Skip to main content
A flow is a journey through the product, built from the claims your documentation makes. Flows lists every flow of every repository in the workspace.

The list

Search matches the title. Filter by Status, Driver and Repository. The filters ride the page address, so Home and the repository console can link straight to a narrowed list.

The five statuses

A flow’s page

Opening a row opens the flow. The header leads with the status, in the same word the list used. Beside it sit the markers a flow can wear: not in specs for a flow the documents no longer derive, dismissed, a tool-defect marker, epic for a flow that chains other flows, and manual for a hand-written test. Then the flow’s id, and a switch between the reading and the stored artifact, which is the test’s YAML when the flow has a test and the flow’s own entry when it has none. Under that come the flow’s title and its goal.

Milestones

The flow’s chain, always shown when the flow has one. Each milestone reads M1, M2 and so on, carries the claim sentence, and links to the section that states it. A milestone lists its cases, the situations that would prove it, only when it has more than one. With exactly one case, the case would repeat the claim almost word for word. A long list shows the first four and collapses the rest behind a count.

The test

Below the chain sits the test itself: its verdict, the filmstrip of screenshots for a browser run, the list of steps with an inspector beside it, and drawers for the transcript, the interfaces the test used and the rulings. Each step row carries an M2 chip beside its driver chip, naming the milestone that step proves, which is what ties a step back to the chain above.

When there is no test

Where the test would be, a flow with none shows one block: the status, and then why, in one sentence. Examples are that a required interface is not mapped, that a driver does not exist yet, or that authoring could not write a test and will try again on the next generation. When what is missing is something you can provide, the block instead names the service or the data and links straight to that dependency on the repository’s Dependencies tab. Blocks that would say the same reason twice are folded into one, because the milestones above already say which obligations are held up.

Stop testing a flow

Don’t test this flow sits in the rulings drawer, after the evidence it is made on. The next Flow generation drops the flow and deletes its tests. Un-dismiss puts it back.

Next steps

Flow generation

Where flows, milestones and cases come from.

Flow run

How the tests execute, and what evidence they leave.