create-epics

To translate approved GDDs into epics, you map the design and architecture to named epics for each architectural module, defining scope, governing ADRs, engine risk, and untraced requirements. This ensures governance and traceability before development begins.

Mapping architecture to epics involves targeting a specific layer or system—such as foundation, core, feature, or presentation—and generating a single epic per target module to define its boundaries and risks. This scoped workflow ensures focused module delivery.

To ensure traceability for untraced requirements and ADRs, each generated epic must explicitly list the governing ADRs, engine risks, and any technical requirements that remain untraced before the epic is finalized. This prevents hidden scope gaps during implementation.

You do not need to manually break epics into stories during translation. The process generates one epic per module, after which you run a dedicated command to create stories for each specific epic. This separates high-level module planning from detailed task breakdown.

A key limitation is that epic creation only defines the architectural module scope, ADRs, and risks; it does not generate actionable stories, requiring a separate step to break down each epic into implementation tasks. Scope is also restricted to one epic per target layer or system.