skip to content
Skylar Payne

Topic hub

Agents

Agents are only useful if you can tell what they did and why. I care less about the magic loop and more about the boring parts: named states, allowed transitions, approval before risky actions, and traces you can read later.

Start here

  1. Break complex work into parts you can grade.
  2. Control the loop with state graphs.
  3. Put approval before risky tools.
  4. Record enough state to resume or replay failures.

Related writing

TILs