AGENT SKILL

domain-modeling

Build and sharpen a project's domain model. Use when the user wants to pin down domain terminology or a ubiquitous language, record an architectural decision, or when another skill needs to maintain the domain model.

License 许可
GPL-3.0
最近更新
2026-11-10
Stars
★ 0
Forks
0

01 前言 Front Matter

name
domain-modeling
description
Build and sharpen a project's domain model. Use when the user wants to pin down domain terminology or a ubiquitous language, record an architectural decision, or when another skill needs to maintain the domain model.

02 资源文件 2

03 SKILL.md 正文

Domain Modeling

Build the model actively: challenge and sharpen domain language, test boundaries, and record terms or durable decisions when they resolve. Merely consuming an existing glossary does not invoke this skill.

Route the domain artifacts

  1. If docs/agents/domain.md exists, read and follow it; it owns repository-specific context and ADR routing.
  2. Check that optional CONTEXT.md, CONTEXT-MAP.md, and ADR paths exist before opening them. Treat absence silently and use the defaults in the format references.
  3. Before writing, follow the repository entrypoint and active-doc registry for any required index updates; include those updates in the same change.

Create domain files lazily. Read CONTEXT-FORMAT.md before adding or editing a glossary entry, and read ADR-FORMAT.md before offering or writing an ADR.

Model during the session

  • Challenge the glossary. Surface conflicts between the user's language and existing definitions.
  • Sharpen fuzzy terms. Propose a precise canonical term when language is vague or overloaded.
  • Test concrete scenarios. Probe edge cases that clarify relationships and boundaries.
  • Check reality. Compare claims with code and governing documents; surface contradictions rather than recording them as fact.
  • Check agent-authored material too. Apply the same scrutiny to autonomous recaps, ledgers, and determinations.

Record resolved changes

When a term resolves, update the relevant CONTEXT.md inline using the format reference. If a caller's mutation boundary is still unresolved, retain the exact pending write for its checkpoint instead of editing early.

Separate terms from decisions before writing:

  • Put only a domain or app-layer concept's tight definition in CONTEXT.md.
  • Put ownership, alternatives, trade-offs, consequences, and implementation shape in the appropriate ADR, spec, issue, or design note.

For an architectural decision, apply ADR-FORMAT.md's qualification and ownership rules. Skip an ADR when the decision does not qualify or an authoritative artifact already owns its decision record.

Supporting-role closeout

When another skill invoked this one, review the session's current decisions once before closeout.

  • Settled facts and user-ratified decisions with no domain change: Domain model unchanged — no new app-layer terms, no ADR-worthy decisions.
  • Any provisional decision with no domain change: Domain model unchanged — contingent on ratification of the provisional decisions.
  • A changed model: list the exact CONTEXT.md term changes and every ADR created or offered, including written paths.

Include the applicable result in the caller's recap or pre-deliverable checkpoint.

04 同仓库更多 Skills