‹ BackHN Continuity

Thread

Agents.md speaks Unix, and you should too

16 points · 9 comments · fmind-dev

  1. axeldunkel · · focus · HN ↗
    Lets assume this real - for humans these condensed tools make sense. For an agent I'd argue it takes him more tokens to understand these condensed infos since he most probably knows the right tools anyways and does not need the shortcuts.
    1. fmind-dev · · focus · HN ↗
      Indeed, you need to invest some tokens in configuring the tools, configs and skills. But once it's done, tools and configs do not consume more tokens, as they are automatically used by your AI Agents. Moreover, skills support progressive disclosure, and I try to keep a budget of max 5000 tokens for global + 5000 tokens for local setups (= Skills + AGENTS.md).

      But yes, I burned too many tokens on this project. That's why I'll be happy if it can help others.

Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.