ship-design
Use when a brainstormed feature needs architectural decisions — presents 2-3 approaches for user to choose before planning
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
メニュー
Use when a brainstormed feature needs architectural decisions — presents 2-3 approaches for user to choose before planning
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Use when a feature plan has been verified and is ready for implementation — executes tasks with atomic commits
Use when you want to auto-run all remaining Ship steps for a feature without manual step-by-step invocation
Use when a feature has been brainstormed and needs a detailed implementation plan with tasks
Use when a plan has been created and needs independent verification against codebase patterns before building
Use when beginning new feature work, adding functionality, or fixing a bug — runs intensive brainstorming to capture requirements before planning
Use when the user asks about Ship commands, workflow, or how to use the framework
| name | ship:design |
| description | Use when a brainstormed feature needs architectural decisions — presents 2-3 approaches for user to choose before planning |
| effort | medium |
| allowed-tools | Read, Write, Edit, Agent, Glob, Grep, AskUserQuestion |
| argument-hint | [feature-name] |
Present architecture options for the active feature so the user can choose an approach before planning.
Feature state is injected by hooks at session start and after compaction — check conversation context for "SHIP ACTIVE FEATURES" or "SHIP FEATURE STATE" blocks first.
$ARGUMENTS is provided, use it as the feature namebrainstormed or planned (replanning with architecture input).planning/features/*/CONTEXT.mdIdentify the 2-3 genuinely distinct viable approaches for THIS feature. They come from the feature's actual decision axes — e.g. extend an existing module vs. build a new one, sync vs. async, buy (use a library) vs. build — not from a fixed menu of philosophies. The count and framing come from the feature: if only two real contenders exist, present two; do not invent a third to fill a slot.
.planning/features/{name}/CONTEXT.md for the requirements and constraints.For small or familiar surfaces, explore inline with Glob/Read/Grep and develop the approaches yourself. For larger or unfamiliar surfaces, launch one parallel sub-agent per candidate approach via the Agent tool — each explores the codebase and reports back in this format:
## Approach: {approach name}
**Summary:** [2-3 sentences]
**Key implementation points:**
- [concrete point, naming key files]
- [concrete point, naming key files]
**Tradeoffs:** [what this buys and costs vs. the alternatives]
**Estimated tasks:** [N]
Once every candidate approach is concrete, present a comparison to the user.
Display a summary of each approach, then form your own recommendation based on the feature's complexity, the codebase's current state, and the trade-offs.
Use AskUserQuestion to ask: "Which architecture approach should we use for this feature?"
After the user chooses:
.planning/features/{name}/CONTEXT.md## Chosen Architecture section:## Chosen Architecture
**Approach:** {chosen approach name}
**Summary:** {summary from the chosen approach}
**Key implementation points:**
{points from the chosen approach}
## DESIGN COMPLETE
Feature: {name}
Chosen: {approach name}
The planner will use this architecture to guide task design.
Next: /ship:plan
$ARGUMENTS