start

Design Partner Start Command

For Claude: Initialize a new design session by capturing context and generating initial ideas.

Process

1. Capture Context - Use AskUserQuestion with 3 questions to understand the design challenge:

   • Who are the users?
   • What job needs to be done?
   • What constraints exist?

2. Create Design Brief - Synthesize answers into structured brief

3. Save State - Write to .claude/dp.local.md

4. Generate Ideas - Invoke the idea-generator agent

Implementation

// Use AskUserQuestion to gather context
const questions = [
  {
    question: "Who are the primary users or audience?",
    header: "Target Users",
    placeholder: "e.g., College students, small business owners, healthcare workers..."
  },
  {
    question: "What job needs to be done? What problem are you solving?",
    header: "Job to Be Done",
    placeholder: "e.g., Stay focused while studying, Track fitness progress, Manage client appointments..."
  },
  {
    question: "What constraints should we consider?",
    header: "Constraints",
    placeholder: "e.g., Mobile-first, Works offline, Limited budget, Must integrate with existing systems..."
  }
];

State Schema

Create .claude/dp.local.md with:

---
session_id: "dp-[timestamp]"
project_name: "[from args or generated]"
created_at: "[ISO timestamp]"
last_updated: "[ISO timestamp]"

design_brief:
  users: "[answer 1]"
  job_to_be_done: "[answer 2]"
  constraints: "[answer 3 as array]"

ideas: []
selected_idea: null

image_generation:
  enabled: false
  default_provider: "openai"
  fallback_order: ["openai", "gemini"]
  default_size: "1024x1024"
  default_quality: "standard"
  daily_budget_usd: 5.00
  warn_threshold_usd: 0.50
  current_daily_spend: 0.00
  last_reset_date: "[today's date]"
  providers:
    openai:
      configured: false
      enabled: false
      last_used: null
      total_images: 0
      total_cost_usd: 0.00
    gemini:
      configured: false
      enabled: false
      last_used: null
      total_images: 0
      total_cost_usd: 0.00

generated_images: []
decisions: []
---

# Design Brief

**Users:** [users]

**Job to Be Done:** [job]

**Constraints:**
- [constraint 1]
- [constraint 2]
- [constraint 3]