- Home
- Skills
- Product Management
- Vibe Helper: Idea-to-Delivery PM Skill Suite
Vibe Helper: Idea-to-Delivery PM Skill Suite
一句话产品想法 → 需求访谈、Brief、竞品调研、用户声音、需求分层、框架、静态原型、交接的完整可开工交付包;1 全流程 + 5 单点 PM skill,中英双语。One orchestrator + 5 focused PM skills turning a one-sentence idea into a development-ready package, bilingual.
Free
Works with the AI Agents you already use
Vibe Helper: Idea-to-Delivery PM Skill Suite
Example session with this skill installed
我想做一个叫「错题本」的 App,帮考研学生拍照归档错题并复习,帮我从调研做到能开工。
- Read your context and instructions
- Compiled the vibe helper
一个 ./错题本/ 交付包:00-状态.md(滚动状态卡)到 07-技术框架.md(AI 轨道另含 08-评测方案.md),可点击的静态原型 prototypes/,backlog.md 与 badcase台账.md;每份文档第一屏是决策面板,论断标 A/B/C 证据等级,交付前 validate.sh 机器自检全部通过。
Connects securely to your tools. The creator never sees your data.
About this skill
从一句话产品想法到可直接开工的交付包:1 个全流程 orchestrator + 5 个单点 PM skill,中英双语。覆盖产品 Brief、竞品调研、用户声音、需求分层与技术框架;做 AI 产品自动进入 AI 轨道:Brief 承诺能力边界、产出人出答案的评测方案、交接 badcase 台账;全程 A/B/C 证据分级、决策面板,支持自主与委托两种决策模式,调研后敢拿证据劝你"别做了"。
English: from a one-sentence idea to a development-ready delivery package — one orchestrator plus five focused PM skills, bilingual (Chinese/English), with an AI track (capability boundaries, a human-answered eval plan, a badcase ledger), A/B/C evidence grading, decision panels, and a delegated decision mode that still makes you say "build or not" out loud.
套件组成 / What's included
- Vibe_helper — 8 阶段一条龙:需求访谈 → 产品 Brief → 竞品调研 → 用户声音 → 需求分层 → 产品与技术框架 → 静态原型 → 交接落地;4 道确认门、最薄一刀、范围守卫、交付 checklist 与 validate.sh 机器自检
- product-brief — Mom Test 问卷、产品 Brief、AI 轨道判定
- competitive-analysis — 竞品与开源调研、对标选择点、评测种子
- user-voice — 用户声音挖掘、需求真伪初判
- requirements-layering — MoSCoW + Gherkin 验收场景 + RICE 排序
- tech-framework — 最薄一刀、产品与技术框架、08-评测方案(AI 轨道)
全流程与单点读写同一套编号文件:跑一半可切出,单点产出也能被全流程复用;每个 skill 自包含,可单独取用。
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Product Management.
产出文档与高保真静态原型,不生成最终产品代码(Stage 8 交接给建站/全栈 skill 或开发者);调研质量取决于 agent 的联网/浏览器能力;生成的文件名与小节标题保留中文模板形式以便 validate.sh 机器自检,正文跟随用户语言。English: outputs documents and static prototypes, not final product code; research quality depends on the agent's web access; generated file names keep Chinese template forms for machine validation while bodies follow your language.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 1 day ago
- Free with an Agensi account
- Software
- Agent with web searchorbrowser
Git
- Optional
- Bash, to run validate.shA site/full-stack skill for Stage 8Claude Code for /slash commands
Stages 3–4 (research) need the agent to have web search or browser access; Stage 8 hands off to a site/full-stack skill when available — stages 1–7 work without one. Generated file names keep Chinese template forms for machine validation.
Permissions
File Scopes
Research stages use the agent's own web search/browser tools against public sites (GitHub, Product Hunt, Reddit, HN, Zhihu, V2EX, app stores); no fixed API hosts are called. bash only runs the bundled validate.sh self-check.