Chapter 21 · CLI
Dynamic workflows, batch and ultracode
Dynamic workflows orchestrate subagents at scale: the session spawns and manages agents as needed instead of following a rigid plan. This approach fits large tasks whose structure emerges along the way.
Video
Steps
- Understand the difference between a dynamic workflow and a fixed plan: the tool decides at runtime how many agents and for what.
- Batch tasks repeat the same action over many items; define one template and run it at scale.
- Set a clear success criterion per agent so the orchestration knows when a subtask finished correctly.
- Track cost and time: many parallel agents consume a lot, and sometimes simple is better.
Common pitfall: Wide orchestration without a stop criterion can spin: agents keep generating work for each other with no one saying enough.
Hands-on exercise: Take a real batch task (updating docs across many files, for example) and run it through a workflow. Record cost, time and quality against doing it by hand.
Filmed demo · Video is added in the media phase
Recap
Comprehension check
When is a dynamic workflow better than a fixed plan?
Recap
- Dynamic orchestration for emergent structure
- Batch: one template over many items
- A clear stop criterion per agent