brainstorming

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Brainstorming Skill

头脑风暴Skill

Overview

概述

Collaborative design exploration for new features, architecture decisions, and complex problem-solving.
针对新功能、架构决策及复杂问题解决的协作式设计探索。

Triggers

触发条件

Activate this skill when:
  • User says "let's brainstorm", "let's ideate", or "let's explore"
  • User runs
    /ideate
    command
  • User wants to discuss design options before implementation
  • A problem has multiple valid solutions needing evaluation
For a complete worked example, see
references/worked-example.md
.
在以下场景激活此Skill:
  • 用户说 "let's brainstorm"、"let's ideate" 或 "let's explore"
  • 用户执行
    /ideate
    命令
  • 用户希望在实现前讨论设计方案
  • 某个问题有多个可行解决方案需要评估
完整示例请查看
references/worked-example.md

Three-Phase Process

三阶段流程

Phase 1: Understanding

阶段1:理解

Goal: Deeply understand the problem before proposing solutions.
Rules:
  • Ask ONE question at a time
  • Wait for response before asking next question
  • Focus on: goals, constraints, existing patterns, user preferences
  • Maximum 5 questions before moving to exploration
Question Types:
  1. "What problem are we solving?" (core need)
  2. "What constraints exist?" (time, tech, compatibility)
  3. "What patterns already exist in the codebase?" (consistency)
  4. "Who/what will consume this?" (users, APIs, other systems)
  5. "What does success look like?" (acceptance criteria)
目标: 在提出解决方案前深入理解问题。
规则:
  • 一次仅提出一个问题
  • 等待用户回复后再提出下一个问题
  • 聚焦于:目标、约束条件、现有模式、用户偏好
  • 最多提出5个问题后进入探索阶段
问题类型:
  1. "我们要解决什么问题?"(核心需求)
  2. "存在哪些约束条件?"(时间、技术、兼容性)
  3. "代码库中已存在哪些模式?"(一致性)
  4. "谁/什么会使用这个功能?"(用户、API、其他系统)
  5. "成功的标准是什么?"(验收标准)

Phase 2: Exploration

阶段2:探索

Goal: Present 2-3 distinct approaches with trade-offs.
Use the approach format from
references/design-template.md
. Present genuinely different approaches with honest trade-offs. Recommend one option with rationale.
目标: 呈现2-3种不同的方案及权衡分析。
使用
references/design-template.md
中的方案格式。展示真正不同的方案并如实说明权衡点。推荐一个选项并给出理由。

Phase 3: Design Presentation

阶段3:设计呈现

Goal: Document the chosen approach in detail with numbered requirements.
Document the chosen approach using the structure in
references/design-template.md
. Sections of 200-300 words max. Use diagrams for complex flows.
Requirements format (MANDATORY):
  • Use numbered requirement identifiers:
    DR-1
    ,
    DR-2
    , ...,
    DR-N
  • Each requirement MUST have an
    **Acceptance criteria:**
    block with concrete, testable criteria
  • At least one requirement MUST address error handling, failure modes, or edge cases
  • These DR-N identifiers are provenance anchors — implementation plans trace tasks to them
Save Location:
docs/designs/YYYY-MM-DD-<feature>.md
目标: 详细记录选定的方案,包含编号的需求项。
使用
references/design-template.md
中的结构记录选定的方案。每个部分最多200-300字。复杂流程使用图表展示。
需求格式(必填):
  • 使用编号的需求标识:
    DR-1
    DR-2
    ……
    DR-N
  • 每个需求必须包含
    **Acceptance criteria:**
    块,内容为具体、可测试的标准
  • 至少有一个需求必须涉及错误处理、故障模式或边缘情况
  • 这些DR-N标识是溯源锚点——实现规划会将任务与它们关联
保存位置:
docs/designs/YYYY-MM-DD-<feature>.md

Iteration Limits

迭代限制

Design iterations: max 3. If Phase 2 (Exploration) cycles through 3 rounds of presenting approaches without the user converging on a choice, pause and summarize the trade-offs for the user to make a final decision.
The user can override:
/ideate --max-iterations 5
设计迭代:最多3次。如果阶段2(探索)经过3轮方案展示后用户仍未选定方案,请暂停并总结权衡点,供用户做出最终决策。
用户可覆盖此限制:
/ideate --max-iterations 5

Anti-Patterns

反模式

Don'tDo Instead
Jump to solution immediatelyAsk clarifying questions first
Present only one optionAlways show 2-3 alternatives
Hide drawbacks of preferred optionBe transparent about trade-offs
Write walls of textUse 200-300 word sections max
Ignore existing patternsReference codebase conventions
Skip documentationSave design to docs/designs/
禁止做法正确做法
直接跳到解决方案先提出澄清问题
只展示一个选项始终展示2-3种替代方案
隐藏首选方案的缺点透明说明权衡点
大段文字堆砌每个部分最多200-300字
忽略现有模式参考代码库约定
跳过文档记录将设计保存到docs/designs/

State Management

状态管理

This skill manages workflow state for context persistence.
此Skill管理工作流状态以保持上下文持久化。

On Start (before Phase 1)

启动时(阶段1之前)

Initialize workflow state using
mcp__exarchos__exarchos_workflow
with
action: "init"
,
workflowType: "feature"
, and the featureId.
This creates a state file tracked by the MCP server.
使用
mcp__exarchos__exarchos_workflow
初始化工作流状态,参数为
action: "init"
workflowType: "feature"
及featureId。
这会创建一个由MCP服务器跟踪的状态文件。

On Design Save (after Phase 3)

设计保存时(阶段3之后)

action: "set", featureId: "<id>", updates: { "artifacts": { "design": "<path>" } }, phase: "plan"
action: "set", featureId: "<id>", updates: { "artifacts": { "design": "<path>" } }, phase: "plan"

Phase Transitions and Guards

阶段转换与防护

This skill is the entry point for the feature workflow (
workflowType: "feature"
). The full lifecycle is:
ideate → plan → plan-review → delegate ⇄ review → synthesize → completed
For the full transition table, consult
@skills/workflow-state/references/phase-transitions.md
.
此Skill是功能工作流
workflowType: "feature"
)的入口点。完整生命周期为:
ideate → plan → plan-review → delegate ⇄ review → synthesize → completed
完整转换表请查阅
@skills/workflow-state/references/phase-transitions.md

Schema Discovery

架构发现

Use
exarchos_workflow({ action: "describe", actions: ["set", "init"] })
for parameter schemas and
exarchos_workflow({ action: "describe", playbook: "feature" })
for phase transitions, guards, and playbook guidance.
使用
exarchos_workflow({ action: "describe", actions: ["set", "init"] })
获取参数架构,使用
exarchos_workflow({ action: "describe", playbook: "feature" })
获取阶段转换、防护及工作指南。

Completion Verification

完成验证

Run the ideation artifact verification:
typescript
mcp__exarchos__exarchos_orchestrate({
  action: "check_design_completeness",
  featureId: "<featureId>",
  designPath: "docs/designs/YYYY-MM-DD-<feature>.md"
})
On
passed: true
:
All completion criteria met — proceed to gate check. On
passed: false
:
Missing artifacts — review output and complete before continuing. If the check is advisory (
advisory: true
), emit a warning but do not block auto-chain.
执行构思工件验证:
typescript
mcp__exarchos__exarchos_orchestrate({
  action: "check_design_completeness",
  featureId: "<featureId>",
  designPath: "docs/designs/YYYY-MM-DD-<feature>.md"
})
passed: true
时:
所有完成条件满足——进入 gate check。
passed: false
时:
缺少工件——查看输出并补充完整后再继续。如果检查是建议性的(
advisory: true
),则发出警告但不阻止自动链式执行。

Adversarial Gate Check (ideate → plan)

对抗性Gate Check(ideate → plan)

After artifact verification passes, run the design completeness gate check. This is the D1 (spec fidelity) lightweight adversarial check at the ideate → plan boundary.
typescript
mcp__exarchos__exarchos_orchestrate({
  action: "check_design_completeness",
  featureId: "<id>",
  designPath: "<path>"
})
The handler returns a structured result:
{ passed, advisory, findings[], checkCount, passCount, failCount }
.
  • passed=true
    :
    Design complete — all requirements have acceptance criteria and error coverage.
  • passed=false, advisory=true
    :
    Findings detected. These are advisory — they do NOT block the auto-chain to
    /plan
    . Present
    result.data.findings
    to the user alongside the transition message.
Gate events (
gate.executed
) are emitted automatically by the handler — no manual event emission is needed.
工件验证通过后,执行设计完整性gate check。这是ideate → plan边界处的D1(规范保真度)轻量级对抗性检查。
typescript
mcp__exarchos__exarchos_orchestrate({
  action: "check_design_completeness",
  featureId: "<id>",
  designPath: "<path>"
})
处理器返回结构化结果:
{ passed, advisory, findings[], checkCount, passCount, failCount }
  • passed=true
    设计完整——所有需求都有验收标准和错误覆盖。
  • passed=false, advisory=true
    检测到问题。这些是建议性的——不会阻止自动链式执行到
    /plan
    。将
    result.data.findings
    展示给用户并附带转换消息。
Gate事件(
gate.executed
)由处理器自动触发——无需手动触发事件。

Transition

转换

After brainstorming completes, auto-continue to planning (no user confirmation):
头脑风暴完成后,自动进入规划阶段(无需用户确认):

Pre-Chain Validation (MANDATORY)

链式前验证(必填)

Before invoking
/plan
:
  1. Verify
    artifacts.design
    exists in workflow state
  2. Verify the design file exists on disk:
    test -f "$DESIGN_PATH"
  3. Run
    mcp__exarchos__exarchos_orchestrate({ action: "check_design_completeness", featureId: "<id>", designPath: "<path>" })
    (advisory — record findings but don't block)
  4. If steps 1 or 2 fail: "Design artifact not found, cannot auto-chain to /plan"
调用
/plan
前:
  1. 验证工作流状态中存在
    artifacts.design
  2. 验证设计文件存在于磁盘:
    test -f "$DESIGN_PATH"
  3. 执行
    mcp__exarchos__exarchos_orchestrate({ action: "check_design_completeness", featureId: "<id>", designPath: "<path>" })
    (建议性——记录问题但不阻止)
  4. 如果步骤1或2失败:"未找到设计工件,无法自动链式执行到/plan"

Chain Steps

链式步骤

  1. Update state:
    action: "set", featureId: "<id>", updates: { "artifacts": { "design": "<path>" } }, phase: "plan"
  2. If
    result.data.passed === false
    and
    result.data.advisory === true
    : Output
    result.data.findings
    summary, then: "Advisory findings noted. Auto-continuing to implementation planning..." If
    result.data.passed === true
    : Output: "Design complete. Auto-continuing to implementation planning..."
  3. Invoke immediately:
    typescript
    [Invoke the exarchos:plan skill with args: <design-path>]
This is NOT a human checkpoint. The human checkpoint occurs at plan-review (plan approval) and synthesize (merge confirmation).
Workflow continues:
/ideate
->
/plan
-> plan-review -> [HUMAN CHECKPOINT] ->
/delegate
->
/review
->
/synthesize
-> [HUMAN CHECKPOINT]
  1. 更新状态:`action: "set", featureId: "<id>", updates: { "artifacts": { "design": "<path>" } }, phase: "plan"
  2. 如果
    result.data.passed === false
    result.data.advisory === true
    :输出
    result.data.findings
    摘要,然后:"已记录建议性问题。自动进入实现规划..." 如果
    result.data.passed === true
    :输出:"设计完成。自动进入实现规划..."
  3. 立即调用:
    typescript
    [调用exarchos:plan skill,参数:<design-path>]
这不是人工检查点。人工检查点位于plan-review(规划审批)和synthesize(合并确认)阶段。
工作流继续:
/ideate
->
/plan
-> plan-review -> [人工检查点] ->
/delegate
->
/review
->
/synthesize
-> [人工检查点]

Exarchos Integration

Exarchos集成

When Exarchos MCP tools are available:
  1. At workflow start: Auto-emitted by
    exarchos_workflow
    action: "init"
    — do NOT manually append
    workflow.started
当Exarchos MCP工具可用时:
  1. 工作流启动时:
    exarchos_workflow
    action: "init"
    自动触发——请勿手动添加
    workflow.started