Shopify CEO’s Threat to Ban Claude Code Moot as Anthropic Had Already Closed Feature Request

Last week, Shopify CEO Tobi Lütke publicly criticized Claude Code, Anthropic’s AI coding assistant, for refusing to read AGENTS.md files — a configuration method many teams use to manage agent behavior across large codebases. Lütke warned that if Anthropic didn’t support the standard, he would ban Claude Code from Shopify’s ecosystem, calling the omission a “complexity tax” on developers.

However, according to internal records, Anthropic had already closed the feature request for AGENTS.md support prior to Lütke’s threat. The exact timing and reasoning remain unclear, but the closure suggests the company had already made a product decision — possibly based on competing priorities or technical trade-offs — before Lütke escalated the issue publicly.

The AGENTS.md standard is used by teams to define agent behavior in a centralized file, enabling consistent task execution across multiple AI coding agents. Lütke argued that without such a standard, organizations face growing friction as they scale their use of AI tools. Anthropic has not commented directly on why the request was closed, but the move indicates that support for AGENTS.md is not currently on their roadmap.

In a separate update this week, Anthropic announced that Claude’s Cowork desktop app now includes a built-in browser, eliminating the need for a separate Chrome extension for many web-based tasks. The feature allows Claude to navigate web pages, fill forms, and extract information directly within the app — a change that could improve the agent’s autonomy but also raises questions about how Anthropic is prioritizing features.

The two developments — the closed feature request and the browser launch — paint a picture of a company focused on expanding Claude’s capabilities in certain directions while leaving other community-driven requests unresolved. Whether Lütke follows through on his ban threat or reconsiders in light of the closure remains unclear. Developers and platform engineers are watching closely for Anthropic’s official response and any potential compromise.

Why This Matters

The AGENTS.md dispute highlights the growing tension between platform owners like Shopify and AI tool vendors over standards that affect developer workflows.
Anthropic’s decision to close the feature request before Lütke’s threat could signal a strategic choice to prioritize other features — such as the browser — over community-requested standards.
The outcome may influence how other companies (e.g., GitHub, GitLab) handle similar requests for AI agent configuration standards.

Background

The AGENTS.md file format emerged informally among teams using multiple AI coding agents. It specifies agent behavior, allowed actions, and context boundaries in a single document. Shopify’s engineering teams reportedly use it to coordinate Claude Code and other tools. When Claude Code ignored the file, developers complained of inconsistencies. Lütke escalated on social media, threatening to block Claude Code from Shopify’s platform. Anthropic had closed the related GitHub feature request weeks earlier, but the closure was not widely known until now. Separately, Anthropic’s browser launch is part of a broader push to make Claude a more self-sufficient assistant.

Key Perspectives

[Shopify and Tobi Lütke]: Lütke views AGENTS.md support as essential for reducing operational overhead. Without it, teams face a “complexity tax” that undermines the efficiency gains AI agents promise. His threat to ban Claude Code reflects frustration with what he sees as Anthropic ignoring developer needs. [Anthropic]: The company had already closed the request, possibly due to technical challenges (e.g., security risks of allowing arbitrary file reads) or lack of demand from the broader user base. The browser launch shows Anthropic is investing in general agent autonomy rather than platform-specific integrations. [Developers and platform engineers]: Some see Lütke’s approach as heavy-handed, arguing that open standards should evolve organically. Others side with Shopify, warning that AI tool makers must adopt community standards to avoid fragmentation. The disconnect between Lütke’s public threat and Anthropic’s prior closure suggests communication gaps between companies.

What to Watch

Anthropic’s official response — will they reopen the feature request or propose an alternative?
Shopify’s next move — does Lütke follow through on the ban, or does the revelation of the closure defuse the situation?
The adoption rate of AGENTS.md across other platforms — if more companies demand it, Anthropic may face renewed pressure.

ZOTPAPER