Token导航 LogoToken导航

Agent Skill · 开发

mvp-planner(MVP 策划师)

mvp-planner 用于处理浏览器自动化、网页检查和页面信息提取,适合在 OpenClaw 中需要让 Agent 打开页面、读取网页或验证前端流程时使用。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。

mvp-planner 是开发类 Agent Skill,可用宿主为 OpenClaw,通过ClawHub获取,权限为执行命令;许可证为 MIT-0,第三方扫描已通过,按安装量在开发类的 21,746 个 Skill 里排第 9,355。

总安装 2,747周安装 118GitHub Stars 公开资料未说明下载 963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
MIT-0
版本
1.0.0

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:mvp-planner(MVP 策划师)
来源仓库:https://github.com/jiubanszd/mvp-planner
安装命令:
openclaw skills install mvp-planner
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 OpenClaw 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。

ClawHubOpenClaw
openclaw skills install mvp-planner

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。

将模糊产品想法转化为可行的启动方案。mvp-planner 属于开发类 Skill,可作为该场景下的辅助能力补充。适合新项目范围定义和 MVP 规划需求。输出包含功能清单、优先级和技术建议。适用宿主包括 OpenClaw,接入前应确认版本、权限和运行环境要求。调用时需提供核心功能和目标用户描述。

建议核对输出内容是否匹配实际开发能力边界。

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 OpenClaw;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 openclaw skills install mvp-planner;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

    它需要执行命令,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

mvp-planner 怎么安装?

在终端执行 openclaw skills install mvp-planner。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

mvp-planner 支持哪些 AI 工具?

公开资料标注支持 OpenClaw。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。

mvp-planner 需要什么权限?

权限级别是执行命令。安装流程涉及命令执行,可能通过 openclaw skills install mvp-planner 联网下载 Skill 或依赖。用户安装前应确认命令来源、仓库内容和执行环境。

mvp-planner 是免费的吗?

来源标注的许可证是 MIT-0。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

mvp-planner 用的人多吗?

公开数据是累计安装 2,747 次,近一周安装 118 次,下载 963 次。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。

mvp-planner 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

  • 专注 OpenClaw 一种宿主,触发方式和使用习惯不用来回切换。
  • 许可证明确为 MIT-0,商用和分发边界好判断。
  • 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

  • 需要执行命令,装到常用环境前要先评估影响面。
  • 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 OpenClaw,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。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

不太适合

  • 不能接受 Skill 在你机器上执行命令的人。
  • 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
mvp-planner

MVP 策划师 · 本页

2,747—执行命令MIT-0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

name
project-kickoff
description
Turn a vague product, webpage, app, workflow, automation, plugin, or skill idea into an actionable kickoff plan. Use when Codex needs to scope a new project from scratch, define the MVP, choose a stack, break work into phases, identify risks, or convert a rough idea into concrete build steps.

Project Kickoff

Overview

Use this skill to turn an early idea into a buildable plan without forcing the user to already know product, design, or engineering language.

Default to momentum. Make reasonable assumptions, state them, and only stop to ask questions when the answer changes architecture, cost, or delivery risk in a non-obvious way.

Core Workflow

Follow this order unless the user explicitly asks for only one part.

  1. Classify the project.
  2. Extract the real goal.
  3. Define the smallest valuable version.
  4. Choose the delivery shape.
  5. Break implementation into phases.
  6. Flag risks, unknowns, and decisions.
  7. End with the clearest next action.

1. Classify The Project

Map the request into one primary type before planning:

  • web-page: landing page, interactive site, dashboard, mini tool, portfolio
  • web-app: authenticated app, CRUD product, SaaS, internal tool
  • skill: reusable Codex skill with workflows, references, or scripts
  • plugin: Codex plugin or integration surface
  • automation: recurring task, scheduled workflow, inbox automation
  • script-tool: CLI helper, local utility, data processor
  • content-system: CMS-backed site, documentation hub, blog, knowledge base

If the request spans multiple types, choose the build target that delivers the first usable value and mention the other types as later phases.

2. Extract The Real Goal

Rewrite the request as a plain-language product goal:

  • Who is it for?
  • What frustrating job should it solve?
  • What should feel better after using it?
  • What makes it interesting enough to deserve building?

If the user gives only a theme, infer a likely user and use case. Mark that inference explicitly.

3. Define The MVP

Reduce the project to the smallest version that still feels complete.

For the MVP:

  • Include only the core loop
  • Prefer one polished flow over many partial features
  • Remove admin panels, auth, analytics, AI, payments, collaboration, and syncing unless they are central to the first value
  • Keep the MVP demonstrable in one sitting

Use this framing:

  • must-have: required for the first usable version
  • nice-to-have: useful but safely postponed
  • later: good future expansions

4. Choose The Delivery Shape

Recommend a concrete implementation shape, not a vague tech list.

Cover:

  • platform: static page, SPA, full-stack app, local script, skill folder, plugin scaffold
  • stack: HTML/CSS/JS, React/Vite, Next.js, Python, Node.js, etc.
  • data model: local state, JSON file, localStorage, SQLite, remote DB, or none
  • deployment shape: local only, static hosting, server deployment, Codex skill install

Bias toward the lowest-complexity stack that supports the MVP cleanly.

Examples:

  • Single interactive webpage -> static HTML/CSS/JS unless complexity clearly needs a framework
  • Small internal dashboard -> React + local mock data first
  • Reusable Codex planning helper -> skill with SKILL.md and optional references
  • Scheduled reporting assistant -> automation before full web app

5. Break Work Into Phases

Always produce a phased build plan with concrete outputs.

Use this default structure:

Phase 0

Clarify the concept, constraints, examples, and success criteria.

Phase 1

Build the MVP shell and the core interaction loop.

Phase 2

Polish UX, persistence, error states, and useful quality improvements.

Phase 3

Ship, validate, and add optional stretch features.

For each phase, include:

  • objective
  • deliverables
  • implementation notes
  • done-when check

6. Flag Risks And Decisions

Separate risk from uncertainty.

Use short bullets for:

  • technical risks: stack mismatch, data complexity, auth, performance, API dependency
  • product risks: too broad MVP, unclear audience, novelty without utility
  • decision points: where one choice affects architecture or long-term cost

When possible, pair each risk with a mitigation.

7. End With The Next Move

Always end with the most helpful next action for the current moment.

Choose one:

  • write the project scaffold
  • generate page structure and component list
  • draft the skill files
  • propose the database schema
  • create the implementation plan
  • start building immediately

If the user seems ready to build, do not stop at planning. Transition into execution.

Output Format

Use this structure unless the user asks for something shorter:

Project Summary

  • one short paragraph

Users And Goal

  • target user
  • user problem
  • expected outcome

MVP

  • must-have
  • nice-to-have
  • later

Recommended Stack

  • platform
  • tech choices
  • persistence or data model
  • deployment shape

Build Plan

  • phase 0
  • phase 1
  • phase 2
  • phase 3

Risks And Decisions

  • key risk bullets

Next Step

  • one clear recommended action

Project-Type Guidance

Read references/project-patterns.md when you need project-specific planning patterns, stack bias, or a sharper MVP cut for one of these types:

  • webpage or microsite
  • web app or dashboard
  • Codex skill
  • automation
  • script or CLI tool

Handling Ambiguity

When details are missing:

  • infer the simplest sensible audience
  • infer the smallest useful scope
  • prefer boring technology over ambitious architecture
  • surface assumptions in one short block

Ask follow-up questions only when the answer changes:

  • hosting or deployment model
  • whether user accounts are required
  • whether external APIs are mandatory
  • whether the work is a local utility or a public product
  • whether persistence must survive across devices

Quality Bar

The plan should be specific enough that another strong builder could start implementation immediately.

Avoid:

  • generic advice with no delivery shape
  • giant feature dumps
  • “it depends” without a recommendation
  • recommending frameworks because they are popular rather than necessary
  • overcomplicating a one-weekend project

Prefer:

  • concrete scope cuts
  • named phases
  • opinionated but reversible recommendations
  • short assumptions
  • one decisive next move

平台分布

OpenClaw

91.91%
按下载量换算885

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

通过

权限和风险

执行命令

安装流程涉及命令执行,可能通过 openclaw skills install mvp-planner 联网下载 Skill 或依赖。用户安装前应确认命令来源、仓库内容和执行环境。

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