Multi-agent system

Instead of one agent doing everything, a multi-agent system splits work across several. One might plan, others research or write in parallel, and another reviews the result. Each agent has a focused role and a smaller, cleaner context.

The benefits are parallelism, specialization, and keeping any one context window from overflowing. The costs are real too: more tokens, more moving parts, and new failure modes where agents miscommunicate, duplicate work, or lose information at handoffs.

A good rule is to start with a single agent and add more only when you hit a clear limit, such as a task that naturally splits into independent parts or a context that keeps overflowing.