菜单
Build and sharpen a project's domain model. Use when the user wants to pin down domain terminology or a ubiquitous language, challenge or record a design decision, or when another skill needs to maintain the domain model.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Execute broad changes one coherent, independently verifiable stage at a time. Use when work spans multiple independently reviewable changes, or when contracts, producer-consumer migrations, generated artefacts, or release packaging create an ordered multi-stage rollout; skip small single-purpose changes.
Commit workflow using the dot git-commit gateway in the maintainer's concise one-line style. Use only after the user explicitly requests a commit or push, including /commit or /commit-push. Never infer repeat authorisation from an earlier commit or push; never run raw git commit.
Patterns for working with git branches, remotes, diffs against the default branch, and rebases. Use when resolving rebase conflicts, continuing interactive rebases, amending commits, or any git operation that would open an interactive editor.
Investigate a topic against primary sources and return cited findings, comparing credible maintainer and contributor perspectives when judgement is involved. Use when the user wants a topic researched, docs, API, or spec facts gathered, an external library or GitHub behaviour verified, competing views compared, or reading legwork delegated to a background agent.
Ask minimal clarifying questions only when ambiguity materially changes implementation. Use for routine underspecification; do not use for user-requested light or full grilling, plan stress-testing, or broad design interviews.
Stress-test a plan or proposed change through focused one-question-at-a-time grilling, with light and full intensity. Use when the user says grill, grill me, grill me lightly, ask me a couple of questions, stress-test this plan, or wants question-led scrutiny before planning.
| 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, challenge or record a design decision, or when another skill needs to maintain the domain model. |
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 decisions down the moment they crystallise. Merely reading existing docs for vocabulary is not this skill; that is a one-line habit any skill can do. This skill is for when you are changing the model, not just consuming it.
This repo does not prescribe a CONTEXT.md or docs/adr/ layout. Record terms and decisions wherever the project already keeps them: an existing glossary or domain doc, a README section, the notes vault, or a short comment next to the code. If nothing exists yet, only create a home for it when there is something durable to write and the user wants it kept. Do not stand up a glossary or decision-record convention the project has not asked for.
Keep a glossary free of implementation detail: it is a shared language, not a spec or a scratch pad. Keep a decision record focused on the decision and its trade-off, not the mechanics.
When the user uses a term that conflicts with the language already in use, call it out immediately. "You've been using 'cancellation' to mean X, but here you seem to mean Y. Which is it?"
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."
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.
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?"
When a term is resolved, write it down right there, wherever the model lives. Don't batch these up; capture them as they happen.
Only offer to write a decision record when all three are true:
If any of the three is missing, skip it. When you do record one, keep it to the decision, the alternatives considered, and why this one won.
