Domain Modeling

Actively sharpen the domain model: challenge terms, stress-test scenarios, write decisions as they crystallise. Reading CONTEXT.md for vocabulary is a one-line habit — this skill is for changing the model.

Where files live

Single context: CONTEXT.md + docs/adr/. Multi-context: CONTEXT-MAP.md points to per-area CONTEXT.md and ADRs. Create files lazily when you have something to write. Formats: [CONTEXT-FORMAT.md](./CONTEXT-FORMAT.md), [ADR-FORMAT.md](./ADR-FORMAT.md).

During the session

• Challenge glossary — call out terms that conflict with CONTEXT.md.
• Sharpen fuzzy language — propose precise canonical terms.
• Stress-test with scenarios — edge cases that force boundary decisions.
• Cross-check code — surface contradictions between stated rules and implementation.
• Cross-check with code graph — if graphify-out/graph.json exists, tell the user "用代码图谱验证模块边界假设" then query module boundaries and dependency density to validate concept boundary assumptions. Use /code-graph query modules for the overview and /code-graph query deps <module> to verify that proposed domain boundaries align with actual code boundaries.
• Update CONTEXT.md inline — one term at a time; glossary only, no implementation detail.

ADRs — offer only when all three hold

1. Hard to reverse
2. Surprising without context
3. Result of a real trade-off with alternatives

Otherwise skip the ADR.