Principles for agent-native CLIs

(twitter.com)

21 points | by blumpy22 2 hours ago

2 comments

  • wolttam 35 minutes ago
    Getting agents used to using `--force` to bypass prompts seems like a bad idea. `--force` is for when the action failed (or would fail) for some reason and you want it to definitely happen this time.

    I think `--yes` or `--yes-do-the-dangerous-thing` is leagues better.

    • tekacs 14 minutes ago
      It also in the case of an LLM can bias it towards using that sort of flag more commonly, which is less than ideal when it then uses a more ordinary Unix command that uses that to mean something dangerous.
    • ihsw 17 minutes ago
      `--non-interactive` has precedent too.
  • sandermvanvliet 7 minutes ago
    Is it me or are all these articles about using AI effectively and building for AI just, you know, things that we should have been doing all along?

    It feels like most of the “rules” are “don’t be an ass to your consumer”.