SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillssuper-pm

super-pm

Office & productivity

View repository

Updated 10-16 06:22

About this skill

Startup Principles

Read existing project materials according to this round's user task; do not automatically create docs/ directories or check for updates online when activating the skill. For version checks and upgrades, explicitly use /super-pm-upgrade. The path is determined by the actual location of the installed skill, rather than inferring the skill file path from a pasted Bash code block.

Routing and Delivery Boundaries

Explicit user intent takes priority. For complete new product planning, you may first recommend /pm-brainstorm, but explicit /pm-demand and other local tasks enter the corresponding skill, which checks for missing evidence itself. Existing relevant materials can be reused, with no need to ask repeated questions. When the user is only discussing and not writing files, provide suggestions only in chat; do not create directories or documents unless the user later requests artifacts.

Cross-Agent Interaction Rules

When the process requires interaction with the user:

  1. If the current environment supports AskUserQuestion, use AskUserQuestion (best experience).
  2. If the current environment does not support AskUserQuestion, you must ask the same question via a regular chat message.
  3. Ask only one question at a time.
  4. After asking, you must stop the current turn and wait for the user's answer.
  5. When only discussing, do not write files; when key input is needed, ask first and wait.
  6. Existing relevant and valid materials can be reused, with no need to ask repeated questions; irrelevant documents cannot substitute for evidence.

Intent Recognition (candidate skills, not mandatory)

When the following keywords are detected, proactively guide the user to start the process:

Product category: product, feature, functionality, module, new product, business model, community, app, mini program, platform, website, tool, SaaS Requirements category: requirement, pain point, user need, requirements pool, requirements document, sort out requirements, collect requirements, validate requirements User category: user, target user, user persona, user experience, user feedback, consumer, customer Market category: market, competitor, industry, market size, track, trend Evidence category: verify sources, evidence audit, data authenticity, basis for conclusions → /pm-evidence-audit; for PRD completeness and feasibility review, use /pm-prd-review Scenario category: from 0 to 1, start from scratch, new project, MVP, I want to build, help me plan, help me design, discuss product, write requirements document, write PRD, write BRD, write product document, review PRD, GEO optimization, AI search optimization, competitor monitoring, user interview, OKR, goal management Motion category: motion, animation, brand animation, logo animation, motion design, brand motion, opening animation, splash, loading animation, hover motion, microinteraction, motion direction, motion demo Strategy category: multiple product lines, product portfolio, resource allocation, strategic decision Perspective category: from Jobs's perspective, how would Jobs see it, Jobs mode, steve jobs perspective, switch to Jobs, if Jobs, Jobs thinking, expression DNA

Response After Triggering

Scenario 1: New Product Planning (vague idea, needs divergence)

Recognition Signals: Mentions a specific product form but the requirements are unclear. For example, "I want to build an XX", "help me plan an XX product", "discuss the product direction", "XX community/app/mini-program".

Use AskUserQuestion to provide options:

🎯 Detected that you are planning a new product: "{user-described product}"

The product direction is still unclear. It is recommended to first brainstorm to diverge your thinking and clarify the user group and core pain points.

A) Start brainstorming (recommended) - run /pm-brainstorm B) Go directly to demand research (only if you have already completed brainstorming or explicitly request to skip) - run /pm-demand C) Not needed, I have other questions

Scenario 2: Clear direction, needs research validation

Recognition signals: The user has expressed a clear demand direction or a specific validation intent. For example, "validate this demand", "research the pain points of XX users", "analyze the target users".

Use AskUserQuestion to provide options:

🎯 Detected that you need demand research!

A) Start demand research (recommended) - run /pm-demand B) Do market analysis first - run /pm-market C) Not needed, I have other questions

Scenario 3: Want to write a document (check prerequisites)

Recognition signals: The user wants to "write an XX document" but has not yet completed the prerequisite research. For example, "write a requirements document", "write a PRD", "write a product document", "write a BRD".

⚠️ Important rule: "Write a document" is an output intent, not the starting point of the process. You must first check whether the prerequisite documents exist:

  • Write a requirements document/PRD → Prerequisites: demand research report, brainstorming plan
  • Write a BRD → Prerequisites: demand research report, market analysis report

🎯 Detected that you want to write a "{document type}", but you need to complete the prerequisite research first:

📊 Current progress: {list existing/missing prerequisite documents}

A) Start from the beginning - run /pm-brainstorm → /pm-demand → finally produce the document B) I already have complete research, write the document directly - run /pm-docs C) Not needed, I have other questions

Scenario 4: Multi-product line management

🎯 Detected that you are managing multiple products!

A) Start product portfolio analysis (recommended) - run /pm-portfolio B) Resource allocation decision - run /pm-resource C) Not needed, I have other questions

Scenario 5: Brand motion needs

Recognition signals: The user mentions motion, animation, logo animation, brand motion, opening animation, splash, loading animation, hover motion, etc.

🎯 Detected that you need brand motion design!

A) Start brand motion plan (recommended) - run /pm-brand-motion B) Complete product positioning first - run /pm-position C) Not needed, I have other questions

Scenario 6: Expert perspective / clash of ideas

Recognition signals: The user mentions "from Jobs's perspective", "how would Jobs see it", "Jobs mode", "steve jobs perspective", "switch to Jobs", "if Jobs", "Jobs thinking", "expression DNA", or hopes to examine a product/decision/idea using a certain expert's thinking framework.

🎯 Detected that you want to invoke an expert perspective!

A) Enter the Steve Jobs thinking advisor (recommended) - run /steve-jobs-perspective B) Not needed, I have other questions

Process recommendation engine

Golden path (when a complete process is needed)

New product from zero to one: /pm-brainstorm → /pm-demand → /pm-market → /pm-priority → /pm-mvp → /pm-docs → /pm-tech → /pm-release

For a complete new product process, brainstorm is recommended by default; when there is an explicit partial task, relevant existing materials, or a user request to skip, you may go directly to the target skill, and mark the steps that have not yet been verified.

Existing Product Optimization and Growth: /pm-aarrr → /pm-growth → /pm-report → /pm-iteration

Strategic Decision-Making (Standalone Use): /pm-business-model → /pm-decision → /pm-portfolio → /pm-resource

Recommend based on current state

Needs Insight Stage：

  • Needs research not completed → recommend /pm-demand
  • Needs completed but no market analysis → recommend /pm-market
  • Market data available but no prioritization → recommend /pm-priority
  • Prioritization available but no MVP → recommend /pm-mvp

Solution Design Stage：

  • MVP completed but no PRD → recommend /pm-docs
  • PRD completed but no prototype → recommend /pm-proto
  • Prototype completed but no brand motion → recommend /pm-brand-motion
  • Brand motion completed but no technical solution → recommend /pm-tech
  • Technical solution completed but no data metrics → recommend /pm-data

Growth Iteration Stage：

  • Product launched → recommend /pm-aarrr
  • Growth analysis completed → recommend /pm-growth
  • Data report available → recommend /pm-iteration

Risk Control Management Stage：

  • Technical solution completed → recommend /pm-risk
  • Risk control completed → recommend /pm-release

Multi-option recommendations

After each skill is completed, provide 2-3 next-step options.

Usage Examples

Example 1: New product from 0 to 1

用户: 我要做一个生鲜电商产品

AI: 🎯 检测到您在规划新产品！
    建议先从头脑风暴开始，明确目标用户和核心痛点。
    是否开始执行 /pm-brainstorm？

Example 2: User wants to write a document but has no prior research

用户: 写个需求文档，关于某个垂直社区产品

AI: 🎯 检测到您要写需求文档，但当前缺少前置调研。
    
    📊 前置依赖检查：
    ❌ 头脑风暴方案
    ❌ 需求调研报告
    
    💡 建议从头走完整流程：
    1. /pm-brainstorm - 明确产品定位和核心场景
    2. /pm-demand - 系统化调研验证需求
    3. /pm-docs - 基于调研产出需求文档
    
    是否从 /pm-brainstorm 开始？

Example 3: Continue across sessions

用户: 继续

AI: 📊 检测到您的进度：
    ✅ 需求调研
    ✅ 市场分析
    ⏳ 优先级排序

    💡 建议继续 /pm-priority

Notes

  1. Risk front-loading: Validate pain points immediately during the needs research stage
  2. Search dependency notes(the skill pack itself does not include a built-in search tool; it relies on external skill/MCP enhancement):
    • AnySearch(standalone skill, not a built-in tool) — best Chinese search quality, supports 17 vertical domains (finance, law, academia, etc.). Installation:git clone https://github.com/konglong87/anysearch.git ~/.claude/skills/anysearch When not installed, it automatically degrades without affecting process execution.
    • Exa MCP(MCP server) — English/semantic search enhancement. Must be configured in the project .claude/settings.json configure mcpServers. When not configured, it degrades to WebSearch.
    • WebSearch(if provided by the current platform) — as the search fallback; if unavailable, you should truthfully inform the user rather than assume success.
    • Priority: AnySearch → Exa MCP → WebSearch, no skipping levels. When using WebSearch, mark "⚠️ degraded mode"
  3. Fallback mechanism: Provide alternatives when files are missing
  4. Markdown storage: All documents are human-readable and editable
  5. Process recommendations: After each skill is completed, recommend 2-3 next steps
  6. View all skills: Open INDEX.md to view the complete skill index
  7. Pre-check: When the user wants to "write document XX", first check whether prerequisite dependencies (needs research, brainstorming, etc.) exist; if missing, guide them to start the process from the beginning

Super-PM - Give one product manager the capabilities of a product team

Common Misconceptions / Red Flags — STOP

Stop immediately and backtrack when the following situations occur:

Misconception Correct approach
Drawing conclusions using "should", "probably", "seems" Must be based on actual data and verification
Claiming completion without running checks Verify first, then state
The user explicitly requested a lightweight process Respect the scope, flag unverified assumptions, and do not fabricate completed steps

Output Quality Check / Verification Checklist

  • Process recommendation engine has been triggered
  • The corresponding skill has been recommended based on the current project status
  • The user has been informed of the next optional paths

⚠️ Ensure the user receives clear next-step recommendations.