Agent setup fix call

You have a specific problem. The agent keeps looping, can't find context, or tool permissions are broken. You've tried a few things and nothing stuck. This is a 30-minute call where we find it and fix it.

How it works

  1. 1.

    You share your setup beforehand: AGENTS.md, tool config, the error or behavior you're seeing. It doesn't need to be polished, just enough context to look at before we start.

  2. 2.

    We spend the call on it. I ask questions, we reproduce the problem if needed, and fix it live where we can.

  3. 3.

    You leave with a short list of concrete changes. Anything we didn't finish live, I write up so you can finish it yourself.

Good fit if

  • Tool permissions are failing or too permissive and you can't figure out why.
  • The agent keeps running commands that no longer exist in the codebase.
  • You're getting inconsistent behavior you can't reliably reproduce.
  • You want a second pair of eyes before committing to a bigger setup change.

Not a good fit if

  • You're starting from scratch and need a full setup (the starter pack or async audit makes more sense).
  • You have several unrelated problems. 30 minutes isn't enough to address all of them properly.