and