Token导航 LogoToken导航TokenDH.com
研究检索需要联网github未标认证来源可访问clear审计通过

implementation-planning实施规划

Agent Skill

implementation-planning 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

399

周安装

16

GitHub Stars

265

下载量

129
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

3

许可证

MIT

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:implementation-planning(实施规划)
来源仓库:https://github.com/rsmdt/the-startup
仓库路径:skills/implementation-planning
安装命令:
npx skills add https://github.com/rsmdt/the-startup --skill implementation-planning
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/rsmdt/the-startup --skill implementation-planning

简介

implementation-planning 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中快速定位候选结果。

  • 适用于需要根据关键词、任务场景或来源线索进行信息搜集与整理的场景。
  • 通过关键词匹配和来源仓库筛选实现信息检索,支持结合具体任务目标使用。
  • 安装命令为 npx skills add https://github.com/rsmdt/the-startup --skill implementation-planning。
  • 使用前需确认权限范围、维护状态,注意是否触发联网或文件读写操作。

SKILL.md

Implementation Plan Skill

Creates actionable implementation plans that break features into executable tasks following TDD principles. Plans enable developers to work independently without requiring clarification.

Success Criteria

A plan is complete when:

  • A developer can follow it independently without additional context
  • Every task produces a verifiable deliverable (not just an activity)
  • All PRD acceptance criteria map to specific tasks
  • All SDD components have corresponding implementation tasks
  • Dependencies are explicit and no circular dependencies exist

When to Activate

Activate when:

  • Create a new PLAN from the template
  • Complete phases in an existing implementation-plan.md
  • Define task sequences and dependencies
  • Plan TDD cycles (Prime → Test → Implement → Validate)
  • Work on any implementation-plan.md file in docs/specs/

Template

The PLAN template is at template.md. Use this structure exactly.

To write template to spec directory:

  1. Read the template: plugins/start/skills/implementation-plan/template.md
  2. Write to spec directory: docs/specs/[NNN]-[name]/implementation-plan.md

PLAN Focus Areas

Your plan MUST answer these questions:

  • WHAT produces value? (deliverables, not activities)
  • IN WHAT ORDER do tasks execute? (dependencies and sequencing)
  • HOW TO VALIDATE correctness? (test-first approach)
  • WHERE is each task specified? (links to PRD/SDD sections)

Keep plans actionable and focused:

  • Use task descriptions, sequence, and validation criteria
  • Omit time estimates—focus on what, not when
  • Omit resource assignments—focus on work, not who
  • Omit implementation code—the plan guides, implementation follows

Task Granularity Principle

Track logical units that produce verifiable outcomes. The TDD cycle is the execution method, not separate tracked items.

Good Tracking Units (produces outcome)

  • "Payment Entity" → Produces: working entity with tests ✓
  • "Stripe Adapter" → Produces: working integration with tests ✓
  • "Payment Form Component" → Produces: working UI with tests ✓

Bad Tracking Units (too granular)

  • "Read payment interface contracts" → Preparation, not deliverable
  • "Test Payment.validate() rejects negative amounts" → Part of larger outcome
  • "Run linting" → Validation step, not deliverable

Structure Pattern

- [ ] **T1.1 Payment Entity** `[activity: domain-modeling]`

  **Prime**: Read payment interface contracts `[ref: SDD/Section 4.2; lines: 145-200]`

  **Test**: Entity validation rejects negative amounts; supports currency conversion; handles refunds

  **Implement**: Create `src/domain/Payment.ts` with validation logic

  **Validate**: Run unit tests, lint, typecheck

The checkbox tracks "Payment Entity" as a unit. Prime/Test/Implement/Validate are embedded guidance.

TDD Phase Structure

Every task follows red-green-refactor within this pattern:

1. Prime Context

  • Read relevant specification sections
  • Understand interfaces and contracts
  • Load patterns and examples

2. Write Tests (Red)

  • Test behavior before implementation
  • Reference PRD acceptance criteria
  • Cover happy path and edge cases

3. Implement (Green)

  • Build to pass tests
  • Follow SDD architecture
  • Use discovered patterns

4. Validate (Refactor)

  • Run automated tests
  • Check code quality (lint, format)
  • Verify specification compliance

Task Metadata

Use these annotations in the plan:

- [ ] T1.2.1 [Task description] `[ref: SDD/Section 5; lines: 100-150]` `[activity: backend-api]`
MetadataDescription
[parallel: true]Tasks that can run concurrently
[component: name]For multi-component features
[ref: doc/section; lines: X-Y]Links to specifications
[activity: type]Hint for specialist selection

Cycle Pattern

For each phase requiring definition, follow this iterative process:

1. Discovery Phase

  • Read PRD and SDD to understand requirements and design
  • Identify activities needed for each implementation area
  • Launch parallel specialist agents to investigate:

- Task sequencing and dependencies - Testing strategies - Risk assessment - Validation approaches

2. Documentation Phase

  • Update the PLAN with task definitions
  • Add specification references ([ref:...])
  • Focus only on current phase being defined
  • Follow template structure exactly

3. Review Phase

  • Present task breakdown to user
  • Show dependencies and sequencing
  • Highlight parallel opportunities
  • Wait for user confirmation before next phase

Ask yourself each cycle:

  1. Have I read the relevant PRD and SDD sections?
  2. Do all tasks trace back to specification requirements?
  3. Are dependencies between tasks clear?
  4. Can parallel tasks actually run in parallel?
  5. Are validation steps included in each phase?
  6. Have I received user confirmation?

Specification Compliance

Every phase should include a validation task:

- [ ] **T1.3 Phase Validation** `[activity: validate]`

  Run all phase tests, linting, type checking. Verify against SDD patterns and PRD acceptance criteria.

For complex phases, validation is embedded in each task's Validate step.

Deviation Protocol

When implementation requires changes from the specification:

  1. Document the deviation with clear rationale
  2. Obtain approval before proceeding
  3. Update SDD when the deviation improves the design
  4. Record all deviations in the plan for traceability

Validation Checklist

See validation.md for the complete checklist. Key gates:

  • All specification file paths are correct and exist
  • Context priming section is complete
  • All implementation phases are defined
  • Each phase follows TDD: Prime → Test → Implement → Validate
  • Dependencies between phases are clear (no circular dependencies)
  • Parallel work is properly tagged with [parallel: true]
  • Activity hints provided for specialist selection [activity: type]
  • Every phase references relevant SDD sections
  • Every test references PRD acceptance criteria
  • Integration & E2E tests defined in final phase
  • Project commands match actual project setup
  • A developer could follow this plan independently

Output Format

After PLAN work, report:

📋 PLAN Status: [spec-id]-[name]

Phases Defined:
- Phase 1 [Name]: ✅ Complete (X tasks)
- Phase 2 [Name]: 🔄 In progress
- Phase 3 [Name]: ⏳ Pending

Task Summary:
- Total tasks: [N]
- Parallel groups: [N]
- Dependencies: [List key dependencies]

Specification Coverage:
- PRD requirements mapped: [X/Y]
- SDD components covered: [X/Y]

Next Steps:
- [What needs to happen next]

Examples

See examples/phase-examples.md for reference.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

04

需要参考平台分布和安装热度时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Claude Code

28.54%
按下载量换算37

OpenCode

23.92%
按下载量换算31

windsurf

14.66%
按下载量换算19

github-copilot

12.5%
按下载量换算16

Codex

8.18%
按下载量换算11

Gemini CLI

3.55%
按下载量换算5

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。

来源信息

继续浏览同类 Skills