Executable Product Layer

The Executable Product Layer adds a product-first entry point to Asagiri without turning it into a visual app builder. The goal is to keep every artifact local, inspectable, and versioned in the repository.

Pipeline

Intent
  -> prototype
  -> flows
  -> contracts
  -> generated specs
  -> executable tasks
  -> work plan

Repository structure

For a product named workspace-saas, Asagiri stores artifacts under:

.asagiri/products/workspace-saas/
  intent.md
  product.yaml
  prototype/
  flows/
  screens/
  contracts/
  extraction/
  generated-specs/

It also writes:

.asagiri/specs/workspace-saas/
.asagiri/tasks/workspace-saas/

Why this exists

• Keep intent-to-code traceability explicit.
• Generate implementation-ready tasks from user flows.
• Make contracts and risks reviewable before coding.
• Keep deterministic, local-first defaults.

Boundaries

This layer is not a replacement for architecture review, human validation, or backend implementation. The prototype is a working simulation, not production code.

Related

• Prototype as executable spec
• Workflow: product to code
• CLI: prototype