Token导航 LogoToken导航

Agent Skill · 开发

spec-to-plan(计划规格)

spec-to-plan 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。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。

spec-to-plan 是开发类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为需要联网;许可证信息还不明确,第三方扫描已通过,按安装量在开发类的 21,746 个 Skill 里排第 20,588。

总安装 218周安装 9GitHub Stars 6下载 71核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

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:spec-to-plan(计划规格)
来源仓库:https://github.com/kambleakash0/agent-skills
仓库路径:skills/spec-to-plan
安装命令:
npx skills add https://github.com/kambleakash0/agent-skills --skill spec-to-plan
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。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/kambleakash0/agent-skills --skill spec-to-plan

这个 Skill 是做什么的

开发类 Skill 面向写代码、改代码和排查问题,一般会读写仓库文件,部分还会执行构建或测试命令。

spec-to-plan 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。它可辅助 Agent 查询项目结构、追踪变更记录、分析提交历史和检查分支差异。

通过 npx skills add 命令从指定仓库安装,具体用法需结合原始 README 进一步确认。安装前建议核实权限范围、维护状态,以及是否会触发联网、命令执行或文件读写操作。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

来源仓库内路径:skills/spec-to-plan

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/kambleakash0/agent-skills --skill spec-to-plan;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要需要联网,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

spec-to-plan 怎么安装?

在终端执行 npx skills add https://github.com/kambleakash0/agent-skills --skill spec-to-plan。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

spec-to-plan 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

spec-to-plan 需要什么权限?

权限级别是需要联网。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。

spec-to-plan 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

spec-to-plan 用的人多吗?

公开数据是累计安装 218 次,近一周安装 9 次,下载 71 次,GitHub 6 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

spec-to-plan 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理开发类任务、又不想每次都手写提示词的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

开发类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
skill-vetter

技能审查员 · OpenClaw

8,779,018986操作浏览器MIT-0通过
ontology

本体论 · OpenClaw

6,636,960576只读MIT-0通过
entra-app-registration

内部应用程序注册 · Codex、Claude

6,631,872764接触敏感数据未明确通过
entra-app-registration

内部应用程序注册 · Codex、Claude

3,757,680201接触敏感数据未明确通过
nano-banana-pro

纳米香蕉专业版 · OpenClaw

3,526,898364接触敏感数据MIT-0有提醒
spec-to-plan

计划规格 · 本页

2186需要联网未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Turn a PRD into an Implementation Plan

This skill translates what needs to be built (the PRD) into how it will be built (a phased implementation plan). It breaks the PRD into tracer-bullet vertical slices, orders them into phases, and writes the result as a Markdown plan file in ./plans/.

When to Use

Use this skill when:

  • You already have a PRD or equivalent spec and now need a concrete implementation plan.
  • You want to think in phases, not just issues: what gets built first, what can be deferred, and how slices build on each other.
  • You intend to drive downstream work (e.g. /slice-the-spec, /incremental-tdd) from a single, durable planning document.

If the requirements are still fuzzy, use /grill-me and /spec-writer first.

Workflow

You can compress steps if the context is already clear, but keep the core structure: PRD → slices → phases → plan file.

  1. Confirm the PRD is in context

- Ensure the PRD is available in the conversation or repo. - If not, ask the user to paste it or point you to the file path or issue. - Skim for: goals, user stories, functional requirements, non-goals, and constraints.

  1. Explore the codebase (light)

- If you have not already explored the codebase for this feature, do a quick pass to understand major modules, integration layers, and architectural patterns. - Note any constraints that will affect slicing (e.g. existing routing, data models, auth, third-party boundaries).

  1. Identify durable architectural decisions Before slicing, identify decisions that are unlikely to change across phases and are worth calling out up front: These go into the Architectural decisions section of the plan so every phase can reference them.

- Route structures / URL patterns. - Database schema shape and key tables. - Core data models and domain concepts. - Authentication / authorization approach. - Third‑party service boundaries and integration points.

  1. Draft vertical slices (tracer bullets)

- Break the PRD into tracer-bullet slices: thin, end‑to‑end pieces that cut through all relevant layers (schema, domain, API, UI, tests). - Apply the vertical-slice rules: - Each slice delivers a narrow but complete path through every necessary layer. - Each slice is demoable or verifiable on its own. - Prefer many thin slices over a few thick ones. - Do not include brittle implementation details (file names, specific function signatures) that are likely to change. - Do include durable decisions like route paths, schema shapes, and model names.

  1. Group slices into phases

- Order slices into 2–5 phases that make sense from a risk and value perspective. - Earlier phases should: - Validate the riskiest assumptions. - Deliver something demoable quickly. - Set up patterns that later phases can follow. - Later phases can expand coverage, add polish, or handle edge cases and scale.

  1. Quiz the user

- Present the proposed phases as a numbered list. For each phase, show: - Title: short, descriptive name. - User stories covered: which PRD stories this phase delivers. - Ask: - Does the granularity feel right (too coarse / too fine)? - Are phases in the right order? - Should any phases be merged or split? - Refine until the user approves the breakdown.

  1. Write the plan file

- Ensure ./plans/ exists; if not, create it. - Name the plan file after the feature (e.g. ./plans/weekly-admin-summary-email.md). - Use and adapt the template below, filling in architectural decisions and one section per phase.

Plan Template

Use this as the default Markdown structure.

# Plan: <Feature Name>
> Source PRD: <brief identifier or link>

## Architectural decisions

Durable decisions that apply across all phases:

- **Routes:** ...
- **Schema:** ...
- **Key models:** ...
- **Auth:** ...
- **Third-party services:** ...
- (add/remove items as appropriate)

## Phases

### Phase 1: <Title>

#### What to build

A concise description of this vertical slice. Describe the **end-to-end behavior**, not layer-by-layer implementation.[1]

- Which user stories this phase covers.
- Which layers it will touch (data, domain, API, UI, background jobs, etc.).[1]

#### Acceptance criteria

- Criterion 1
- Criterion 2
- Criterion 3

***

### Phase 2: <Title>

#### What to build

...

#### Acceptance criteria

- ...

<!-- Repeat for each phase -->

Behavior and Rules

  1. Stay at the plan level. Do not drop into detailed implementation plans or specific file edits; that is work for execution skills like /tdd.
  2. Think in risk and value. Earlier phases should de-risk core architecture and ship something demoable quickly.
  3. Keep slices vertical. Avoid phases that are purely “set up database” or “build UI” without delivering end-to-end value.
  4. Make decisions explicit. Put durable architectural decisions in the plan header so downstream skills and humans can reuse them.
  5. Write for future you and agents. Plans should be readable, skimmable, and easy to hand to another engineer or AI agent as a starting point.

平台分布

Codex

38.09%
按下载量换算27

Claude

28.2%
按下载量换算20

Cursor

16.51%
按下载量换算12

Gemini CLI

8.77%
按下载量换算6

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。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills