object-sketch

To extract an object model from domain logic, the Skill sketches structured Module blocks containing Extract, Domain-logic, and Concept sections from upstream context like extracted-domain-logic.md. It grounds the initial object model by citing source passages for traceability.

The best way to structure module partitioning from scattered concepts is to map them into a canonical object-sketch.md template. This generates readable Module blocks and supports parked [Unallocated] or [Rejected] modules when concepts lack immediate placement.

You do not strictly need upstream domain files, but the Skill operates optimally when sources such as extracted-domain-logic.md or key-abstractions.md exist. These files provide the seed sourcing rules needed to ground the module completeness.

You map key abstractions to specific domain modules by running the object-sketch agent on engagement context. It processes key-abstractions.md to produce structured Module blocks with traceable Extract evidence linked to the original source passages.

Yes, you can park unallocated or rejected concepts during object modeling. The Skill explicitly supports creating [Unallocated] and [Rejected] modules to maintain structural completeness without forcing concepts into incorrect domain logic partitions.