Should you have an AGENTS.md?

Published on 24 February 2026 • Estimated reading time: 2 minutes

Small paper on the usefulness of AGENTS.md files: https://arxiv.org/abs/2602.11988. Findings boil down to:

1. It costs more to have them—makes perfect sense, since we’re putting more into context.
2. Human-curated versions of AGENTS.md outperform auto-generated ones—again, makes sense. The AI can’t always know what’s useful; it’s just summarizing and guessing.
3. The paper tested only solving tasks, and results vary based on the model, but seldom is there a big swing either way for it being good or bad—this is kinda surprising to me.
4. Testing is improved by having AGENTS.md—makes sense, we’re telling the agent what to do each time. AGENTS.md is augmenting our prompt.
5. Not in this paper, but a prior work cited here tested security and saw improvements from AGENTS.md again—makes sense; it’s augmenting the prompt with security details.

Ultimately, my takeaway is that AGENTS.md is not magic (no one should be surprised by that)—all the AGENTS.md is doing is just augmenting your prompt with more info. More info costs money with AIs, and more info can help or hinder if it’s good or bad info. Like anything you use, know how the tool works to get the best results and do not blindly trust AIs (even /init is an AI).

Where do I land after reading this? Keep it—not because your individual LLM use is better or worse, but as a team, you commit the AGENT.md to your repo and your team’s experience with AI tools becomes more consistent. Also, I’m way too old and forget things, so having an AGENTS.md with things I should tell the agent—rather than the individual prompt—costs less than running the same thing three times and paying a lot more in totality because I forgot something.