planning
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
ChineseImplementation Planning
实施规划
Create comprehensive, actionable implementation plans for features and projects through structured requirement gathering, codebase analysis, and detailed planning.
通过结构化的需求收集、代码库分析和详细规划,为功能和项目创建全面、可执行的实施计划。
What This Skill Provides
本Skill提供的能力
This skill enables creating comprehensive, actionable implementation plans that include:
- Structured requirement gathering - Systematic questions to understand features completely
- Codebase analysis - Automated exploration of project structure and patterns
- Multi-phase planning - Requirements, research, architecture, implementation details
- Parallel agent coordination - Leverage specialized agents for complex analysis
本Skill可创建全面、可执行的实施计划,包含以下内容:
- 结构化需求收集 - 系统性提问以全面理解功能需求
- 代码库分析 - 自动探索项目结构与模式
- 多阶段规划 - 需求、调研、架构、实施细节全流程覆盖
- 多Agent并行协同 - 借助专业Agent完成复杂分析
When to Use This Skill
何时使用本Skill
Use this skill when:
- User requests a detailed implementation plan
- Feature requires architectural decisions
- Need to analyze existing codebase before implementation
- Multiple teams or modules are involved
- Security, performance, or scalability considerations exist
在以下场景使用本Skill:
- 用户需要详细的实施计划
- 功能开发需要做出架构决策
- 实施前需要分析现有代码库
- 涉及多个团队或模块
- 需要考虑安全、性能或可扩展性因素
Core Planning Workflow
核心规划工作流
Follow this six-phase workflow to create comprehensive plans:
遵循以下六阶段工作流创建全面计划:
Phase 1: Requirement Gathering
阶段1:需求收集
FIRST: Ask for a feature name (kebab-case):
Examples: "oauth-authentication", "user-profile", "data-export"The feature name becomes the plan filename:
.opencode/plans/plan-<feature-name>.mdTHEN: Ask requirement questions:
- What is the goal or feature to be implemented?
- What are the key requirements and constraints?
- Are there specific technical requirements or preferences?
- What is the expected outcome or deliverables?
- Are there dependencies or integration points?
Important: Don't proceed without the feature name and clear answers.
第一步:询问功能名称(使用kebab-case格式):
Examples: "oauth-authentication", "user-profile", "data-export"功能名称将作为计划文件名:
.opencode/plans/plan-<feature-name>.md第二步:询问需求相关问题:
- 要实现的目标或功能是什么?
- 核心需求与约束条件有哪些?
- 是否有特定的技术要求或偏好?
- 预期成果或交付物是什么?
- 是否存在依赖项或集成点?
重要提示:未获取功能名称与明确的需求答复前,请勿推进后续流程。
Phase 2: Context Analysis
阶段2:上下文分析
Analyze the project automatically:
Project Structure:
bash
find . -type f -name "*.ts" -o -name "*.tsx" -o -name "*.js" -o -name "*.jsx" -o -name "*.py" -o -name "*.go" -o -name "*.java" | head -50Git Context:
bash
git status
git log --oneline -10Codebase Analysis:
-
Use the explore agent to identify:
- Relevant files and modules
- Existing patterns and conventions
- Technology stack and frameworks
- Architecture patterns
-
Load appropriate skills based on detected technologies:
- TypeScript, Python, Go, etc.
- Framework-specific skills (React, Vue, etc.)
-
Use parallel subagents where beneficial:
- Frontend analysis (if applicable)
- Backend analysis (if applicable)
- Database schema review (if applicable)
- API design review (if applicable)
自动分析项目:
项目结构:
bash
find . -type f -name "*.ts" -o -name "*.tsx" -o -name "*.js" -o -name "*.jsx" -o -name "*.py" -o -name "*.go" -o -name "*.java" | head -50Git上下文:
bash
git status
git log --oneline -10代码库分析:
-
使用explore agent识别:
- 相关文件与模块
- 现有模式与规范
- 技术栈与框架
- 架构模式
-
根据检测到的技术加载对应Skill:
- TypeScript、Python、Go等
- 框架专属Skill(React、Vue等)
-
在合适场景使用并行子Agent:
- 前端分析(如适用)
- 后端分析(如适用)
- 数据库架构评审(如适用)
- API设计评审(如适用)
Phase 3: Research and Verification
阶段3:调研与验证
Web Search for uncertain approaches:
- Official documentation
- Current best practices
- Known issues or gotchas
- Relevant examples or patterns
Consult Documentation:
- Framework documentation
- Library API references
- Internal project documentation
针对不确定的方案进行网络搜索:
- 官方文档
- 当前最佳实践
- 已知问题与注意事项
- 相关示例或模式
查阅文档:
- 框架文档
- 库API参考
- 内部项目文档
Phase 4: Planning with Subagents
阶段4:借助子Agent规划
Utilize specialized agents in parallel where appropriate:
- Architecture Agent - System design decisions
- Security Agent - Security considerations
- Performance Agent - Performance optimization strategies
- Frontend Agent - UI/UX implementation (if applicable)
- Backend Agent - Server-side implementation (if applicable)
Run agents in parallel when work is independent to maximize efficiency.
在合适场景并行使用专业Agent:
- Architecture Agent - 系统设计决策
- Security Agent - 安全考量
- Performance Agent - 性能优化策略
- Frontend Agent - UI/UX实施(如适用)
- Backend Agent - 服务端实施(如适用)
当任务相互独立时,并行运行Agent以提升效率。
Phase 5: Create Implementation Plan
阶段5:创建实施计划
Generate a detailed plan with these sections:
- Overview - Summary, goals, success criteria, high-level approach
- Technical Architecture - Components, data flow, integration points
- Implementation Steps - Step-by-step tasks with file changes and code snippets
- Dependencies - External packages, internal modules, API integrations
- Testing Strategy - Unit, integration, E2E tests required
- Security Considerations - Auth impacts, validation, best practices
- Performance Considerations - Impact, optimization, scalability
- Migration Path - Breaking changes, data migration, rollback strategy
- Documentation Updates - Code docs, README, API docs
- Rollout Plan - Deployment steps, feature flags, monitoring
See for detailed section templates.
references/plan-structure.md生成包含以下模块的详细计划:
- 概述 - 摘要、目标、成功标准、整体方案
- 技术架构 - 组件、数据流、集成点
- 实施步骤 - 包含文件变更与代码片段的分步任务
- 依赖项 - 外部包、内部模块、API集成
- 测试策略 - 所需的单元测试、集成测试、E2E测试
- 安全考量 - 认证影响、校验机制、最佳实践
- 性能考量 - 影响范围、优化方案、可扩展性
- 迁移路径 - 破坏性变更、数据迁移、回滚策略
- 文档更新 - 代码文档、README、API文档
- 上线计划 - 部署步骤、功能开关、监控方案
详细模块模板请参考****。
references/plan-structure.mdPhase 6: Output Plan File and User Verification
阶段6:输出计划文件并获取用户确认
Present plan summary to user for verification:
📋 Plan Summary
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Feature: <feature-name>
Description: <brief description>
Key Components:
• <Component 1>
• <Component 2>
• <Component 3>
Implementation Steps: <count> steps
Estimated Complexity: <Low/Medium/High>
Dependencies:
• <Dependency 1>
• <Dependency 2>
Security Considerations: <Yes/No>
Performance Considerations: <Yes/No>
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Please review the plan summary above.
Would you like to proceed with this plan? (yes/no)Wait for user confirmation:
- If user confirms: Proceed with implementation using
/implement - If user wants changes: Ask what needs to be modified and update the plan
- If user rejects: Ask if they want to gather different requirements
向用户展示计划摘要以获取验证:
📋 Plan Summary
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Feature: <feature-name>
Description: <brief description>
Key Components:
• <Component 1>
• <Component 2>
• <Component 3>
Implementation Steps: <count> steps
Estimated Complexity: <Low/Medium/High>
Dependencies:
• <Dependency 1>
• <Dependency 2>
Security Considerations: <Yes/No>
Performance Considerations: <Yes/No>
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Please review the plan summary above.
Would you like to proceed with this plan? (yes/no)等待用户确认:
- 若用户确认:使用指令推进实施
/implement - 若用户需要修改:询问具体修改内容并更新计划
- 若用户拒绝:询问是否需要重新收集需求
Plan Quality Standards
计划质量标准
The plan file must be:
- Comprehensive - Cover all aspects of implementation
- Actionable - Clear steps that can be followed directly
- Detailed - Include code snippets, file paths, specific instructions
- Ready to implement - No ambiguity or missing information
计划文件必须满足:
- 全面性 - 覆盖实施全流程各方面
- 可执行性 - 步骤清晰可直接落地
- 细节丰富 - 包含代码片段、文件路径、具体指令
- 可直接实施 - 无模糊表述或缺失信息
Execution Checklist
执行检查清单
- Start by asking for feature name
- Ask requirement questions thoroughly
- Use agents in parallel when tasks are independent
- Load skills proactively based on detected technologies
- Web search for verification of best practices
- Think step-by-step through the phases
- Be thorough enough for another developer to implement
- Verify user understanding of the plan before proceeding to implementation
- 首先询问功能名称
- 全面询问需求相关问题
- 任务独立时并行使用Agent
- 根据检测到的技术主动加载对应Skill
- 通过网络搜索验证最佳实践
- 按阶段逐步推进规划
- 内容足够详尽,确保其他开发者可直接实施
- 实施前获取用户对计划的确认
Additional Resources
额外资源
Reference Files
参考文件
- - Detailed templates for all 10 plan sections
references/plan-structure.md - - Comprehensive question catalog for different feature types
references/requirement-questions.md - - Strategies for parallel agent usage
references/agent-coordination.md - - Plan status management and workflow integration
references/plan-tracking.md
- - 10个计划模块的详细模板
references/plan-structure.md - - 针对不同功能类型的全面问题库
references/requirement-questions.md - - 多Agent并行协同策略
references/agent-coordination.md - - 计划状态管理与工作流集成
references/plan-tracking.md
Quick Reference: Feature Name Examples
快速参考:功能名称示例
| Feature Type | Good Name | Bad Name |
|---|---|---|
| Authentication | | |
| User Profile | | |
| Data Export | | |
| API Integration | | |
| Database Migration | | |
Feature names should be:
- Descriptive and specific
- Lowercase with hyphens (kebab-case)
- Unique within the project
- 2-4 words typically
| 功能类型 | 规范名称 | 不规范名称 |
|---|---|---|
| 认证 | | |
| 用户资料 | | |
| 数据导出 | | |
| API集成 | | |
| 数据库迁移 | | |
功能名称应满足:
- 描述性强、具体明确
- 小写字母加连字符(kebab-case格式)
- 项目内唯一
- 通常包含2-4个单词
Summary
总结
Use this skill to create comprehensive implementation plans that:
- Gather complete requirements upfront
- Analyze existing codebase automatically
- Research best practices and verify approaches
- Coordinate specialized agents in parallel
- Generate detailed, actionable plans
- Present plans to users for confirmation before proceeding
The resulting plans are production-ready and can be handed to any developer for implementation without additional questions.
使用本Skill创建的实施计划具备以下特点:
- 提前收集完整需求
- 自动分析现有代码库
- 调研最佳实践并验证方案
- 并行协同专业Agent
- 生成详细、可执行的计划
- 实施前向用户展示计划并获取确认
最终生成的计划可直接用于生产环境,无需额外沟通即可交付给任意开发者实施。