name: "Solo Founder All-Hands"
description: "One sentence in, 8 AI 'departments' collaborate, full launch plan out in 2 minutes — your one-person company"

agents_dir: "agency-agents"

llm:
  provider: deepseek
  model: deepseek-chat

concurrency: 4

inputs:
  - name: idea
    description: "Your one-sentence idea (e.g. a tool that auto-generates invoices for freelancers)"
    required: true

steps:
  # ── Layer 1: CEO sets direction ──
  - id: ceo_kickoff
    role: "strategy/nexus-strategy"
    name: "CEO"
    emoji: "👔"
    task: |
      You are the CEO of this one-person company. Kick off the all-hands.

      Founder's idea: {{idea}}

      Output:
      1. Product positioning in one sentence
      2. What pain point we're solving
      3. Who the target user is
      4. What success looks like in 6 months

      Tone: like a CEO at a morning standup — concise, concrete, no corporate fluff.
    output: vision

  # ── Layer 2: four departments research in parallel ──
  - id: market_research
    role: "product/product-trend-researcher"
    name: "Market Researcher"
    emoji: "📊"
    task: |
      The CEO just set direction: {{vision}}

      As the trend researcher, do rapid market research:
      1. Market size and growth rate for this category
      2. Top 3 competitors + their strengths/weaknesses
      3. Where the market gap is
      4. Trend forecast for the next 12 months

      Every point must be concrete — real numbers, real reasoning, no generic filler.
    output: market_insight
    depends_on: [ceo_kickoff]

  - id: user_research
    role: "design/design-ux-researcher"
    name: "User Researcher"
    emoji: "🔍"
    task: |
      The CEO just set direction: {{vision}}

      As the user researcher, deliver user insights:
      1. Core user personas (2 typical users, specific about age, profession, mindset)
      2. How they currently solve this problem (existing solutions and their flaws)
      3. Top 3 pain points (scene-specific, not abstract)
      4. Key trigger that makes them willing to pay
    output: user_insight
    depends_on: [ceo_kickoff]

  - id: tech_feasibility
    role: "engineering/engineering-backend-architect"
    name: "Tech Lead"
    emoji: "💻"
    task: |
      The CEO just set direction: {{vision}}

      As the tech architect, assess feasibility:
      1. Recommended stack (something one person can own — explain why this not that)
      2. MVP architecture (clear module relationships, no over-engineering)
      3. 4-week build plan (week-by-week with deliverables)
      4. Real technical risks and mitigations (the ones you'll actually hit, not textbook lists)
    output: tech_plan
    depends_on: [ceo_kickoff]

  - id: brand_design
    role: "design/design-brand-guardian"
    name: "Brand Director"
    emoji: "🎨"
    task: |
      The CEO just set direction: {{vision}}

      As the brand designer, deliver the brand foundation:
      1. Product name candidates (3 options + rationale; memorable and domain-available)
      2. Brand tone and keywords
      3. Slogan candidates (3 options; conversational and sticky)
      4. Visual direction (palette + typography with specific hex codes)
    output: brand_plan
    depends_on: [ceo_kickoff]

  # ── Layer 3: three planning functions run in parallel on research ──
  - id: product_plan
    role: "product/product-manager"
    name: "Product Manager"
    emoji: "🧭"
    task: |
      Research is in:

      **Market**: {{market_insight}}
      **User**: {{user_insight}}

      As the PM, output MVP plan:
      1. 3 core features (who uses each, what problem it solves)
      2. Top 5 user stories
      3. Priority + tradeoffs (explain what you're cutting and why)
      4. Launch criteria (concrete, measurable definition of "usable")
    output: product_spec
    depends_on: [market_research, user_research]

  - id: marketing_plan
    role: "marketing/marketing-content-creator"
    name: "Marketing Lead"
    emoji: "📣"
    task: |
      Research is in:

      **Market**: {{market_insight}}
      **User**: {{user_insight}}
      **Brand**: {{brand_plan}}

      As the marketing lead, output a cold-start plan:
      1. Launch copy (hooky headline, shareable body)
      2. Top 3 launch channels + tactics (specific, not generic)
      3. Zero-budget growth strategy (executable — no "make great content" fluff)
      4. Week-by-week goals and actions for month one
    output: marketing_spec
    depends_on: [market_research, user_research, brand_design]

  - id: finance_plan
    role: "finance/finance-fpa-analyst"
    name: "CFO"
    emoji: "💰"
    task: |
      Research is in:

      **Market**: {{market_insight}}
      **Tech**: {{tech_plan}}

      As the financial advisor, output the business model:
      1. Pricing strategy (free vs paid tiers, pricing logic)
      2. Cost structure (monthly fixed costs, line-by-line)
      3. Revenue forecast (3/6/12 months with key assumptions and math)
      4. Break-even analysis (how many users / how much revenue to survive)
    output: finance_spec
    depends_on: [market_research, tech_feasibility]

  # ── Layer 4: CEO decides ──
  - id: ceo_decision
    role: "strategy/nexus-strategy"
    name: "CEO"
    emoji: "👔"
    task: |
      All-hands reports are in:

      **Market**: {{market_insight}}
      **User**: {{user_insight}}
      **Tech**: {{tech_plan}}
      **Brand**: {{brand_plan}}
      **Product**: {{product_spec}}
      **Marketing**: {{marketing_spec}}
      **Finance**: {{finance_spec}}

      As the CEO, make the final call. Output "One-Person Company Launch Plan":

      ## Decision
      Go or No-Go? One-sentence reason.

      ## Product Name
      Pick one from brand options and explain why.

      ## Week One Action List
      Day-by-day, concrete enough to execute.

      ## Time Allocation
      How to split your time (product / engineering / marketing %).

      ## Risk Watch
      Top 3 risks + mitigation.

      ## Note to Self
      As a solo founder, what to hold on to and what to avoid.

      Every section must have a clear decision and rationale. No hedging. Action items must be executable — no "do great work" filler.
    output: launch_plan
    depends_on: [product_plan, marketing_plan, finance_plan]