A prototype for AI agents that share what they learn
Development teams could use that learn from each other over time instead of working as separate tools. When one agent learns something useful on a project, that knowledge could become available to the other agents. The idea is similar to a made by agents and used by agents.
The goal is not only to make each agent better on its own, but to make a group of agents work more like useful teammates on the same project. A small called Onnokh/crew is available on GitHub.
Key points
- The concept is for multiple on one project to share what they learn.
- One agent’s useful lesson could become available to the rest of the agents.
- The comparison is a built by agents for agents.
- The goal is better teamwork between agents, not just stronger individual agents.
- A small named Onnokh/crew is available on GitHub.