product-engineer-agent（product 工程师 Agent）

Product Engineer Agent

Design and develop new product concepts with comprehensive specifications.

This skill uses 5 specialized agents that analyze product ideas from different engineering perspectives, then synthesizes into a complete product specification.

What It Produces

Prerequisites

• GOOGLE_API_KEY - For generating product visuals (uses image-generation skill)
• Works with any product category

Workflow

Step 1: Gather Product Idea (REQUIRED)

⚠️ DO NOT skip this step. Use interactive questioning — ask ONE question at a time.

Question Flow

⚠️ Use the AskUserQuestion tool for each question below. Do not just print questions in your response — use the tool to create interactive prompts with the options shown.

Q1: Problem

"I'll help you design this product! First — what problem does it solve? *(The core user need)*"

*Wait for response.*

Q2: User

"Who is the target user? *(Who will use this product?)*"

*Wait for response.*

Q3: Features

"Any must-have features or key requirements? *(Or say 'help me figure it out')*"

*Wait for response.*

Q4: Constraints

"Any constraints to consider? - Budget range - Size/form factor - Materials - Manufacturing method - Or describe"

*Wait for response.*

Q5: Visuals

"Do you want me to generate visuals? - 🎨 Concept renders (what it looks like) - 🔧 Engineering drawings (exploded views, cross-sections) - Both - No visuals (spec document only)"

*Wait for response.*

Quick Reference

Parse visual preferences:

• "yes", "visuals", "show me", "render", "drawings" → Generate all visuals
• "concept only" → Just concept render + lifestyle
• "engineering only" → Just exploded/technical views
• "no visuals" or not mentioned → Skip visual generation
• Unclear → Default to generating visuals (they add value)

Step 2: Run Specialized Engineering Agents in Parallel

Deploy 5 agents, each analyzing from a different perspective:

Agent 1: Industrial Designer

Focus: Form, ergonomics, aesthetics, user interaction

Consider:
- Physical form factor and dimensions
- Ergonomics and human factors
- Visual aesthetics and brand expression
- User interaction points (buttons, displays, etc.)
- Packaging and unboxing experience

Agent 2: Mechanical Engineer

Focus: How it works, materials, mechanisms

Consider:
- Core mechanism / how it functions
- Materials selection (strength, weight, cost)
- Manufacturing feasibility
- Durability and lifecycle
- Assembly and serviceability

Agent 3: User Researcher

Focus: User needs, pain points, usability

Consider:
- User journey with the product
- Pain points addressed
- Potential usability issues
- Onboarding and learning curve
- Accessibility considerations

Agent 4: Manufacturing Advisor

Focus: Feasibility, cost, production

Consider:
- Manufacturing methods (injection molding, CNC, etc.)
- Tooling requirements and costs
- Unit cost estimates at various volumes
- Supply chain considerations
- Quality control points

Agent 5: Innovation Scout

Focus: Existing solutions, patents, differentiation

Consider:
- Similar products in market
- Patent landscape (potential conflicts)
- Unique differentiators
- Technology trends to leverage
- Blue ocean opportunities

Step 3: Synthesize into Product Specification

Combine all agent outputs into a structured specification:

{
  "product": {
    "name": "Product Name",
    "tagline": "One-line description",
    "problem_solved": "Core problem it addresses",
    "target_user": "Who it's for",
    "category": "Product category"
  },
  "design": {
    "form_factor": "Physical description",
    "dimensions": "L x W x H",
    "weight": "Estimated weight",
    "materials": ["Material 1", "Material 2"],
    "colors": ["Primary options"],
    "key_interactions": ["How users interact with it"]
  },
  "features": {
    "must_have": [
      {"feature": "Feature 1", "rationale": "Why it's essential"}
    ],
    "should_have": [
      {"feature": "Feature 2", "rationale": "High value add"}
    ],
    "could_have": [
      {"feature": "Feature 3", "rationale": "Nice to have"}
    ]
  },
  "technical": {
    "mechanism": "How it works",
    "power_source": "Battery/plug/manual/etc.",
    "electronics": "Any electronic components",
    "software": "Any software/firmware needed"
  },
  "manufacturing": {
    "primary_method": "Main manufacturing process",
    "estimated_bom": [
      {"component": "Part 1", "estimated_cost": "$X"}
    ],
    "unit_cost_estimates": {
      "100_units": "$XX",
      "1000_units": "$XX",
      "10000_units": "$XX"
    },
    "complexity": "Low/Medium/High"
  },
  "market": {
    "similar_products": ["Competitor 1", "Competitor 2"],
    "differentiators": ["What makes this unique"],
    "price_positioning": "Budget/Mid/Premium",
    "target_msrp": "$XX"
  },
  "next_steps": [
    "1. Validate with potential users",
    "2. Create detailed CAD model",
    "3. Build first prototype",
    "4. Patent search (if applicable)"
  ]
}

Step 4: Generate Product Visuals (If Requested)

Only generate visuals if user requested them in Step 1.

If user wants visuals, generate using the image-generation skill:

From Industrial Designer:

• product_concept.png - Main concept render (studio lighting, clean background)
• product_context.png - Product in-use/lifestyle shot

From Mechanical Engineer:

• product_exploded.png - Exploded view showing all components
• product_section.png - Cross-section (if internal mechanism is key)

Visual Generation Order:

1. Concept render first (shows overall design)
2. In-context shot (shows usage)
3. Exploded view (shows engineering)
4. Cross-section (if needed)

If user didn't request visuals: Skip to Step 5 with spec document only.

Step 5: Deliver Complete Package

Delivery message (with visuals):

"✅ Product design complete!

Product: [Name] Problem: [What it solves] Key Differentiator: [What makes it unique]

Estimated unit cost: $XX at 1,000 units Suggested MSRP: $XX

Generated visuals:

• Concept render ✓
• Lifestyle/context shot ✓
• Exploded assembly view ✓

Next steps:

1. [First recommended action]
2. [Second recommended action]

Want me to:

• Deep dive on any section?
• Generate additional views or angles?
• Explore alternative designs?
• Estimate costs for different volumes?
• Compare to specific competitors?"

Delivery message (spec only, no visuals):

"✅ Product specification complete!

Product: [Name] Problem: [What it solves] Key Differentiator: [What makes it unique]

Estimated unit cost: $XX at 1,000 units Suggested MSRP: $XX

Next steps:

1. [First recommended action]
2. [Second recommended action]

Want me to:

• Generate visuals? (concept renders, engineering drawings)
• Deep dive on any section?
• Explore alternative designs?
• Estimate costs for different volumes?"

Integration with Other Skills

This skill works well with:

Generate PDF Report

After completing the product specification, offer to generate a PDF:

"Would you like me to generate a PDF report of this product specification?"

python3 ${CLAUDE_PLUGIN_ROOT}/skills/media-utils/scripts/report_to_pdf.py \
  --input product_spec.md \
  --output product_spec.pdf \
  --title "Product Specification" \
  --style technical

Agents

Output Files

When generating a complete product design, you'll receive:

product_spec.md           ← Complete specification document
product_concept.png       ← 3D concept render
product_context.png       ← Lifestyle/in-use shot
product_exploded.png      ← Exploded assembly view
product_section.png       ← Cross-section (if applicable)

Example Prompts

Basic:

"Design a new portable phone charger that's more convenient"

With context:

"I want to create a kitchen gadget that helps with meal prep. Target audience is busy parents. Budget under $30 retail."

Iteration:

"Take my existing product idea and suggest improvements: [description]"

Competitive:

"Design something better than [competitor product]"

With visuals:

"Design a smart water bottle and show me what it would look like"

Engineering focus:

"Design a modular desk organizer and show me the exploded assembly view"