Working with AI agents in a team

Problem

An AI agent that writes code is fast, but it is just as fast at doing something you do not want: committing without being asked, deleting files or following its own style. In a team there is a second problem. Everyone makes their own agreements with the agent, so the result differs per developer.

What I did

• Shared rules. The team's agreements live in one repository that every developer loads. Personal preferences are kept apart and win when they conflict.
• A hook as a safety net. An agent can ignore a rule written in text. So a hook blocks dangerous git commands before they run.
• A model per task. Mechanical work goes to a small model, normal research to a mid-size model, and only choices where judgment decides the outcome go to the largest model.
• Several issues in parallel. One session divides the work and starts a separate agent per issue.
• A fixed verification step. An agent's work is only done once it has been checked: lint, the full test suite and a check of the diff against the rules.

TODO: add an example of what went wrong and which rule was added afterwards.

What I measured

TODO: there are no numbers here yet. Add something verifiable, or drop this heading.

What I learned

• A rule that only exists as text is a request. What must never happen belongs in a hook.
• Rules grow out of mistakes. Every correction gets a fixed place, so the same mistake does not come back.
• Working in parallel only saves time if the check at the end is as strict as it is for a single issue.