domain-modeling-skill

What I do

I actively build and sharpen the project's domain model as you design. This is the active discipline — challenging terms, inventing edge-case scenarios, and writing the glossary and decisions down the moment they crystallise.

Merely reading CONTEXT.md for vocabulary is NOT this skill — that's a one-line habit any skill can do. I am for when you're changing the model, not just consuming it.

1. Challenge terms against the existing glossary and surface contradictions immediately
2. Sharpen fuzzy language by proposing precise canonical terms
3. Stress-test with scenarios that probe edge cases and force precision about concept boundaries
4. Cross-reference with code — when the user states how something works, verify the code agrees
5. Update CONTEXT.md inline — capture resolved terms immediately, never batch them
6. Offer ADRs sparingly — only when a decision is hard-to-reverse, surprising, and a real trade-off

I am the docs engine (model-invoked). grilling-skill --docs mode inlines a compact copy of my capture convention (glossary + three-criteria ADR gate) so the interview captures docs directly; I remain the canonical full engine.

When to use me

Use this skill when:

• The user introduces or disputes domain terminology
• A ubiquitous language needs to be established or sharpened
• An architectural decision has just been made and may warrant an ADR
• Another skill (e.g. grilling-skill --docs mode) captures docs inline during a session
• Code contradicts the user's stated understanding of how the system works

Trigger phrases:

• "let's define our terms"
• "that should be an ADR"
• "update CONTEXT.md"
• "is that the right word for"
• "capture this decision"
• "ubiquitous language"

File Structure

Single-context repos (most repos)

/
├── CONTEXT.md
├── docs/
│   └── adr/
│       ├── 0001-event-sourced-orders.md
│       └── 0002-postgres-for-write-model.md
└── src/

Multi-context repos

If a CONTEXT-MAP.md exists at the root, the repo has multiple contexts. The map points to where each one lives:

/
├── CONTEXT-MAP.md
├── docs/
│   └── adr/                          # system-wide decisions
├── src/
│   ├── ordering/
│   │   ├── CONTEXT.md
│   │   └── docs/adr/                 # context-specific decisions
│   └── billing/
│       ├── CONTEXT.md
│       └── docs/adr/

Create files lazily

Only create files when you have something to write:

• No CONTEXT.md → create one when the first term is resolved
• No docs/adr/ → create it when the first ADR is needed

Inference rules:

• If CONTEXT-MAP.md exists → read it to find contexts
• If only a root CONTEXT.md exists → single context
• If neither exists → create a root CONTEXT.md lazily when the first term resolves

When multiple contexts exist, infer which one the current topic relates to. If unclear, ask.

Core Workflow

Step 1: Challenge against the glossary

When the user uses a term that conflicts with the existing language in CONTEXT.md, call it out immediately:

"Your glossary defines 'cancellation' as X, but you seem to mean Y — which is it?"

Step 2: 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."

Step 3: 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 the boundaries between concepts.

Step 4: 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?"

Step 5: Update CONTEXT.md inline

When a term is resolved, update CONTEXT.md right there. Do not batch these up — capture them as they happen. Use the CONTEXT.md Format below.

CONTEXT.md must be totally devoid of implementation details. Do not treat it as a spec, a scratch pad, or a repository for implementation decisions. It is a glossary and nothing else.

Step 6: Offer ADRs sparingly

Only offer to create an ADR when all three are true:

1. Hard to reverse — the cost of changing your mind later is meaningful
2. Surprising without context — a future reader will wonder "why did they do it this way?"
3. The result of a real trade-off — there were genuine alternatives and you picked one for specific reasons

If any of the three is missing, skip the ADR. Use the ADR Format below.

CONTEXT.md Format

Structure

# {Context Name}

{One or two sentence description of what this context is and why it exists.}

## Language

**Order**:
{A one or two sentence description of the term}
_Avoid_: Purchase, transaction

**Invoice**:
A request for payment sent to a customer after delivery.
_Avoid_: Bill, payment request

**Customer**:
A person or organization that places orders.
_Avoid_: Client, buyer, account

Rules

• Be opinionated. When multiple words exist for the same concept, pick the best one and list the others under _Avoid_.
• Keep definitions tight. One or two sentences max. Define what it IS, not what it does.
• Only include project-specific terms. General programming concepts (timeouts, error types, utility patterns) don't belong. Before adding a term, ask: is this a concept unique to this context, or a general programming concept? Only the former belongs.
• Group terms under subheadings when natural clusters emerge. If all terms belong to a single cohesive area, a flat list is fine.

Multi-context map

# Context Map

## Contexts

- [Ordering](./src/ordering/CONTEXT.md) — receives and tracks customer orders
- [Billing](./src/billing/CONTEXT.md) — generates invoices and processes payments

## Relationships

- **Ordering → Billing**: Ordering emits `OrderPlaced` events; Billing consumes them to generate invoices
- **Ordering ↔ Billing**: Shared types for `CustomerId` and `Money`

ADR Format

ADRs live in docs/adr/ and use sequential numbering: 0001-slug.md, 0002-slug.md, etc. Create the directory lazily.

Template

# {Short title of the decision}

{1-3 sentences: what's the context, what did we decide, and why.}

That's it. An ADR can be a single paragraph. The value is in recording that a decision was made and why — not in filling out sections.

Optional sections

Only include these when they add genuine value. Most ADRs won't need them:

• Status frontmatter (proposed | accepted | deprecated | superseded by ADR-NNNN) — useful when decisions are revisited
• Considered Options — only when the rejected alternatives are worth remembering
• Consequences — only when non-obvious downstream effects need to be called out

Numbering

Scan docs/adr/ for the highest existing number and increment by one.

What qualifies for an ADR

All three of these must be true (hard to reverse + surprising + real trade-off). If a decision is easy to reverse, skip it — you'll just reverse it. If it's not surprising, nobody will wonder why. If there was no real alternative, there's nothing to record.

Qualifies	Example
Architectural shape	"We're using a monorepo." "The write model is event-sourced, the read model is projected into Postgres."
Integration patterns between contexts	"Ordering and Billing communicate via domain events, not synchronous HTTP."
Technology choices that carry lock-in	Database, message bus, auth provider, deployment target — not every library, just ones that would take a quarter to swap out
Boundary and scope decisions	"Customer data is owned by the Customer context; others reference it by ID only." The explicit no-s are as valuable as the yes-s.
Deliberate deviations from the obvious path	"We're using manual SQL instead of an ORM because X." Stops the next engineer from "fixing" something that was deliberate.
Constraints not visible in code	"We can't use AWS because of compliance." "Response times must be under 200ms because of the partner API contract."
Rejected alternatives when non-obvious	If you considered GraphQL and picked REST for subtle reasons, record it — otherwise someone will suggest GraphQL again in six months.

Rules

• Never put implementation details in CONTEXT.md. It is a glossary, not a spec.
• Capture inline, not in batches. Resolved terms and decisions are written the moment they crystallise.
• Be opinionated about language. Pick the best word; list the others under _Avoid_.
• Don't create ADRs reflexively. The three-criteria gate exists to keep signal high.
• Verify claims against code. The user's mental model and the code can drift; your job is to catch it.

Integration with Other Skills

Skill	Integration
grilling-skill	The interview engine whose resolved terms I write down; its --docs mode inlines a compact copy of my capture convention
continuous-learning-skill	ADRs and glossary entries feed into long-term learning storage
architecture-review-subagent	Review findings may surface decisions worth recording as ADRs

Example Usage

During a grilling session

User: "When an order is cancelled, we refund the original payment method."
Agent: "Quick glossary check — your CONTEXT.md defines 'cancellation' as voiding before fulfillment.
        You just described refunding a *delivered* order, which sounds different. Is that
        'return'? I'll add both terms if they're distinct."

The skill will:

1. Cross-reference the claim against the existing CONTEXT.md
2. Surface the contradiction
3. Propose precise terms (cancellation vs return)
4. Update CONTEXT.md inline once resolved

Offer an ADR

Agent: "We've decided ordering and billing communicate via domain events rather than
        synchronous HTTP. That's hard to reverse, surprising (most teams default to HTTP),
        and a real trade-off (eventual consistency). Want me to record it as ADR-0003?"

References

• grilling-skill --docs - User-facing mode that inlines a compact copy of my capture convention
• grilling-skill - The interview engine I capture docs from
• continuous-learning-skill - Long-term storage for decisions