/deep-research is the quickest way to see a dynamic workflow
Dynamic workflows make more sense once you have watched one run. Claude Code ships with /deep-research: it fans out searches, cross-checks sources, and writes a cited report.
Run the bundled one first
Dynamic workflows click once you have watched one run, and Claude Code ships with a good one. To investigate a question across many sources:
/deep-research What changed in the Node.js permission model between v20 and v22?
What it does
It fans web searches out across several angles, fetches and cross-checks the sources it finds, votes on each claim, and synthesises a cited report — with the claims that did not survive cross-checking filtered out. That last part is the bit worth internalising: it is not one model summarising search hits, it is independent agents checking each other before anything is reported.
Watch it work
The run starts in the background. /workflows lists runs; select one and press Enter to see each phase with its agent count, token total, and elapsed time, and drill into any agent to read its prompt and result. Watching the agent count climb is the clearest picture you will get of what "orchestration moved into code" actually means.
Then go deeper
Once /deep-research makes sense, the dynamic-workflows tutorial covers when to reach for a workflow over subagents or skills, and how to have Claude write and save one for your own task.