Thread

Claude Code now reads AGENTS.md if there is no Claude.md

741 points · 285 comments · datadrivenangel

askonomm · 1789765915 · focus · HN ↗
Finally doing something standards-compliant instead of forcing users into a proprietary workflow.
1. swyx · 1789769164 · focus · HN ↗
  fwiw, i am with thariq <a href="https://x.com/trq212/status/2092302273099796842" rel="nofollow">https://x.com/trq212/status/2092302273099796842 in that prompts should be tuned for models and in fact blindly applying agents.md is probably an antipattern unless you want all models to basically converge to some common ill defined of instruction following - good local minima, bad global minima for model diversity and exploration of intelligence.
  aka, sometimes it really is too early to force a standard
  1. orlp · 1789770891 · focus · HN ↗
    If you want this it's trivial to add an AGENTS.md that simply says "if you're Claude read CLAUDE.md, if you're Astra read ASTRA.md". A common entry point is good regardless.
    jwolfe · 1789771100 · focus · HN ↗
    This wastes both tokens and turns. But yes it's probably the best option we have today.
    recursivegirth · 1789771461 · focus · HN ↗
    Wasting turns? That is silly, use a better harness. Also token usage can mitigated by incremental discovery instead of stuck 5k+ worth of tokens in the AGENT/Claude md file.
    jwolfe · 1789772131 · focus · HN ↗
    Every turn means more tokens in ways that are not obvious to most people and lead to tons of unnecessary cache reads.
    No harness can batch your agents.md read with the reads the contents of the file tell it to read.
    pishpash · 1789772149 · focus · HN ↗
    It can try its own file and fall back to generic like here. What's wrong with that?

Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.