Home / Lessons / Assign the work / 8 of 27

One lead agent, and specialists

One agent owns the slice. The others get narrow tasks.

  1. One lead agent owns the slice.
  2. It hands out narrow, separate tasks.
  3. Each works in its own copy, so nobody overwrites anybody.
  4. One result comes back to the lead, and the lead answers for it.

One lead agent owns the slice. It hands narrow, independent tasks to builders, testers and reviewers. Each works in its own Git worktree, so two agents never overwrite each other.

Builders implement. Checkers verify. One integrator owns the final diff and the commit history. The builder never approves its own work.

The lead owns the outcome, specialists own their assignments, checkers produce independent evidence, and I decide whether the result passes a human gate.

  1. MeI decide whether the result passes a human gate.
  2. Lead agentOwns the slice and its outcome.
  3. BuilderA narrow task, in its own loop.
  4. TesterA narrow task, in its own loop. Produces independent evidence.
  5. Security reviewerA narrow task, in its own loop. Produces independent evidence.
Their work passes through automated checks into an evidence pack, and stops at a gate only I can open.

How I do it

The number of agents matters less than how clear their assignments, shared state and authority are.

The rule

Give every agent one task and one owner, and keep the one who built it away from the one who approves it.

The skill I use for this: /subagent-driven-development. Made by Jesse Vincent.

Read next

The ideas behind it

This lesson as plain text, for your agent. Or all 27 on one page.

Try “secrets”, “merge”, “handover” or “database”.