brainstorm by buiducnhat/agent-skills
npx skills add https://github.com/buiducnhat/agent-skills --skill brainstorm使用此技能,通过结构化对话将粗略的想法转化为清晰、可评审的设计产出。
目标是:
根据共享的 Context Loading Protocol 加载项目背景。同时检查与想法相关的关键实现文件,并注意来自现有架构、依赖项和约定的约束。
此步骤应保持专注。仅收集当前想法所需的内容。
按顺序提出有针对性的问题以消除歧义。遵循 AskUserQuestion 指令。
不要过早跳入实现细节。
提出 2-5 个可行的方案。
对于每个方案,包括:
首先提出你推荐的方案,并解释为什么它最适合项目背景和约束。
在呈现所有方案后,使用 AskUserQuestion 让用户选择他们偏好的方案(例如,“方案 A — 推荐”、“方案 B”、“方案 C”、“需要更多细节才能决定”)。
一旦需求清晰,以小章节(每章约 200-300 词)呈现设计,并在每个章节后进行验证。
建议的章节顺序:
广告位招租
在这里展示您的产品或服务
触达数万 AI 开发者,精准高效
在每个章节后,使用 AskUserQuestion 询问是继续、调整还是回顾。提供具体选项(例如,“进入下一章节”、“调整此章节”、“回到前一章节”)。
(仅当用户在步骤 6 中明确选择“撰写产出物”后才执行此步骤。)
将结果持久化到标准化位置:
docs/brainstorms/YYMMDD-HHmm-<topic-slug>/docs/brainstorms/YYMMDD-HHmm-<topic-slug>/SUMMARY.mddocs/brainstorms/YYMMDD-HHmm-<topic-slug>/section-01-<slug>.mddocs/brainstorms/YYMMDD-HHmm-<topic-slug>/section-02-<slug>.mdSUMMARY.md 应包含:
在你和用户处理完需求及任何澄清性问题后,是时候决定如何处理这些信息了。
使用 AskUserQuestion 向用户呈现三个高层次的后续行动:
write-plan。如果用户选择 撰写产出物,则按照步骤 5 在 docs/brainstorms/... 下创建相应的文件。产出物草稿存在后,再次使用 AskUserQuestion 并提供选项来验证它们:
如果用户选择了 立即撰写计划,则使用当前的 brainstorming 上下文启动到 write-plan 的交接;无需额外的产出物验证。
如果用户选择了 结束会话,则直接停止。目前收集的信息被认为是足够的。
(对于以上所有问题,请遵循 AskUserQuestion 指令:一次只问一个问题,并提供可选择的选项。)
在最终确定 SUMMARY.md 之前,确认:
每周安装次数
85
仓库
GitHub 星标数
26
首次出现
11 天前
安全审计
安装于
claude-code73
cursor57
antigravity48
github-copilot47
gemini-cli42
opencode38
Use this skill to convert rough ideas into clear, reviewable design outputs through structured dialogue.
The goal is to:
Load project context per the shared Context Loading Protocol. Also check key implementation files relevant to the idea and note constraints from existing architecture, dependencies, and conventions.
Keep this pass focused. Only gather what is needed for the current idea.
Ask targeted questions sequentially to remove ambiguity. Follow the AskUserQuestion mandate.
Do not jump to implementation details too early.
Propose 2-5 viable approaches.
For each approach, include:
Lead with your recommended option and explain why it best fits the project context and constraints.
After presenting all approaches, use AskUserQuestion to let the user pick their preferred approach (e.g., "Approach A — recommended", "Approach B", "Approach C", "Need more details before deciding").
Once requirements are clear, present the design in small sections (about 200-300 words each), validating after each section.
Suggested section order:
After each section, use AskUserQuestion to ask whether to proceed, adjust, or revisit. Provide concrete options (e.g., "Proceed to next section", "Adjust this section", "Go back to a previous section").
(Only perform this step after the user has explicitly chosen “Write artifacts” during Step 6.)
Persist results to the standardized location:
docs/brainstorms/YYMMDD-HHmm-<topic-slug>/docs/brainstorms/YYMMDD-HHmm-<topic-slug>/SUMMARY.mddocs/brainstorms/YYMMDD-HHmm-<topic-slug>/section-01-<slug>.mddocs/brainstorms/YYMMDD-HHmm-<topic-slug>/section-02-<slug>.mdSUMMARY.md should contain:
After you and the user have worked through requirements and any clarifying questions, it's time to decide what to do with the information.
Use AskUserQuestion to present the user with three high‑level next actions:
write-plan handoff.If the user picks Write artifacts , create the appropriate files under docs/brainstorms/... per Step 5. Once the draft artifacts exist, use AskUserQuestion again to validate them with options:
If the user picked Write plan immediately , initiate a handoff to write-plan using the current brainstorming context; no additional artifact validation is required.
(For all questions above, follow the AskUserQuestion mandate: ask one question at a time with selectable options.)
Before finalizing SUMMARY.md, confirm:
Weekly Installs
85
Repository
GitHub Stars
26
First Seen
11 days ago
Security Audits
Gen Agent Trust HubPassSocketPassSnykPass
Installed on
claude-code73
cursor57
antigravity48
github-copilot47
gemini-cli42
opencode38
站立会议模板:敏捷开发每日站会指南与工具(含远程团队异步模板)
10,500 周安装
first-ask AI助手:先理解再执行,提升GitHub Copilot等AI编程工具协作效率
7,700 周安装
create-tldr-page:AI辅助生成简洁命令文档的GitHub Copilot技能
7,700 周安装
GitHub Copilot 规范转 Issue 工具:自动从规范文件创建 GitHub Issue 功能
7,700 周安装
纯文本转Markdown工具 - 一键转换文本文档为格式规范的Markdown文件
7,700 周安装
GitHub Copilot 指令推荐工具 - 自动分析项目并推荐优质AI编程指令
7,700 周安装
结构化自主规划工具 - GitHub Copilot项目开发计划制定与分解助手
7,700 周安装
If the user picked End session , simply stop. The information collected so far is considered sufficient.