菜单
Use this skill when the user has a rough idea, wants to explore a feature, or says "let's brainstorm" or "help me design." The goal is to refine an idea into a validated design doc before any code is written.
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
Use this skill when ending a session and you want the next session to pick up where you left off. The goal is to capture enough context for seamless session continuity.
Use this skill when starting feature work that needs isolation from the current workspace, or when you need to work on multiple branches simultaneously. The goal is to create an isolated worktree with a verified clean baseline.
Use this skill when fixing a bug caused by invalid data reaching deep into execution. The goal is to add validation at every layer the data passes through, making the bug structurally impossible to reproduce.
Use this skill when documentation has drifted from the codebase, needs an inventory, or should be archived.
Autonomous implementation pipeline. Reads a plan file, reviews it adversarially,
>-
| name | onex-brainstorming |
| description | Use this skill when the user has a rough idea, wants to explore a feature, or says "let's brainstorm" or "help me design." The goal is to refine an idea into a validated design doc before any code is written. |
| disable-model-invocation | true |
Use this skill when the user has a rough idea, wants to explore a feature, or says "let's brainstorm" or "help me design." The goal is to refine an idea into a validated design doc before any code is written.
Transform vague ideas into fully-formed designs through collaborative dialogue. This skill ensures alignment on what to build before committing to how to build it.
01-codebase-research)Bounded context check.
Announce what you will read before reading anything. Follow the 01-codebase-research rule. Read only README.md, CLAUDE.md, and one relevant directory listing.
Clarify with one question at a time. Ask questions to understand purpose, constraints, success criteria, and affected components. Only one question per message. Prefer multiple-choice when the answer space is finite. Stop asking when you have enough context to propose approaches.
Propose 2-3 approaches with trade-offs. Present each approach with a name, pros, cons, and a clear recommendation with a one-sentence rationale. Apply YAGNI — remove features not needed for the stated goal.
Present design in 200-300 word sections. After an approach is chosen, present the design incrementally. After each section, ask "Does this look right so far?" and wait for confirmation. Cover: architecture, components, data flow, error handling, testing strategy.
Write the design file.
Save the validated design to docs/plans/YYYY-MM-DD-<topic>-design.md using today's date and a kebab-case topic slug.
Hand off with artifact path.
Output the exact handoff line referencing the saved file path:
**Next step**: In Cursor Composer, invoke rule 11-writing-plans and provide the path docs/plans/YYYY-MM-DD-<topic>-design.md as context.
A Markdown design document saved to docs/plans/ containing:
docs/plans/YYYY-MM-DD-<topic>-design.md