speckit-plan-zh

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

用户输入

User Input

text
$ARGUMENTS
在继续之前,您必须考虑用户输入(如果不为空)。
text
$ARGUMENTS
Before proceeding, you must consider the user input (if not empty).

大纲

Outline

  1. 设置: 从仓库根目录运行
    .specify/scripts/powershell/setup-plan.ps1 -Json
    并解析 JSON 以获取 FEATURE_SPEC、IMPL_PLAN、SPECS_DIR、BRANCH。对于参数中的单引号,如 "I'm Groot",请使用转义语法:例如 'I'''m Groot'(或者如果可能的话使用双引号:"I'm Groot")。
  2. 加载上下文: 读取 FEATURE_SPEC 和
    .specify/memory/constitution.md
    。加载 IMPL_PLAN 模板(已复制)。
  3. 执行计划工作流程: 遵循 IMPL_PLAN 模板中的结构来:
    • 填写技术上下文(将未知项标记为"需要澄清")
    • 从章程中填写章程检查部分
    • 评估门禁(如果有未正当化的违规则报错)
    • 阶段 0: 生成 research.md(解决所有"需要澄清")
    • 阶段 1: 生成 data-model.md、contracts/、quickstart.md
    • 阶段 1: 通过运行代理脚本更新代理上下文
    • 设计后重新评估章程检查
  4. 停止并报告: 命令在阶段 2 规划后结束。报告分支、IMPL_PLAN 路径和生成的工件。
  1. Setup: Run
    .specify/scripts/powershell/setup-plan.ps1 -Json
    from the root directory of the repository and parse the JSON to obtain FEATURE_SPEC, IMPL_PLAN, SPECS_DIR, BRANCH. For single quotes in parameters, such as "I'm Groot", use escape syntax: e.g., 'I'''m Groot' (or use double quotes if possible: "I'm Groot").
  2. Load Context: Read FEATURE_SPEC and
    .specify/memory/constitution.md
    . Load the copied IMPL_PLAN template.
  3. Execute Plan Workflow: Follow the structure in the IMPL_PLAN template to:
    • Fill in technical context (mark unknown items as "Needs Clarification")
    • Fill in the constitution check section from the constitution
    • Evaluate gateways (report ERROR if there are unjustified violations)
    • Phase 0: Generate research.md (resolve all "Needs Clarification" items)
    • Phase 1: Generate data-model.md, contracts/, quickstart.md
    • Phase 1: Update agent context by running agent scripts
    • Re-evaluate constitution check after design
  4. Stop and Report: The command ends after Phase 2 planning. Report the branch, IMPL_PLAN path, and generated artifacts.

阶段

Phases

阶段 0: 大纲与研究

Phase 0: Outline & Research

  1. 从上述技术上下文中提取未知项:
    • 对于每个"需要澄清" → 研究任务
    • 对于每个依赖项 → 最佳实践任务
    • 对于每个集成 → 模式任务
  2. 生成并分发研究代理:
    text
    对于技术上下文中的每个未知项:
      任务: "研究 {未知项} 用于 {功能上下文}"
    对于每个技术选择:
      任务: "查找 {技术} 在 {领域} 中的最佳实践"
  3. research.md
    中整合发现结果
    ,使用格式:
    • 决策: [选择了什么]
    • 理由: [为什么选择]
    • 考虑的替代方案: [还评估了什么]
输出: 解决了所有"需要澄清"的 research.md
  1. Extract unknown items from the above technical context:
    • For each "Needs Clarification" → Research task
    • For each dependency → Best practice task
    • For each integration → Pattern task
  2. Generate and distribute research agents:
    text
    For each unknown item in the technical context:
      Task: "Research {unknown item} for {feature context}"
    For each technical choice:
      Task: "Find best practices for {technology} in {domain}"
  3. Integrate findings in
    research.md
    using the format:
    • Decision: [What was chosen]
    • Rationale: [Why it was chosen]
    • Alternatives Considered: [What else was evaluated]
Output: research.md with all "Needs Clarification" items resolved

阶段 1: 设计与契约

Phase 1: Design & Contracts

前提条件:
research.md
完成
  1. 从功能规格中提取实体
    data-model.md
    :
    • 实体名称、字段、关系
    • 来自需求的验证规则
    • 如适用的状态转换
  2. 从功能需求生成 API 契约:
    • 对于每个用户操作 → 端点
    • 使用标准的 REST/GraphQL 模式
    • 将 OpenAPI/GraphQL 模式输出到
      /contracts/
  3. 代理上下文更新:
    • 运行
      .specify/scripts/powershell/update-agent-context.ps1 -AgentType claude
    • 这些脚本检测正在使用的 AI 代理
    • 更新相应的代理特定上下文文件
    • 仅添加当前计划中的新技术
    • 保留标记之间的手动添加内容
输出: data-model.md、/contracts/*、quickstart.md、代理特定文件
Prerequisite:
research.md
is completed
  1. Extract entities from feature specification
    data-model.md
    :
    • Entity names, fields, relationships
    • Validation rules from requirements
    • State transitions if applicable
  2. Generate API contracts from functional requirements:
    • For each user action → Endpoint
    • Use standard REST/GraphQL schema
    • Output OpenAPI/GraphQL schema to
      /contracts/
  3. Agent Context Update:
    • Run
      .specify/scripts/powershell/update-agent-context.ps1 -AgentType claude
    • These scripts detect the AI agent being used
    • Update the corresponding agent-specific context files
    • Only add new technologies from the current plan
    • Keep manually added content between markers
Output: data-model.md, /contracts/*, quickstart.md, agent-specific files

关键规则

Key Rules

  • 使用绝对路径
  • 如果检查点失败或存在未解决的澄清项,则报 ERROR
  • Use absolute paths
  • Report ERROR if checkpoint fails or there are unresolved clarification items