You Are Part of the Harness: Building a 100+ Agent Swarm in Web3 (Part 4)
Part 4 of my HackerNoon series on building a 100+ agent swarm is live. Part 1 covered context management. Part 2 covered skill engineering. Part 3 covered prohibitions and deterministic enforcement. This one is about the constraint I didn’t expect: me.
AI agents amplify your habits. All of them.
I skipped planning for twenty years and got away with it. I believed in DRY but never enforced it with tooling. I knew principles mattered but never wrote them down. These were bad habits I could tolerate when I was the only one writing code. With 100+ agents, those habits scaled too. And at scale, they broke everything.
The article covers:
- Planning separation — separating the spec from execution, because agents execute unplanned code faster than you ever could
- Principle codification — writing down the rules you carry in your head, because agents cannot infer what you never stated
- DRYP enforcement — applying Don’t Repeat Your Prompt to keep agent instructions composable instead of contradictory
The biggest improvement to my agent system wasn’t a better model or a new tool. It was curbing my own bad habits. The harness was fine. The person holding it was the bottleneck.