iurysza Domain Modeling
Actively build and sharpen the project's domain model as you design. This is the active discipline—challenging terms, inventing edge-case scenarios, visualizing relationships, and writing the glossary and decisions down when they crystallize. Merely reading ai-artifacts/CONTEXT.md for vocabulary is not this skill; this skill is for changing the model, not just consuming it.
Artifact home
Keep agent-authored project knowledge under ai-artifacts/. This includes context glossaries, context maps, ADRs, plans, specs, analyses, reports, reviews, and other Markdown primarily created for or by an LLM.
Use another location only when the user or repository explicitly requires it, such as a human-facing README.md or established product documentation. Do not relocate existing documentation unless asked.
Embed ASCII or Mermaid diagrams directly in the Markdown they explain. Do not create standalone diagram files.
File structure
Most repositories have a single context:
/
├── ai-artifacts/
│ ├── CONTEXT.md
│ └── docs/
│ └── adr/
│ ├── 0001-event-sourced-orders.md
│ └── 0002-postgres-for-write-model.md
└── src/
If ai-artifacts/CONTEXT-MAP.md exists, the repository has multiple contexts:
/
├── ai-artifacts/
│ ├── CONTEXT-MAP.md
│ ├── docs/
│ │ └── adr/ ← system-wide decisions
│ └── src/
│ ├── ordering/
│ │ ├── CONTEXT.md
│ │ └── docs/adr/ ← context-specific decisions
│ └── billing/
│ ├── CONTEXT.md
│ └── docs/adr/
└── src/ ← application source code
ai-artifacts/src/ mirrors the source tree only to make context ownership obvious. Application code remains in the repository's real source directories.
Create files lazily—only when there is something worth recording. Create ai-artifacts/CONTEXT.md when the first term is resolved. Create an ADR directory when the first qualifying decision is made. Create ai-artifacts/CONTEXT-MAP.md only when the project has multiple genuine contexts.
During the session
Challenge against the glossary
When the user uses a term that conflicts with the relevant CONTEXT.md, call it out immediately. "Your glossary defines 'cancellation' as X, but you seem to mean Y—which is it?"
Sharpen fuzzy language
When the user uses vague or overloaded terms, propose a precise canonical term. "You're saying 'account'—do you mean the Customer or the User? Those are different things."
Discuss concrete scenarios
When domain relationships are being discussed, stress-test them with specific scenarios. Invent scenarios that probe edge cases and force the user to be precise about boundaries between concepts.
Cross-reference with code
When the user states how something works, check whether the code agrees. If you find a contradiction, surface it: "Your code cancels entire Orders, but you just said partial cancellation is possible—which is right?"
Visualize the model
Include an embedded diagram when boundaries, relationships, states, or flows are easier to understand visually.
- Use Mermaid for graphs, state transitions, sequences, and multi-context relationships.
- Use ASCII art for small structures or when plain-text portability matters more.
- Use domain terms from the glossary.
- Show the domain model, not classes, framework components, or incidental infrastructure.
- Keep the diagram beside the prose it clarifies.
Update context artifacts inline
When a term or relationship is resolved, update the relevant file under ai-artifacts/ immediately. Do not batch resolved domain knowledge. Use CONTEXT-FORMAT.md.
A CONTEXT.md is a glossary, not a spec, scratch pad, or implementation guide. Its diagrams may show conceptual relationships, boundaries, and states, but must not contain implementation details.
Offer ADRs sparingly
Only offer to create an ADR when all three are true:
- Hard to reverse—the cost of changing the decision later is meaningful.
- Surprising without context—a future reader will wonder why it was done this way.
- The result of a real trade-off—genuine alternatives existed and one was chosen for specific reasons.
If any condition is missing, skip the ADR. Use ADR-FORMAT.md.