GitHub’s Changelog for October 1, 2026 says dynamic workflows are available in public preview in GitHub Copilot CLI, the GitHub Copilot app, and the GitHub Copilot SDK (Changelog).
A dynamic workflow is a program that defines how a task is carried out—automated steps plus one or more agents, running one after another, in parallel, or both. It lives inside a GitHub Copilot extension. Preview status can change; treat this as early access, not a finished generally available product.
What you can orchestrate (as GitHub lists)
GitHub names these capabilities on the Changelog:
- Run commands, use tools, or call other services
- Divide a goal and run independent tasks in parallel
- Pass structured results from one stage to the next
- Have subagents verify each other’s findings
- Ask you for input, if your client supports it
- Pause at a checkpoint so you can review results and resume when ready
You can author workflows yourself or have Copilot write them; Copilot includes built-in authoring guidance. Dynamic workflows are available on all Copilot plans—GitHub publishes no dollar prices in this Changelog.
Unlike /fleet
Unlike /fleet, where Copilot delegates work to subagents and coordinates them in parallel, a dynamic workflow carries out a process defined in code (Changelog). The two are different.
How to turn it on
- Copilot app: Dynamic workflows are always available with no setup required.
- Copilot CLI: In the latest Copilot CLI, enable experimental features with the
--experimentalcommand-line option or/experimental onin an interactive session; update with/update.
Ask “What dynamic workflows are available?” to discover what’s on your setup.
Not computer use, not cloud remote control
Copilot computer use is a separate public preview—local desktop GUI control on macOS and Windows. Dynamic workflows are coded multi-agent orchestration in CLI, app, and SDK. This is also not a Manus-style remote or cloud computer you drive from another device.
The Changelog’s “when to use” bullets (incident timelines, release checks, parallel PR review, dual-model comment checks, codebase sweeps, research→plan→change, long pauseable runs) are illustrative vendor examples, not verified customer case studies or performance guarantees.
Who should care
Developers who want repeatable, code-defined multi-agent pipelines—with structured handoffs, verification, and pause-resume—should start at the Oct 1 Changelog. Keep public preview, use the CLI experimental path if you’re on CLI, and treat example scenarios as illustrations only.

The Campfire
CommentsThanks. Your comment is waiting for the desk. Once it is approved it appears here within about a minute.
Add a comment