← Back to context

Comment by goalieca

2 days ago

> I add to my team’s CLAUDE.md multiple times a week.

This concerns me because fighting tooling is not a positive thing. It’s very negative and indicates how immature everything is.

The Claude MD is like the documentation you hand to a new engineer on your team that explains details about your code that they wouldn't otherwise know. It's not bad to need one.

  • But that documentation shouldn’t need to be updated nearly every other day.

    • Consider that every time you start a session with Claude Code. It's effectively a new engineer. The system doesn't learn like a real person does, so for it to improve over time you need to manually record the insights that for a normal human would be integrated by the natural learning process.

      4 replies →

    • I certainly could be updating the documentation for new devs very frequently - the problem with devs is that they don't bother reading the documentation.

      1 reply →

    • If you are consistent with how you do your projects you shouldn't need to update CLAUDE.md nearly every day. Early on, I was adjusting it nearly every day for maybe a couple of projects but now I have very little need to make any adjustments.

      Often the challenge is users aren't interacting with Claude Code about their rules file. If Claude Code doesn't seem to be working with you ask it why it ignore a rule. Often times it provides very useful feedback to adjust the rules and no longer violate them.

      Another piece of advice I can give is to clear your context window often! Early in my start in this I was letting the context window auto compact but this is bad! Your model is it's freshest and "smartest" when it has a fresh context window.

      1 reply →

    • Have you never looked at your work's Confluence? Worse, have you never spent time at a company where the documentation wasn't frequently updated?

      1 reply →

You might be misunderstanding what a CLAUDE.md is. It’s not about fighting the model, rather it’s giving the model a shortcut to get the context it needs to do its work. You don’t have to have one. Ours is 100% written by Claude itself.

  • That's not the same thing as adding rules by yourself based on your experiences with Claude.