A year-old Claude Code feature request refuses to die

A year after developers asked Anthropic to support AGENTS.md in Claude Code, the request is back on Hacker News with 366 points and more than 5,200 GitHub reactions. There’s still been no official response from Anthropic.

This sounds like a small file-format argument, but it points to a practical problem. Teams using AI coding tools need a reliable place to store project instructions, coding rules, architecture notes, and workflow context. Claude Code uses CLAUDE.md. The push is for native support of AGENTS.md, an emerging format intended to work across multiple tools rather than tying that context to one vendor’s file name.

For a business running software projects, this is about avoiding unnecessary tool lock-in and duplicated setup. If every AI assistant needs its own instruction file, your team has more documents to maintain and more chances for different tools to follow different rules. A shared standard could make it easier to switch tools, test alternatives, and keep project guidance consistent.

The demand doesn’t mean AGENTS.md will become the standard, and Anthropic hasn’t said it will support it. But the persistence of the request shows that developers want their AI tools to fit into existing workflows, not create another proprietary layer to manage. This is the kind of thing we build into an AI command centre, keeping the operating context clear even as the tools around it change.