Ask HN: Is it better to have no Agent.md than a bad one?

Please share your real word experiences. What is a bad one and why?

5 points | by parvardegr 1 day ago

4 comments

  • allinonetools_ 9 hours ago
    In my experience, a vague or outdated Agent.md causes more damage than not having one, because people assume it is accurate and stop asking questions. A simple, honest doc that is kept current is far more useful than a detailed one nobody maintains.
  • nullcathedral 1 day ago
    One approach (Claude Code) is to evolve it over time. Start small and run /insights often and use that to refine the CLAUDE.md as needed.

    https://github.com/trailofbits/claude-code-config?tab=readme...

  • e5bed97b3349 18 hours ago
    Yes, zero instructions are better than wrong instructions because outdated or incorrect context actively forces the AI to write broken code.

    Start with nothing, and only add a single sentence to your Agent.md after the AI repeatedly makes the same specific mistake.

    • DANmode 18 hours ago
      Without further instruction, it falls back to the default logic.
  • verdverm 1 day ago
    My approach is to keep them minimal. Make them a table of contents and spark notes, spread them across the repo, keep them up to date as the repo changes.

    Something is better than nothing. It's pretty hard to have bad ones, also takes effort to craft good ones. The bad ones are the ones with incorrect information.

    • parvardegr 18 hours ago
      Hmm sounds reasonable
      • verdverm 8 hours ago
        If paired this with system prompts derived from Claude Code. This makes gemini-flash nearly on par with the big boy models. Saves a ton of time and money