Anthropic adds support for the AGENTS.md instructions spec to Claude Code; OpenAI contributed AGENTS.md to the Agentic AI Foundation last year (Thomas Claburn/The Register)
Devs cheer relief from compatibility headache — Anthropic on Friday surprised the developer community by supporting rival OpenAI's mechanism for passing marching orders to AI agents.
Anthropic made an unexpected announcement on Friday that could simplify the lives of developers who use multiple AI platforms. The company has decided to support OpenAI's AGENTS.md mechanism for passing instructions to AI agents within its Claude Code tool.
Claude Code engineer Thariq Shihipar confirmed the update in a social media post, stating that starting from version 2.1.277, Claude will automatically look for and utilize AGENTS.md if it cannot find a CLAUDE.md file in a specific folder. Developers can adjust this behavior using the /config command within the platform.
Prior to this update, users who relied on both Claude Code and OpenAI Codex or other non-Anthropic tools had to maintain separate sets of Markdown instructions - CLAUDE.md and AGENTS.md - for their AI agents, along with any project-specific variations of these files. These documents provided the context for coding agents to operate, detailing expected behavior, preferred tool usage, and coding conventions.
While CLAUDE.md and AGENTS.md share a similar structure, they are not identical - CLAUDE.md may contain instructions specific to Anthropic's model while AGENTS.md aims to be more universally applicable. Developers often adopted workarounds like symlinks to keep these files in sync, as these documents can change frequently.
OpenAI contributed AGENTS.md to the Agentic AI Foundation last year, which is overseen by the Linux Foundation, in an effort to promote the standard. As of December 2025, over 60,000 open source projects had implemented AGENTS.md in their workflows.
For Anthropic, this move aligns with the company's desire to establish its own standards. However, the decision seems to recognize that the benefits of compatibility outweigh the advantages of maintaining separate specifications. Industry experts and OpenAI executives welcomed the collaboration, with OpenAI's head of core product, Thibault Sottiaux, expressing his support and Thariq Shihipar reciprocating with a handshake emoji, symbolizing a united approach.
Written by urgent.news from The Register Science's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
- Anthropic decides to support OpenAI's markdown instructions spec theregister.com
