Skip to main content
The System Flow for Product Engineering describes how data from Linear, GitHub, PagerDuty, and other sources is ingested, normalized, and projected into the domain. This flow ensures that every view in the workbench reflects current reality.
1

Intake

Connectors pull events from Linear, Jira, GitHub, Figma, PagerDuty, Sentry, and Datadog via webhooks and polling.
2

Load

Raw payloads are ingested into Spine Fabric. Each payload retains its native schema.
3

Normalize

Events are mapped to canonical entities: Issue, PullRequest, Incident, Release.
4

Schema Mapping

Field-level mappings align status, priority, and ownership across tools.
5

Context Assembly

Related entities are linked. A GitHub release connects to Linear issues and Datadog monitors.
6

Projection

The domain view is rendered: Sprint Board, Roadmap, Incident Queue, or Release Pipeline.
7

OODA

Twin and agents observe the projected state, orient, and propose actions.
8

Governance

Proposed actions are checked against authorization, scope, and policy.
9

Execution

Approved capabilities write back to source systems.
10

State Update

The updated state is re-ingested, and the cycle continues.