Skip to content
Agentflow

Product Design

product-design owns:

What must the product observably enable and guarantee?

Use when

Section titled “Use when”
  • the user problem is unclear
  • a feature idea needs refinement
  • behavior or scope is changing
  • acceptance outcome is ambiguous

Don’t use when

Section titled “Don’t use when”

The product behavior is already explicit and the unresolved question is architecture, interface mechanics, or implementation.

Model

Section titled “Model”
actor + trigger
   ↓
job / outcome
   ↓
observable behavior
   ↓
meaningful breaks
   ↓
constraints
   ↓
smallest complete scope
   ↓
acceptance proof

Input → output

Section titled “Input → output”
raw idea / product change / constraints
            ↓
       product-design
            ↓
job + behavior graph + scope + proof

Try it

Section titled “Try it”
Use $product-design to define recurring-task behavior.
Keep the scope to the smallest complete user loop.

Common handoff

Section titled “Common handoff”
product-design
  → design-graph when interface structure becomes non-trivial
  → engineering-design when system guarantees need architecture

Canonical runtime reference

Section titled “Canonical runtime reference”

Read the complete product-design instructions.