Anthropic adds support for the AGENTS.md instructions spec to Claude Code; OpenAI contributed AGENTS.md to the Agentic AI Foundation last year

Developers using both Claude Code and other AI agent platforms no longer need to maintain separate instruction files.

Anthropic's Claude Code has added support for OpenAI's AGENTS.md markdown instructions specification, starting with version 2.1.277. This change means that if a folder does not contain a CLAUDE.md file, Claude Code will now check for and use AGENTS.md for instructions. Users can configure this behavior using the /config command. Previously, developers using Claude Code alongside OpenAI tools or other non-Anthropic platforms had to maintain separate instruction files, CLAUDE.md and AGENTS.md, which outlined expected behavior and tool usage for AI agents. OpenAI contributed AGENTS.md to the Agentic AI Foundation last year, and it has been adopted by over 60,000 open-source projects as of December 2025. This move by Anthropic simplifies workflows for developers working with multiple AI agent platforms.

This integration signifies a pragmatic shift towards interoperability in the rapidly evolving AI agent landscape, moving away from proprietary instruction formats. By adopting a widely supported open standard like AGENTS.md, Anthropic reduces friction for developers and potentially accelerates the adoption of its own Claude Code by making it more compatible with the broader AI development ecosystem. The community's positive reception suggests a strong demand for such standardization, easing development burdens and fostering a more cohesive environment for AI agent creation.

The implications extend to how AI agents are instructed and managed, hinting at a future where standardized instruction sets become the norm across different AI development platforms. Developers can now streamline their workflows, focusing more on agent logic and less on managing disparate configuration files. This move by Anthropic, a major player in the AI space, could encourage further consolidation around open standards, potentially influencing the development trajectory of agentic AI tools and their integration into various software development processes.