Shopify CEO Considers Banning Claude Code Over Developer Workflow Issues

Shopify CEO Tobi Lütke says he is considering banning Claude Code across Shopify unless Anthropic changes how the coding agent handles project instruction files. Lütke argued that Claude Code’s preference for CLAUDE.md can create a “split brain” problem when different developers use tools that instead rely on standards such as AGENTS.md and .agents/skills.

The dispute is essentially about standardization in AI-assisted software development. Modern coding agents need project-specific instructions describing coding conventions, available tools and how an application should be modified. If different AI agents read different configuration files, teams can end up maintaining multiple instruction systems, increasing complexity and potentially causing inconsistent behavior between developers.

Lütke’s comments are particularly notable because Shopify has otherwise been strongly supportive of AI adoption. He has previously made effective AI usage a baseline expectation for employees and has required teams to consider whether AI can accomplish work before requesting additional resources or headcount.

The controversy therefore isn’t really about Shopify rejecting AI. Instead, it demonstrates that companies using multiple AI coding agents increasingly need common standards and interoperable tooling. Developers want to switch between systems such as Claude Code and other coding agents without having to maintain separate project instructions for every platform.

For Anthropic, the incident highlights an important issue as AI coding agents become part of professional software development. The quality of an agent is no longer determined only by how well it writes code; compatibility with existing developer workflows, configuration standards and team infrastructure is becoming equally important. If common standards emerge, companies may find it much easier to deploy multiple AI agents without creating additional operational complexity.